| |
Main
Date: 06 Sep 2008 00:57:46
From: Sanny
Subject: Pls analyze this game of Beginner Level
|
What do you say, What were the wrong moves by Beginner that it should
have avoided?
Game Played between Jester and beginner at GetClub.com
--------------------------------------------------------------------------------
Jester : (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM26050&game=Chess
--------------------------------------------------------------------------------
White -- Black
(Jester ) -- (beginner)
1. e2-e4{8} d7-d5{0}
2. e4-d5{6} Qd8-d5{0}
3. Nb1-c3{6} Qd5-d8{0}
4. Bf1-b5{10} c7-c6{18}
5. Bb5-c4{10} Ng8-f6{8}
6. d2-d3{18} Bc8-g4{82}
7. f2-f3{12} Bg4-e6{6}
8. Bc4-e6{12} f7-e6{8}
9. Qd1-e2{10} Qd8-d6{8}
10. Ng1-h3{10} Nb8-d7{8}
11. Ke1-g1{12} g7-g6{8}
12. Nh3-g5{16} e6-e5{6}
13. Kg1-h1{12} Bf8-g7{6}
14. Nc3-e4{12} Nf6-e4{8}
15. Ng5-e4{10} Qd6-d4{6}
16. Bc1-e3{12} Qd4-d5{10}
17. Qe2-f2{12} b7-b6{8}
18. Qf2-h4{10} Nd7-f6{20}
19. Be3-d2{12} Nf6-e4{6}
20. f3-e4{10} Qd5-d4{8}
21. Bd2-c3{14} Qd4-e3{12}
22. Qh4-g4{12} Bg7-f6{24}
23. Qg4-e6{12} Qe3-g5{0}
24. Rf1-f2{10} Ke8-d8{6}
25. Qe6-c6{12} Ra8-b8{8}
26. Ra1-f1{12} Rh8-g8{14}
27. Qc6-d5{12} Kd8-c8{28}
28. Qd5-g8{10} Kc8-b7{10}
29. Qg8-d5{16} Kb7-c8{12}
30. a2-a4{14} h7-h5{14}
31. a4-a5{28} Rb8-b7{20}
32. a5-b6{14} a7-b6{0}
33. Qd5-c6{10} Rb7-c7{6}
34. Qc6-b6{10} Rc7-d7{8}
35. Rf1-a1{12} Rd7-c7{10}
36. Qb6-b5{10} Rc7-a7{18}
37. Qb5-e8{14} Kc8-b7{6}
38. Qe8-d7{14} Kb7-b8{0}
39. Qd7-a7{14} Kb8-c8{0}
40. Qa7-a8{10} Kc8-c7{2}
--------------------------------------------------------------------------------
Jester : (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM26050&game=Chess
Please let me know your opinion about the above moves.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
|
| |
Date: 17 Sep 2008 04:21:03
From: Simon Krahnke
Subject: Re: Another game to analyze.
|
* Guest <[email protected] > (03:46) schrieb:
> It's something I used back in my 8 bit micro days. I just got fond of it
> and kept using it. [...]
Thanks for the explanation. I think I read something like that in the wiki.
>> The new Intel 6core will have 16MB 3rd level cache. I guess, one could
>> fit a whole engine in there. I need more money.
>
> Ir probably *nneds* that much cache just to get tolerable performance.
Sure, they won't waste transistors on caches they won't need. But things
like chess engines might a good abuse for that.
> I'm not sure I'd want one though...
>
> So little of what I do would be able to use all 6 cores that it'd just be
> extra electrical power being wasted, along with the extra money it'd cost to
> buy it.
If you don't need it, you can switch it off. They are more efficient, so
electrical power isn't an issue for the tasks that need it. The only
relevant factor is the money to buy it (and the space it takes up),
that's why I need more money.
mfg, simon .... right now running 3 computers w/ 6 cores
|
| |
Date: 17 Sep 2008 02:42:48
From: Simon Krahnke
Subject: Re: Another game to analyze.
|
* Guest <[email protected] > (2008-09-16) schrieb:
> Speaking of which, that reminds me of the JPC emulator. It stands for "Java
> PC emulator".
Maybe I should try that some time.
> Of course, where as DosBox emulates DOS as well, JPC just emulates the
> hardware and you need an OS to go along with it. So you'll also have to
> include FreeDos or OpenDos or ReactOS, in addition to your program.
Also still have to try ReactOS.
> I used to prefer the 64 element map / unmap approach.
>
> Occasionally I'd do a 16x16 for fun, but I was comfortable with the map /
> unmap approach.
Could you explain that? I never heard these terms before.
>> I didn't find anything about attacks on the wiki.
>
> It may not have it. I haven't checked.
>
> The last I really saw anything about the attack lists etc. was on the old
> CCC board.
>
> Try the archives.
>
> http://www.stmintz.com/ccc/
>
> RGCC archives might have some, but I can't say how much.
So much on the to-do-list, I rather go ahead taking my current approach
and see how far it goes.
> Unless you are working on a 4 bit processor or something like an 8051 with
> embedded memory, I don't see much difference between 512 and 4k. Not when
> you then start talking about 100k programs, 256meg trans tables, History
> Heuristic, Ref. tables, etc.
Memory issues are mostly about caching, there is plenty RAM, but it's
too slow. Cache sizes still aren't very big.
The new Intel 6core will have 16MB 3rd level cache. I guess, one could
fit a whole engine in there. I need more money.
mfg, simon .... l
|
| | |
Date: 16 Sep 2008 20:46:43
From: Guest
Subject: Re: Another game to analyze.
|
"Simon Krahnke" <[email protected] > wrote in message
news:[email protected]...
>* Guest <[email protected]> (2008-09-16) schrieb:
>
>> I used to prefer the 64 element map / unmap approach.
>>
>> Occasionally I'd do a 16x16 for fun, but I was comfortable with the map /
>> unmap approach.
>
> Could you explain that? I never heard these terms before.
It's something I used back in my 8 bit micro days. I just got fond of it
and kept using it.
Your board is a 64 element array.
You use two additional arrays.
Map[64] maps the 64 element board square onto a 120 array.
UnMap[120] maps the 120 element array back onto the 64 element indexing
method. Any illegal square gets returned as -1.
So to do a King or Knight move generation, it'd looke like:
NextOff=&Offsets[Piece];
MSqr=Map[Sqr];
do
{
if ((u = UnMap[MSqr + *NextOff]) >= 0)
if (BoardColor[u] != Side)
StoreMove(pos, u);
} while (*(++NextOff));
I'm not saying it's the best method.
It worked well on the 8 bit micro's I was using and I just kept using it for
years and years.
I didn't like 0x88 method and there wasn't any reason to go back to a plain
10x12 or 16x12 method.
It's not too different from the 0x88 method in concept except we are trading
a bit check for an array access. Depending on the cpu's instructions (6502
etc.) it can be easier to do the array access and branch on negative.
And actually we don't even need the 64 arrays. We could work directly on
the 10x12 directly and the 'UnMap' is just an array holding its own indexes
except for the border being -1.
These days I probably wouldn't use it in a new program, but... (shrug)
>> Unless you are working on a 4 bit processor or something like an 8051
>> with
>> embedded memory, I don't see much difference between 512 and 4k. Not
>> when
>> you then start talking about 100k programs, 256meg trans tables, History
>> Heuristic, Ref. tables, etc.
>
> Memory issues are mostly about caching, there is plenty RAM, but it's
> too slow. Cache sizes still aren't very big.
Even the AMD x2's are 64k L1 data cache.
Excluding Trans table & maybe the History Heuristic, that'd be enough for
nearly all the data to stay in the cache.
The move list might overflow that a bit, but the L2 would catch that.
The old rotated bitboard method used a lot of data and was a problem.
But regular boards and some of the memory efficient bitboard methods use a
small enough amount that it and probably most of your other data will fit
into the L1 cache.
> The new Intel 6core will have 16MB 3rd level cache. I guess, one could
> fit a whole engine in there. I need more money.
Ir probably *nneds* that much cache just to get tolerable performance.
I'm not sure I'd want one though...
So little of what I do would be able to use all 6 cores that it'd just be
extra electrical power being wasted, along with the extra money it'd cost to
buy it.
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| |
Date: 15 Sep 2008 19:39:25
From: Simon Krahnke
Subject: Re: Another game to analyze.
|
* Guest <[email protected] > (2008-09-14) schrieb:
> "Simon Krahnke" <[email protected]> wrote in message
> news:[email protected]...
>>* Guest <[email protected]> (22:12) schrieb:
>>
>>> "Simon Krahnke" <[email protected]> wrote in message
>>> news:[email protected]...
>>>>* Guest <[email protected]> (16:05) schrieb:
>
>>> Heck, if you want to cheat a little, provide a DOS based executable and
>>> just
>>> distribute that. People can use something like DosBox to run it.
>>
>> If you replace DOS by Win32 and Dosbox by Wine that's the trick Rybka
>> uses.
>
> I thought of that, but I wasn't sure about Wine on the Mac.
I wasn't sure too. But Wikipedia tells me that Wine for Mac OS X is "well
maintained". That probably means Intel Macs, since Wine doesn't include
any processor emulation.
> Also DOS & DOSBox have the advantage that they can work on non-x86
> platforms, such as the old Macs. (At a factor of 10 performance or so, but
> it'll work.)
But that probably isn't much better than Java.
>>>>> 2) threads / processes. Used for multi-cpu programming.
>>>>
>>>> Can also be used to seperate communication from the search.
>>>
>>> It can, but doesn't have to.
>>
>> It's can be easier that way. You got clean seperation between the two
>> issues.
>
> Dealing with processes & threads can be a bit more complicated among
> platforms.
Sure. That's why you mentioned it in this list.
But once this is done, it can also solve the input problem.
It's very easy in Java.
> It's easier though.... And for what little extra benefit you'll get, it may
> not be worth the trouble. Especially in the beginning.
Right, when you've never done a single threaded search, don't try a
multi threaded one.
> I've also repeated read that it's much easier to use two processors than
> four. About any method will do okay with just two, but as soon as it's run
> on a quad, you might actually get a slow-down.
>
> So it might be a good idea to do development on a quad isntead of the more
> common & cheaper duals.
Good, I don't own any dual cores.
>>> a very strong, very fast program with mailbox style.
>>
>> Let's hope so. I'm just taking a try on this.
>
> Oh you can. But it will require some creativity.
>
> I admit I was never that creative though. In fact, I was never really that
> fond of 0x88 even. When I used to do mailbox, I used to do classic 120 or
> 16x16 based.
Finalfun uses 0x88.
> It'll involve piece lists at a minimum, of course. Then I've read people
> talking about maintaining lists that can determine attacks. I wasn't paying
> much attention, but I think it's an aspect of 0x88 boards and some other
> data structures & piece lists.
I didn't find anything about attacks on the wiki.
> I was just meaning portability, though. GCC is so portable and so available
> that it's feasible to use it for whatever platform you want. Which will
> likely be x86 based.
Fine.
> Modern bitboard methods, like Gerd's Kindergarten method only take a few K
> for their data to work. (Exact amounts depend on how you do it.) Far less
> than other methods. And they give tolerable performance. (There are faster
> methods, but they use a lot more data. Personally I don't think it's a good
> trade-off.)
>
> Much less data than the old rotated method.
But still way more data than without bitboards.
mfg, simon .... l
|
| | |
Date: 15 Sep 2008 18:49:37
From: Guest
Subject: Re: Another game to analyze.
|
"Simon Krahnke" <[email protected] > wrote in message
news:[email protected]...
>* Guest <[email protected]> (2008-09-14) schrieb:
>> Also DOS & DOSBox have the advantage that they can work on non-x86
>> platforms, such as the old Macs. (At a factor of 10 performance or so,
>> but
>> it'll work.)
>
> But that probably isn't much better than Java.
True.
But that's worst case (for non-x86 hardware), where as with Java, it
probably closer to average.
Speaking of which, that reminds me of the JPC emulator. It stands for "Java
PC emulator".
They claim that it runs about 20% of native speed.
Imagine that... An emulator in Java that can emulate the x86 and PC
hardware at about 1/5th the speed of native.
Pretty darn good. Emulators such as Qemu, DosBox do about that well.
I haven't investigated it. I played the demo some time back but never
actually tried to set it up to do anything real, so I can't vouch for that
kind of performance.
I totally forgot about it when I mentioned DosBox above.
Of course, where as DosBox emulates DOS as well, JPC just emulates the
hardware and you need an OS to go along with it. So you'll also have to
include FreeDos or OpenDos or ReactOS, in addition to your program.
>> I admit I was never that creative though. In fact, I was never really
>> that
>> fond of 0x88 even. When I used to do mailbox, I used to do classic 120
>> or
>> 16x16 based.
>
> Finalfun uses 0x88.
I used to prefer the 64 element map / unmap approach.
Occasionally I'd do a 16x16 for fun, but I was comfortable with the map /
unmap approach.
>
>> It'll involve piece lists at a minimum, of course. Then I've read people
>> talking about maintaining lists that can determine attacks. I wasn't
>> paying
>> much attention, but I think it's an aspect of 0x88 boards and some other
>> data structures & piece lists.
>
> I didn't find anything about attacks on the wiki.
It may not have it. I haven't checked.
The last I really saw anything about the attack lists etc. was on the old
CCC board.
Try the archives.
http://www.stmintz.com/ccc/
RGCC archives might have some, but I can't say how much.
>> Modern bitboard methods, like Gerd's Kindergarten method only take a few
>> K
>> for their data to work. (Exact amounts depend on how you do it.) Far
>> less
>> than other methods. And they give tolerable performance. (There are
>> faster
>> methods, but they use a lot more data. Personally I don't think it's a
>> good
>> trade-off.)
>>
>> Much less data than the old rotated method.
>
> But still way more data than without bitboards.
True.
An 0x88 board will be 128 bytes.
Piece lists will be another 32 bytes.
Attack lists will be at least another 32 bytes.
Maybe one or two more advanced data structures.
Figure maybe 256 bytes total.
Unless you are working on a 4 bit processor or something like an 8051 with
embedded memory, I don't see much difference between 512 and 4k. Not when
you then start talking about 100k programs, 256meg trans tables, History
Heuristic, Ref. tables, etc.
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| |
Date: 14 Sep 2008 19:59:45
From: Simon Krahnke
Subject: Re: Today again game improve Twice
|
* Sanny <[email protected] > (12:22) schrieb:
> Computer never does error. It will always play with same results. So
> once it has learnt what to do it will keep obeying the rules forever.
> While Humans get tired and make mistakes.
Unlike humans computers don't learn.
mfg, simon .... l
|
| |
Date: 14 Sep 2008 03:22:42
From: Sanny
Subject: Re: Today again game improve Twice
|
> =A0 I wonder, what is the "full potential" ofGetClub? =A0Odds are that wi=
th more men
> remaining on the board, both the human
> opponent and the program are more
> prone to errors.
Humans get tired and will be more prone to error. Computers have no
tiredness so in complex positions Computer will be able to get better
result.
Computer never does error. It will always play with same results. So
once it has learnt what to do it will keep obeying the rules forever.
While Humans get tired and make mistakes.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 13 Sep 2008 23:26:47
From: help bot
Subject: Re: Today again game improve Twice
|
On Sep 14, 1:20=A0am, Sanny <[email protected] > wrote:
> White -- Black
> (easy) -- (help bot)
>
> 1. e2-e4{0} c7-c6{6}
A cowardly move; this guy crawls on his
knees, begging for a draw.
> 2. d2-d4{0} d7-d5{2}
> 3. Nb1-c3{0} d5-e4{2}
> 4. Nc3-e4{0} Bc8-f5{2}
At one time, a favorite line of Gary
Kasparov-- a real fighting machine!
> 5. f2-f3{322}
"Only so!" Other moves might lead to
some hope of advantage for White.
> ... Bf5-e4{68}
> 6. f3-e4{24} e7-e5{28}
> 7. Ng1-f3{24} e5-d4{82}
> 8. Qd1-d4{0} Qd8-d4{14}
> 9. Nf3-d4{28} Nb8-d7{34}
> 10. Bc1-g5{134} Bf8-c5{54}
> 11. Nd4-f5{88} g7-g6{46}
> 12. Nf5-h6{20}
As you can see, White dominates the
all-important corners, leaving Black
nothing but scraps.
> ... Ng8-h6{80}
> 13. Bg5-h6{0} Ke8-c8{14}
> 14. Ke1-c1{22} Nd7-e5{36}
> 15. Rd1-d8{20} Rh8-d8{14}
> 16. h2-h3{28} f7-f6{80}
> 17. Bh6-d2{0} b7-b5{110}
> 18. g2-g4{20} Ne5-f3{52}
> 19. Bf1-d3{40} Nf3-e5{46}
> 20. Rh1-f1{20} Bc5-e7{66}
> 21. Bd2-c3{42} Ne5-d3{66}
> 22. c2-d3{66} Rd8-d3{20}
All Rook and pawn endings are drawn, thus,
here the two players decided to retire early, to
grab a brewsky... . No, wait.
> 23. Bc3-f6{308} Be7-f6{76}
> 24. Rf1-f6{0} Rd3-h3{12}
> 25. Rf6-c6{24} Kc8-b7{38}
> 26. Rc6-c5{42} Kb7-b6{28}
> 27. Rc5-g5{42} Kb6-c6{56}
> 28. e4-e5{28} a7-a6{48}
> 29. e5-e6{54} Kc6-d6{26}
> 30. b2-b3{88} Kd6-e6{22}
> 31. Rg5-c5{48} Ke6-d6{24}
> 32. b3-b4{72} Rh3-h4{42}
> 33. Rc5-g5{58} Rh4-h2{40}
> 34. Kc1-b1{22} Rh2-g2{24}
> 35. a2-a3{62} Rg2-d2{80}
> 36. Rg5-c5{20} Rd2-d4{42}
> 37. g4-g5{40} Rd4-d5{60}
Surely, you don't expect GetClub to fall
for /that/?
> 38. Rc5-d5{58}
Okay-- so I was mistaken. (There's one
born every minute.)
> Second Game by Taylor Kingston
> White -- Black
> (easy) -- (Stinky Garlnoot)
> 1. e2-e4{0} c7-c5{2}
> 2. c2-c3{0} d7-d5{8}
> 3. e4-d5{0} Qd8-d5{6}
> 4. d2-d4{0} Ng8-f6{6}
> 5. b2-b3{134}
A clear improvement on main line theory,
this new move tempts Black into trying to
exploit more weaknesses than he can
possibly tackle all at once.
> So you can see how both players quickly got into end game where
> GetClub is weak and win the game.
>
> Still this was Easy Level and yet competed well.
>
> If they had not exchanged major pieces in the opening then it would
> have been a difficult game.
>
> As GetClub's end game is not good.
Who let the cat out of the bag? This was
supposed to be my secret weapon.
> Do you see any mistake in the games of Help Bot & Taylor Kingston.
>
> Help Bot was thinking 45 sec / move while Taylor Kingston was thinking
> 74 sec / move
Hey, when the program goes into a long
think, I switch to another "tab" (such as
this one, where I read rgc postings). To
get back and forth requires time, since
the chess engine gobbles too many
resources on my machine. It often
happens that the display will not show
the chess board for several seconds
or more.
> While the Easy Level was thinking 55 sec / move.
>
> So all players were roughly taking same time.
>
> I was expacting GetClub will win But since =A0they exchanged the pieces
> GetClub could not play with full potential.
I wonder, what is the "full potential" of
GetClub? Odds are that with more men
remaining on the board, both the human
opponent and the program are more
prone to errors.
-- help bot
|
| |
Date: 13 Sep 2008 22:20:45
From: Sanny
Subject: Re: Today again game improve Twice
|
Today 2 games were played by Help Bot & Taylor Kingston against easy
level.
I find Taylor Kingston too 74 sec / move while Help Bot was taking 45
sec / move.
Both win the game against Easy Level. But they exchanged the Queen in
the opening itself to avoid Big fight.
Since the major pieces were gone early, the game came into END GAME.
And since GetClub END GAME is weak both win the game easily.
Both finished the middle game quickly.
I am posting both games below.
First Game by Help Bot.
---------------------------------------
He exhanged most of the pieces in the opening and win the End Game.
Game Played between help bot and easy at GetClub.com
--------------------------------------------------------------------------------
help bot: (Black)
easy: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM27067&game=Chess
--------------------------------------------------------------------------------
White -- Black
(easy) -- (help bot)
1. e2-e4{0} c7-c6{6}
2. d2-d4{0} d7-d5{2}
3. Nb1-c3{0} d5-e4{2}
4. Nc3-e4{0} Bc8-f5{2}
5. f2-f3{322} Bf5-e4{68}
6. f3-e4{24} e7-e5{28}
7. Ng1-f3{24} e5-d4{82}
8. Qd1-d4{0} Qd8-d4{14}
9. Nf3-d4{28} Nb8-d7{34}
10. Bc1-g5{134} Bf8-c5{54}
11. Nd4-f5{88} g7-g6{46}
12. Nf5-h6{20} Ng8-h6{80}
13. Bg5-h6{0} Ke8-c8{14}
14. Ke1-c1{22} Nd7-e5{36}
15. Rd1-d8{20} Rh8-d8{14}
16. h2-h3{28} f7-f6{80}
17. Bh6-d2{0} b7-b5{110}
18. g2-g4{20} Ne5-f3{52}
19. Bf1-d3{40} Nf3-e5{46}
20. Rh1-f1{20} Bc5-e7{66}
21. Bd2-c3{42} Ne5-d3{66}
22. c2-d3{66} Rd8-d3{20}
23. Bc3-f6{308} Be7-f6{76}
24. Rf1-f6{0} Rd3-h3{12}
25. Rf6-c6{24} Kc8-b7{38}
26. Rc6-c5{42} Kb7-b6{28}
27. Rc5-g5{42} Kb6-c6{56}
28. e4-e5{28} a7-a6{48}
29. e5-e6{54} Kc6-d6{26}
30. b2-b3{88} Kd6-e6{22}
31. Rg5-c5{48} Ke6-d6{24}
32. b3-b4{72} Rh3-h4{42}
33. Rc5-g5{58} Rh4-h2{40}
34. Kc1-b1{22} Rh2-g2{24}
35. a2-a3{62} Rg2-d2{80}
36. Rg5-c5{20} Rd2-d4{42}
37. g4-g5{40} Rd4-d5{60}
38. Rc5-d5{58} Kd6-d5{4}
39. Kb1-b2{106} Kd5-e5{38}
40. a3-a4{64} Ke5-f5{16}
41. a4-b5{68} a6-b5{6}
42. Kb2-b3{106} Kf5-g5{10}
43. Kb3-c3{40} Kg5-f5{26}
44. Kc3-d4{32} g6-g5{8}
45. Kd4-e3{116} h7-h5{14}
46. Ke3-f2{82} Kf5-f4{12}
47. Kf2-g1{24} Kf4-f3{22}
48. Kg1-f1{30} g5-g4{8}
49. Kf1-e1{38} g4-g3{16}
50. Ke1-d1{40} Kf3-e3{14}
51. Kd1-c2{74}
--------------------------------------------------------------------------------
help bot: (Black)
easy: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM27067&game=Chess
Second Game by Taylor Kingston
------------------------------------------
He too exchanged Queen in the opening Stage and win in the End Game.
Game Played between Stinky Garlnoot and easy at GetClub.com
--------------------------------------------------------------------------------
Stinky Garlnoot: (Black)
easy: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM27022&game=Chess
--------------------------------------------------------------------------------
White -- Black
(easy) -- (Stinky Garlnoot)
1. e2-e4{0} c7-c5{2}
2. c2-c3{0} d7-d5{8}
3. e4-d5{0} Qd8-d5{6}
4. d2-d4{0} Ng8-f6{6}
5. b2-b3{134} Nb8-c6{130}
6. d4-c5{20} Qd5-d1{120}
7. Ke1-d1{0} e7-e5{4}
8. b3-b4{26} a7-a5{36}
9. Bf1-b5{78} Bc8-d7{102}
10. Bb5-c6{28} Bd7-c6{36}
11. f2-f3{64} b7-b6{122}
12. c5-b6{22} a5-b4{18}
13. Ng1-h3{182} Bf8-c5{398}
14. b6-b7{0} Bc6-b7{22}
15. Bc1-g5{100} Nf6-d5{46}
16. Rh1-e1{28} f7-f6{32}
17. Bg5-d2{38} b4-b3{104}
18. a2-a3{48} Ra8-d8{68}
19. a3-a4{22} Nd5-b6{58}
20. Nh3-f4{32} Ke8-f7{50}
21. a4-a5{90} Nb6-c4{42}
22. Nf4-h5{126} b3-b2{86}
23. Ra1-a2{32} Nc4-d2{42}
24. Kd1-c2{38} Nd2-c4{24}
25. a5-a6{24} Bb7-d5{80}
26. Ra2-a4{26} Rd8-b8{132}
27. a6-a7{22} Nc4-e3{46}
28. Re1-e3{30} Bd5-b3{22}
29. Kc2-d3{26} Rb8-d8{34}
30. Ra4-d4{56} e5-d4{32}
31. Re3-e1{48} Bc5-a7{20}
32. Nh5-f4{138} Rd8-c8{64}
33. c3-d4{20} Rc8-c1{26}
34. Kd3-d2{24} Rh8-c8{64}
35. Nf4-d3{26} Rc1-c2{116}
36. Kd2-e3{0} Rc8-e8{20}
37. Ke3-f4{24} Re8-e1{48}
38. Nd3-e1{46} Rc2-e2{6}
39. Ne1-d3{32} Ba7-b8{56}
40. Nd3-e5{0} f6-e5{16}
41. d4-e5{28} Bb8-e5{42}
42. Kf4-g5{84} Bb3-e6{30}
43. f3-f4{28}
--------------------------------------------------------------------------------
Stinky Garlnoot: (Black)
easy: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM27022&game=Chess
--------------------------
So you can see how both players quickly got into end game where
GetClub is weak and win the game.
Still this was Easy Level and yet competed well.
If they had not exchanged major pieces in the opening then it would
have been a difficult game.
As GetClub's end game is not good.
Do you see any mistake in the games of Help Bot & Taylor Kingston.
Help Bot was thinking 45 sec / move while Taylor Kingston was thinking
74 sec / move
While the Easy Level was thinking 55 sec / move.
So all players were roughly taking same time.
I was expacting GetClub will win But since they exchanged the pieces
GetClub could not play with full potential.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 13 Sep 2008 20:54:46
From: Sanny
Subject: Re: Another game to analyze.
|
> Board= 128 bytes
> u64 Pieces[13]
> u64 Colors[3]
>
> For bitboard attack methods:
> Attacks=3584
> u64 R0_Attacks[64]
> u64 R90_Attacks[64]
> u64 R45_Attacks[64]
> u64 R135_Attacks[64]
> u64 Knight_Attacks[64]
> u64 King_attacks[64]
> u8 FirstRankAttacks[8][64]
>
> Under 4k of bitboard specific data.
Can you provide some link where I can get more information about using
above techniques.
If I generate moves using above how fast the move generation would be?
Bye
Sanny
|
| | |
Date: 14 Sep 2008 10:02:22
From: Guest
Subject: Re: Another game to analyze.
|
"Sanny" <[email protected] > wrote in message
news:[email protected]...
>>
>> Under 4k of bitboard specific data.
>
> Can you provide some link where I can get more information about using
> above techniques.
The Chess Programming Wiki is a good place to check....
http://chessprogramming.wikispaces.com/
Search for "Kindergarten" You'll probably want the 32 bit friendly version.
You probably should search for "BitScan" too. These portable versions
aren't fast, but if you use them sparingly, it wont be too bad. (Although
you might have some extra issues in Java.)
If you've got any questions about these techiniques, I'd suggest asking in
this forum:
http://www.talkchess.com/forum/viewforum.php?start=0&f=7&topic_view=flat
That way the people who developed these newer bitboard methods can respond
directly.
(You may be tempted to ask in the wiki discussion area. Don't. It's too
fragmented. More people will see your message and be able to help if you
post in the forum.)
> If I generate moves using above how fast the move generation would be?
Slower. For bitboards, move generation is slower than the advanced 0x88
methods. At least it's harder to get fast.
Fortunately, move generation is but a small percentage of a chess program's
total runtime.
Where you win is usually the evaluator and doing attack detection.
BitBoards are not magic. They aren't going to autoatmically make your
prorgram faster or better.
They also have a higher learning curve. Meaning it takes longer to get
comfortable with them and put them to good use.
Sanny, I'm going to tell you right here and now.... It's not going to help
your Java program. Your program already has too many other issues.
You can do a fast, strong chess program using the mailbox style. If you
don't want to get fancy and use advanced data structures, even a simple
10x12 board with piecelists etc. will do quite well. It wont be as fast,
but you can still do quite well.
You can not easily take a mailbox program and convert it into bitboard. It
will require a complete rewrite. It requires a completely new way of
thinking, too.
Which you probably need to do anyway, since GC has so many tactical problems
there's got to be something wrong in it. (I tell most people new to chess
programming that they'll throw away their first couple of chess programs.
It takes at least that long before they start getting comfortable with chess
programming and the techniques it requires.)
You could do a hybrid aproach, just to get the feel of things. Do both
mailbox & bitboard. That way you can use whatever data structure you want
in a particular place while you are still learning and are uncomfortable
with bitboards. It can take quite a while before you get comfortable with
bitboards and start thinking in those terms.
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| | | |
Date: 14 Sep 2008 13:54:56
From: Guest
Subject: Re: Another game to analyze.
|
Sanny
I should mention that if you want to keep using mailbox and do piece lists,
attack tables, etc. you can find information in the ChessProgrammingWiki,
along with the new Computer Chess Club forum, and you can check the CCC
archives.
http://chessprogramming.wikispaces.com/
http://www.talkchess.com/forum/viewforum.php?start=0&f=7&topic_view=flat
http://www.stmintz.com/ccc/index.php
With the attack tables and some other stuff I can't remember, you can make
move generation and attack detection very quick indeed. Faster than what
most people can do bitboards. And the code would be java friendly, where as
bitboards aren't.
But, a good chess program is likely to spend under 10% of its time doing
move generation. (The exact amount will vary, of course, depending on the
program itself.)
Fast attack detection is certainly handy in the eval, though.
To be entirely honest, I'd suggest you start over. Use a mailbox (0x88 is
popular, but others exist), do the piece lists, attack tables, etc.
Streamline the search & eval until you can do real well on all the tactical
tests.
Then start adding enhancements to the search & eval.
"Guest" <[email protected] > wrote in message news:[email protected]...
> "Sanny" <[email protected]> wrote in message
> news:[email protected]...
>>>
>>> Under 4k of bitboard specific data.
>>
>> Can you provide some link where I can get more information about using
>> above techniques.
>
> The Chess Programming Wiki is a good place to check....
>
> http://chessprogramming.wikispaces.com/
>
> Search for "Kindergarten" You'll probably want the 32 bit friendly
> version.
>
> You probably should search for "BitScan" too. These portable versions
> aren't fast, but if you use them sparingly, it wont be too bad. (Although
> you might have some extra issues in Java.)
>
>
> If you've got any questions about these techiniques, I'd suggest asking in
> this forum:
>
> http://www.talkchess.com/forum/viewforum.php?start=0&f=7&topic_view=flat
>
> That way the people who developed these newer bitboard methods can respond
> directly.
>
> (You may be tempted to ask in the wiki discussion area. Don't. It's too
> fragmented. More people will see your message and be able to help if you
> post in the forum.)
>
>
>> If I generate moves using above how fast the move generation would be?
>
> Slower. For bitboards, move generation is slower than the advanced 0x88
> methods. At least it's harder to get fast.
>
> Fortunately, move generation is but a small percentage of a chess
> program's total runtime.
>
> Where you win is usually the evaluator and doing attack detection.
>
> BitBoards are not magic. They aren't going to autoatmically make your
> prorgram faster or better.
>
> They also have a higher learning curve. Meaning it takes longer to get
> comfortable with them and put them to good use.
>
>
> Sanny, I'm going to tell you right here and now.... It's not going to help
> your Java program. Your program already has too many other issues.
>
> You can do a fast, strong chess program using the mailbox style. If you
> don't want to get fancy and use advanced data structures, even a simple
> 10x12 board with piecelists etc. will do quite well. It wont be as fast,
> but you can still do quite well.
>
> You can not easily take a mailbox program and convert it into bitboard.
> It will require a complete rewrite. It requires a completely new way of
> thinking, too.
>
> Which you probably need to do anyway, since GC has so many tactical
> problems there's got to be something wrong in it. (I tell most people new
> to chess programming that they'll throw away their first couple of chess
> programs. It takes at least that long before they start getting
> comfortable with chess programming and the techniques it requires.)
>
> You could do a hybrid aproach, just to get the feel of things. Do both
> mailbox & bitboard. That way you can use whatever data structure you want
> in a particular place while you are still learning and are uncomfortable
> with bitboards. It can take quite a while before you get comfortable with
> bitboards and start thinking in those terms.
>
>
>
>
>
>
> ----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet
> News==----
> http://www.pronews.com The #1 Newsgroup Service in the World! >100,000
> Newsgroups
> ---= - Total Privacy via Encryption =---
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| |
Date: 14 Sep 2008 02:43:49
From: Simon Krahnke
Subject: Re: Another game to analyze.
|
* Guest <[email protected] > (22:12) schrieb:
> "Simon Krahnke" <[email protected]> wrote in message
> news:[email protected]...
>>* Guest <[email protected]> (16:05) schrieb:
>>
>>>> Can a Program in C with modifications run on other platforms?
>>>
>>> To a large degree, yes.
>>
>> That means you have to recompile or relink the executable.
>
> You provide three executables, and that takes care of 99% of all platforms.
> Very few modifications are needed to provide that level of portability.
I was just saying you can't use the same executable file on every
platform. Java cross platform is about executables, C cross platform is
about source files.
> Heck, if you want to cheat a little, provide a DOS based executable and just
> distribute that. People can use something like DosBox to run it.
If you replace DOS by Win32 and Dosbox by Wine that's the trick Rybka uses.
> Oh, one other thing.... It's also possible to take a C program and compile
> that for Java bytecode. It involves compiling for MIPS and then converting
> that. Cumbersome, but supposedly works.
:-)
> There are also other compilers that can do that to Java bytecode too.
>
> Just pointing out that you don't have to actually use Java to run it on the
> JVM in a browser.
There's a Ruby implementation for the JVM.
>>> The few exceptions are:
>>>
>>> 1) detect whether there is any keyboard input pending. Used for
>>> pondering.
>>
>> And analyzing.
>
> Analyzing is similar to pondering as far as programming. It's just an
> 'infinite' time limit.
>
> However, many programs make do without analyzing. If you want a deep
> anaylsis you just set a long time and then interrupt it.
But that is basically the same problem, stopping the search. Of course
there is always the way of killing the process. But that again is not
platform independent.
>>> 2) threads / processes. Used for multi-cpu programming.
>>
>> Can also be used to seperate communication from the search.
>
> It can, but doesn't have to.
It's can be easier that way. You got clean seperation between the two issues.
> Also, realistically, going dual core is only going to improve the program's
> performance by a small amount. Less than 1 ply of search. Going quad-core
> may give you a little more than one ply.
But then again it's silly two use only one processor when you got two of
them.
> Point is, threads & processes are a luxury, not a necessity like checking
> for available keyboard input without actually reading it.
But my point above was, that this solves the "keyboard" issue. Your
protocol thread just does a synchronous read on the input and is then
sleeping as long as there is no input.
>>> All the rest can be done in plain generic ISO C.
>>
>> Except some compiler stuff like intrinsics or inline assembler.
>
> Which isn't needed for most stuff.
>
> For mailbox boards (or 0x88 style), it's not needed at all. And you can do
> a very strong, very fast program with mailbox style.
Let's hope so. I'm just taking a try on this.
> For bitboards, there's been a lot of advancements in bitboard 'magic' stuff
> and it's not as big of an issue as it used to be. The only one area where
> it's still very helpful is in bits canning. BSF or BSR. Not a massive
> performance improvement or C code though. Noticable, but not massive.
Sure, but if you want to use them, the compiler matters.
> As for intrinsics & inline... Well, considering the Mac has moved to x86
> processors, you have no significant need to support any processor other than
> that. And considering GNU C is available on all three major platforms, a
> couple of inline asm lines can be added that will work on 99% of all
> systems.
But most windows programmers don't seem to use GCC.
> But you certainly don't have to do bitboards... Mailbox style is still
> acceptable and can be strong and fast. It's more a matter of personal
> choice, not inherent abilities.
But bitboards can be so nice in just about everything a chess engine
does. Only they always seem to use vast amounts of memory.
>>> And "other platforms" only involve three: Windows, Linux and Mac. Not
>>> exactly a wide variety like it used to be.
>>
>> And that can even mean Windows and Unix.
>
> Not sure what you are meaning here, since I said Windows.
"Unix"
In think all the above issues can be solved the same way on Linux,
MacOS, FreeBSD, Solaris ...
>>> With upcomming Javascript improvements, it might be possible to write a
>>> reasonable javascript chess program. Not one of the toy chess programs
>>> that
>>> have currently been done.
>>
>> It would be cool, if Javascript would be faster than Java. Javascript
>> has been fixed to the browser too long now.
>
> It might happen.... That's what people are talking about.
>
> If it goes from an interpreted script language to a JiT compiled language,
> that would give it that kind of performance.
That's what Ruby is currently going through, too. With version 1.9 it's
taking the step from a plain interpreted language to a virtual machine.
mfg, simon .... l
|
| | |
Date: 13 Sep 2008 20:57:17
From: Guest
Subject: Re: Another game to analyze.
|
"Simon Krahnke" <[email protected] > wrote in message
news:[email protected]...
>* Guest <[email protected]> (22:12) schrieb:
>
>> "Simon Krahnke" <[email protected]> wrote in message
>> news:[email protected]...
>>>* Guest <[email protected]> (16:05) schrieb:
>> Heck, if you want to cheat a little, provide a DOS based executable and
>> just
>> distribute that. People can use something like DosBox to run it.
>
> If you replace DOS by Win32 and Dosbox by Wine that's the trick Rybka
> uses.
I thought of that, but I wasn't sure about Wine on the Mac.
Also DOS & DOSBox have the advantage that they can work on non-x86
platforms, such as the old Macs. (At a factor of 10 performance or so, but
it'll work.)
>>>> The few exceptions are:
>>>>
>>>> 1) detect whether there is any keyboard input pending. Used for
>>>> pondering.
>>>
>>> And analyzing.
>>
>> Analyzing is similar to pondering as far as programming. It's just an
>> 'infinite' time limit.
>>
>> However, many programs make do without analyzing. If you want a deep
>> anaylsis you just set a long time and then interrupt it.
>
> But that is basically the same problem, stopping the search. Of course
> there is always the way of killing the process. But that again is not
> platform independent.
Right.
But all that's required is some way to check for the existence of keyboard
input without trying to read it (which pauses the program.)
Whether it's pondering or anaylsis, it's the same issue as far as the
programmer is concerned.
It's the same thing you need to do for Winboard, for example.
It's a common problem and there are already common solutions that can be cut
& pasted.
>>>> 2) threads / processes. Used for multi-cpu programming.
>>>
>>> Can also be used to seperate communication from the search.
>>
>> It can, but doesn't have to.
>
> It's can be easier that way. You got clean seperation between the two
> issues.
Dealing with processes & threads can be a bit more complicated among
platforms.
Windows, Linux & Mac do things differently enough that it can cause issues.
There are also performance issues depending on how you do it.
Or so I've repeatedly read. I've seen discussions in the TalkChess forums
about it.
>
>> Also, realistically, going dual core is only going to improve the
>> program's
>> performance by a small amount. Less than 1 ply of search. Going
>> quad-core
>> may give you a little more than one ply.
>
> But then again it's silly two use only one processor when you got two of
> them.
It's easier though.... And for what little extra benefit you'll get, it may
not be worth the trouble. Especially in the beginning.
I've also repeated read that it's much easier to use two processors than
four. About any method will do okay with just two, but as soon as it's run
on a quad, you might actually get a slow-down.
So it might be a good idea to do development on a quad isntead of the more
common & cheaper duals.
>
>> Point is, threads & processes are a luxury, not a necessity like checking
>> for available keyboard input without actually reading it.
>
> But my point above was, that this solves the "keyboard" issue. Your
> protocol thread just does a synchronous read on the input and is then
> sleeping as long as there is no input.
It can solve it, yes. But it requires more effort and complications to do
it.
Dealing with threads or processes portably can be much more trouble than
having one single routine written portably called CheckForInputAvailable().
>> For mailbox boards (or 0x88 style), it's not needed at all. And you can
>> do
>> a very strong, very fast program with mailbox style.
>
> Let's hope so. I'm just taking a try on this.
Oh you can. But it will require some creativity.
I admit I was never that creative though. In fact, I was never really that
fond of 0x88 even. When I used to do mailbox, I used to do classic 120 or
16x16 based.
It'll involve piece lists at a minimum, of course. Then I've read people
talking about maintaining lists that can determine attacks. I wasn't paying
much attention, but I think it's an aspect of 0x88 boards and some other
data structures & piece lists.
I really don't know the details that much. As I said, I wasn't paying much
attention because I didn't care that much. I probably should have paid
attention, but at the time I just didn't care.
You could check the ChessProgrammingWiki and ask in the TalkChess forums.
>
>> For bitboards, there's been a lot of advancements in bitboard 'magic'
>> stuff
>> and it's not as big of an issue as it used to be. The only one area
>> where
>> it's still very helpful is in bits canning. BSF or BSR. Not a massive
>> performance improvement or C code though. Noticable, but not massive.
>
> Sure, but if you want to use them, the compiler matters.
I was just meaning portability, though. GCC is so portable and so available
that it's feasible to use it for whatever platform you want. Which will
likely be x86 based.
One codebase takes care of it. You wont need to write new BSF or BSR inline
routines for every one.
>
>> As for intrinsics & inline... Well, considering the Mac has moved to x86
>> processors, you have no significant need to support any processor other
>> than
>> that. And considering GNU C is available on all three major platforms, a
>> couple of inline asm lines can be added that will work on 99% of all
>> systems.
>
> But most windows programmers don't seem to use GCC.
Most windows programmers don't have to. Just one programmer... whoever's
program it is or if they know somebody who can compile it for them.
Besides, it's a good idea to run your program through at least two different
compilers, just to help detect bugs. I used to do Microsoft, GNU C and
OpenWatcom.
>
>> But you certainly don't have to do bitboards... Mailbox style is still
>> acceptable and can be strong and fast. It's more a matter of personal
>> choice, not inherent abilities.
>
> But bitboards can be so nice in just about everything a chess engine
> does. Only they always seem to use vast amounts of memory.
I prefer bitboards too. (That's one of the reasons I wasn't paying
attention to the 0x88 stuff I mentioned above.)
Modern bitboard methods, like Gerd's Kindergarten method only take a few K
for their data to work. (Exact amounts depend on how you do it.) Far less
than other methods. And they give tolerable performance. (There are faster
methods, but they use a lot more data. Personally I don't think it's a good
trade-off.)
Much less data than the old rotated method.
Other than that, you've got the board itself and a few other general data
arrays and whatever you want to do for eval.
Let's see....
Board= 128 bytes
u64 Pieces[13]
u64 Colors[3]
For bitboard attack methods:
Attacks=3584
u64 R0_Attacks[64]
u64 R90_Attacks[64]
u64 R45_Attacks[64]
u64 R135_Attacks[64]
u64 Knight_Attacks[64]
u64 King_attacks[64]
u8 FirstRankAttacks[8][64]
Under 4k of bitboard specific data.
And that's pretty much it for a basic bitboard program. It can go up from
there, of course.
Actually, I think you might be able to even lower it a little. I've lost
track of all the new things they've managed to do the past year or two.
There for a while, they were developing a new method every few days.
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| |
Date: 13 Sep 2008 14:16:24
From: help bot
Subject: Re: Today again game improve Twice
|
On Sep 13, 6:41=A0am, Sanny <[email protected] > wrote:
> Next time play with Easy Level and it will make moves of Normal Level.
> So saving your time but not the efforts to win them.
>
> Now, Taylor Kingston will find it much difficult to win the Beginner &
> Easy Level.
You obviously have no idea whom you're
dealing with. Mr. Kingston likely has two
machines set up, side by side, along with
a stack of opening books a quarter-mile
high. On top of this, he will "think" -- if
that is not too much of a stretch -- for five
minutes on a move, even against the
lower levels. Your program has no
chance against this titanic effort, this
operation Barbarossa.
-- help bot
|
| |
Date: 13 Sep 2008 21:20:37
From: Simon Krahnke
Subject: Re: Another game to analyze.
|
* Guest <[email protected] > (16:05) schrieb:
>> Can a Program in C with modifications run on other platforms?
>
> To a large degree, yes.
That means you have to recompile or relink the executable.
> The few exceptions are:
>
> 1) detect whether there is any keyboard input pending. Used for pondering.
And analyzing.
> 2) threads / processes. Used for multi-cpu programming.
Can also be used to seperate communication from the search.
> All the rest can be done in plain generic ISO C.
Except some compiler stuff like intrinsics or inline assembler.
> And "other platforms" only involve three: Windows, Linux and Mac. Not
> exactly a wide variety like it used to be.
And that can even mean Windows and Unix.
> With upcomming Javascript improvements, it might be possible to write a
> reasonable javascript chess program. Not one of the toy chess programs that
> have currently been done.
It would be cool, if Javascript would be faster than Java. Javascript
has been fixed to the browser too long now.
mfg, simon .... l
|
| | |
Date: 13 Sep 2008 15:12:43
From: Guest
Subject: Re: Another game to analyze.
|
"Simon Krahnke" <[email protected] > wrote in message
news:[email protected]...
>* Guest <[email protected]> (16:05) schrieb:
>
>>> Can a Program in C with modifications run on other platforms?
>>
>> To a large degree, yes.
>
> That means you have to recompile or relink the executable.
So?
He didn't say within a browser. He just asked if it can run on other
platforms.
The answer is definetly yes.
You provide three executables, and that takes care of 99% of all platforms.
Very few modifications are needed to provide that level of portability.
Heck, if you want to cheat a little, provide a DOS based executable and just
distribute that. People can use something like DosBox to run it.
Oh, one other thing.... It's also possible to take a C program and compile
that for Java bytecode. It involves compiling for MIPS and then converting
that. Cumbersome, but supposedly works.
There are also other compilers that can do that to Java bytecode too.
Just pointing out that you don't have to actually use Java to run it on the
JVM in a browser.
>
>> The few exceptions are:
>>
>> 1) detect whether there is any keyboard input pending. Used for
>> pondering.
>
> And analyzing.
Analyzing is similar to pondering as far as programming. It's just an
'infinite' time limit.
However, many programs make do without analyzing. If you want a deep
anaylsis you just set a long time and then interrupt it.
This is really the only needed OS specific thing. And that'll only be a few
lines of code that gets written once and never looked at again. You
wouldn't even need to write it, since most chess programmers use the same
code as everybody else. It's some old PD code that gets shared around quite
a bit.
>> 2) threads / processes. Used for multi-cpu programming.
>
> Can also be used to seperate communication from the search.
It can, but doesn't have to.
Also, realistically, going dual core is only going to improve the program's
performance by a small amount. Less than 1 ply of search. Going quad-core
may give you a little more than one ply.
Not a significant amount for a typical program and strength of average
player.
Point is, threads & processes are a luxury, not a necessity like checking
for available keyboard input without actually reading it.
>> All the rest can be done in plain generic ISO C.
>
> Except some compiler stuff like intrinsics or inline assembler.
Which isn't needed for most stuff.
For mailbox boards (or 0x88 style), it's not needed at all. And you can do
a very strong, very fast program with mailbox style.
For bitboards, there's been a lot of advancements in bitboard 'magic' stuff
and it's not as big of an issue as it used to be. The only one area where
it's still very helpful is in bits canning. BSF or BSR. Not a massive
performance improvement or C code though. Noticable, but not massive.
As for intrinsics & inline... Well, considering the Mac has moved to x86
processors, you have no significant need to support any processor other than
that. And considering GNU C is available on all three major platforms, a
couple of inline asm lines can be added that will work on 99% of all
systems.
But you certainly don't have to do bitboards... Mailbox style is still
acceptable and can be strong and fast. It's more a matter of personal
choice, not inherent abilities.
>
>> And "other platforms" only involve three: Windows, Linux and Mac. Not
>> exactly a wide variety like it used to be.
>
> And that can even mean Windows and Unix.
Not sure what you are meaning here, since I said Windows.
>
>> With upcomming Javascript improvements, it might be possible to write a
>> reasonable javascript chess program. Not one of the toy chess programs
>> that
>> have currently been done.
>
> It would be cool, if Javascript would be faster than Java. Javascript
> has been fixed to the browser too long now.
It might happen.... That's what people are talking about.
If it goes from an interpreted script language to a JiT compiled language,
that would give it that kind of performance.
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| |
Date: 13 Sep 2008 15:34:18
From: Simon Krahnke
Subject: Re: Another game to analyze.
|
* help bot <[email protected] > (09:32) schrieb:
> I'm thinking that if I had a chunk of code
> to efficiently generate a list of all the legal
> moves from any given position, I could
> quite easily beat the GetClub program by
> just bean-counting; that is, by simply
> tallying up the material score and doing a
> brute force search, then breaking any tie
> at the instant I need to move by ordering
> equal moves in a very simple manner. I
> would expect to win on tactics alone, by
> just getting the math right (unlike the GC
> engine).
And I doubt that. You nearly always have to break a tie, and that
doesn't help the brute-force search, so it won't get very deep.
> I cannot stomach trying to write code to
> deal with en passant captures, castling,
> or pawn promotions--
Actually castling is the most complicated of these cases. The other two
can be implemented quite easy.
> that's menial labor,
> best suited for economists, politicians
> and lawyers-- the lowest caste. I would
> want to focus on ideas, on strategy,
You mean like an economist or politician?
mfg, simon .... l
|
| |
Date: 13 Sep 2008 03:41:27
From: Sanny
Subject: Today again game improve Twice
|
Today again the game was improved.
So in comming games you have to fight heavily to win.
Play Chess at: http://www.GetClub.com/Chess.html
May be this game you win easily. But new games you play you have to
think a lot.
Next time play with Easy Level and it will make moves of Normal Level.
So saving your time but not the efforts to win them.
Now, Taylor Kingston will find it much difficult to win the Beginner &
Easy Level.
As today again the game was improved Twice.
The Glory of GetClub Chess has returned.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 13 Sep 2008 00:32:02
From: help bot
Subject: Re: Another game to analyze.
|
On Sep 13, 2:10=A0am, Sanny <[email protected] > wrote:
> > =A0 The Java platform seems to have slowed
> > Sanny's program down quite a bit, and his
> > "Bugster" programming language may not
> > be ideal for chess.
>
> True, If the same program is written in C then the program will be
> 5-10 times faster. But "C" program do not work on all computers. Like
> those having Linux/ Mac and other.
>
> Can a Program in C with modifications run on other platforms?
I have no idea. The comment above was
a joke-- there is no "Bugster" programming
language.
Tonight I beat up on the GetClub program
a couple of times, but in my current game
I blundered, losing a piece for nothing and
the struggle continues as I occupy my
time with other things. This could very
well end up being another hundred-mover,
since I traded into a drawable Rook ending
fully expecting to win via sheer force of will
(and of course, my superior intellect).
I'm thinking that if I had a chunk of code
to efficiently generate a list of all the legal
moves from any given position, I could
quite easily beat the GetClub program by
just bean-counting; that is, by simply
tallying up the material score and doing a
brute force search, then breaking any tie
at the instant I need to move by ordering
equal moves in a very simple manner. I
would expect to win on tactics alone, by
just getting the math right (unlike the GC
engine).
I cannot stomach trying to write code to
deal with en passant captures, castling,
or pawn promotions-- that's menial labor,
best suited for economists, politicians
and lawyers-- the lowest caste. I would
want to focus on ideas, on strategy, to
try to improve my bean-counter program
beyond simple tactics, but not until it
was proved to work properly with regard
to tactics.
-- help bot
|
| |
Date: 12 Sep 2008 23:10:12
From: Sanny
Subject: Re: Another game to analyze.
|
> =A0 The Java platform seems to have slowed
> Sanny's program down quite a bit, and his
> "Bugster" programming language may not
> be ideal for chess.
True, If the same program is written in C then the program will be
5-10 times faster. But "C" program do not work on all computers. Like
those having Linux/ Mac and other.
Can a Program in C with modifications run on other platforms?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| | |
Date: 13 Sep 2008 09:05:26
From: Guest
Subject: Re: Another game to analyze.
|
>
>"Sanny" <[email protected]> wrote in message
>news:cdee8970-4fd8-42c7-bd7d->[email protected]...
>> The Java platform seems to have slowed
>> Sanny's program down quite a bit, and his
>> "Bugster" programming language may not
>> be ideal for chess.
>
>True, If the same program is written in C then the program will be
>5-10 times faster. But "C" program do not work on all computers. Like
>those having Linux/ Mac and other.
Supposedly it can be lower than that, if you use the right data structures.
The "Just in Time" java compilers that browsers use do impose some overhead,
but the final code ends up being not too far off from regular compiled code.
Maybe x2. Quite a bit faster than your x5-x10 estimate. (If you wanted to
know for sure, just compile your program with a real native code compiler
instead of bytecode. By doing that, whatever is there that makes it slow is
mostly your own programming, not Java.)
Java will always have some performance penalty, but supposedly it can be
managed. After all, there are already a lot of other Java programs that run
fast.
You might want to try them and compare their performance to yours...
>Can a Program in C with modifications run on other platforms?
To a large degree, yes.
The few exceptions are:
1) detect whether there is any keyboard input pending. Used for pondering.
2) threads / processes. Used for multi-cpu programming.
All the rest can be done in plain generic ISO C.
And "other platforms" only involve three: Windows, Linux and Mac. Not
exactly a wide variety like it used to be.
With upcomming Javascript improvements, it might be possible to write a
reasonable javascript chess program. Not one of the toy chess programs that
have currently been done.
>
>Bye
>Sanny
>
>Play Chess at: http://www.GetClub.com/Chess.html
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| | | |
Date: 13 Sep 2008 22:34:54
From: David Richerby
Subject: Re: Another game to analyze.
|
[Cross-post trimmed.]
Guest <[email protected] > wrote:
> And "other platforms" only involve three: Windows, Linux and Mac.
> Not exactly a wide variety like it used to be.
Um. Solaris, HP-UX, AIX, the BSDs, Playstation, the various mobile phone
operating systems, ...
Dave.
--
David Richerby Flammable Dictator (TM): it's like
www.chiark.greenend.org.uk/~davidr/ a totalitarian leader but it burns
really easily!
|
| | | | |
Date: 13 Sep 2008 17:45:04
From: Guest
Subject: Re: Another game to analyze.
|
"David Richerby" <[email protected] > wrote in message
news:A+t*[email protected]...
> [Cross-post trimmed.]
>
> Guest <[email protected]> wrote:
>> And "other platforms" only involve three: Windows, Linux and Mac.
>> Not exactly a wide variety like it used to be.
>
> Um. Solaris, HP-UX, AIX, the BSDs, Playstation, the various mobile phone
> operating systems, ...
Yes, they exist.
How many of them are likely to try to run GetClub chess?
Under 1%.
You might as well be talking about trying to run it on an 8 bit micro.
That's why I put "other platforms" in quotes. It wasn't meant to be all
inclusive, but only covering the 99% of the platforms.
>
>
> Dave.
>
> --
> David Richerby Flammable Dictator (TM): it's
> like
> www.chiark.greenend.org.uk/~davidr/ a totalitarian leader but it
> burns
> really easily!
>
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| | | | | |
Date: 14 Sep 2008 00:38:06
From: David Richerby
Subject: Re: Another game to analyze.
|
Guest <[email protected] > wrote:
> "David Richerby" <[email protected]> wrote:
>> Guest <[email protected]> wrote:
>>> And "other platforms" only involve three: Windows, Linux and Mac.
>>> Not exactly a wide variety like it used to be.
>>
>> Um. Solaris, HP-UX, AIX, the BSDs, Playstation, the various mobile
>> phone operating systems, ...
>
> Yes, they exist.
>
> How many of them are likely to try to run GetClub chess?
>
> Under 1%.
Actually, if you look at the table at
http://en.wikipedia.org/wiki/Usage_share_of_desktop_operating_systems
which summarizes various surveys for desktop operating systems,
there's an implicit "other" column missing (i.e., the difference
between the totals for the various versions of Windows, MacOS and
Linux, and 100%).
Source Linux Other
===============================
Net Applications 0.93% 1.56%
W3 Counter 2.00% 4.43%
XiTi Monitor 0.98% 1.77%
Onestat 0.42% 0.65%
So, in all cases, "other" is more significant in total than Linux.
Also, that's desktop operating systems, so Playstation and mobile
phones aren't included. Those are big.
> You might as well be talking about trying to run it on an 8 bit micro.
Nonsense.
Dave.
--
David Richerby Addictive Sadistic Bulb (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a light bulb but it wants to hurt
you and you can never put it down!
|
| | | | | | |
Date: 13 Sep 2008 19:51:36
From: Guest
Subject: Re: Another game to analyze.
|
"David Richerby" <[email protected] > wrote in message
news:2Gr*[email protected]...
> Guest <[email protected]> wrote:
>> "David Richerby" <[email protected]> wrote:
>>> Guest <[email protected]> wrote:
>>>> And "other platforms" only involve three: Windows, Linux and Mac.
>>>> Not exactly a wide variety like it used to be.
>>>
>>> Um. Solaris, HP-UX, AIX, the BSDs, Playstation, the various mobile
>>> phone operating systems, ...
>>
>> Yes, they exist.
>>
>> How many of them are likely to try to run GetClub chess?
>>
>> Under 1%.
>
> Actually, if you look at the table at
>
> http://en.wikipedia.org/wiki/Usage_share_of_desktop_operating_systems
>
> which summarizes various surveys for desktop operating systems,
> there's an implicit "other" column missing (i.e., the difference
> between the totals for the various versions of Windows, MacOS and
> Linux, and 100%).
>
> Source Linux Other
> ===============================
> Net Applications 0.93% 1.56%
> W3 Counter 2.00% 4.43%
> XiTi Monitor 0.98% 1.77%
> Onestat 0.42% 0.65%
>
> So, in all cases, "other" is more significant in total than Linux.
> Also, that's desktop operating systems, so Playstation and mobile
> phones aren't included. Those are big.
1) True, in that particular list, 'Other' is bigger than I thought. but it
still amounts to under 2%, where as I said 1 percent.
Big fat hairy deal.
If you look here,
http://marketshare.hitslink.com/report.aspx?qprid=8
Windows, Mac & Linux are 99.45% of he total. Everything else is under 1%.
includes playstation... 0.04% Massive. The iPhone... 0.30%. Huge!
Realistically, you can deal with Windows & Mac and get 98% of the market
that is likely to run GC. Throw in Linux and you get another 1%.
The rest makes as much sense supporting as if you tried to support all those
8 bit micros that some people still do as a hobby.
If you look here
http://marketshare.hitslink.com/report.aspx?qprid=10
you can see that for the MacOS, 2.42% are still PowerPC MacOS based. So
that does bring things down to 96.74% for Windows & x86 based Mac's.
So because of the PPC Mac, my estimate was off.
But mentioning the iPhone & other phones as likely platforms for GC.... Not
hardly.
Keep your target audience in mind....
>
>
>> You might as well be talking about trying to run it on an 8 bit micro.
>
> Nonsense.
>
>
> Dave.
>
> --
> David Richerby Addictive Sadistic Bulb (TM): it's
> www.chiark.greenend.org.uk/~davidr/ like a light bulb but it wants to
> hurt
> you and you can never put it down!
>
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| | |
Date: 13 Sep 2008 13:45:06
From: jefk
Subject: Re: Another game to analyze.
|
Sanny wrote:
> Can a Program in C with modifications run on other platforms?
>
>
depends how you compile it; Linux is fully compatible with C,
but for windows you need a windoze compiler;
hey, that makes sense, doesnt it ?
:)
jef
PS Rybka 3 confirms:
chess is still a draw; least when black is choosing proper defense lines;
like the RL Zaitsev or (old) Queens Indian (eg with Bb7! after g3)
in general i suppose this is a logical outcome for a game
with so much degrees of freedom for the other (black side);
in 4-in-a-row which has been solved, after several moves,
the nr of options for the second side is drastically reduced,
so it is not suprising a forced win can be found if
the first side starts in the middle ; not with chess though,
1.e4 seems the most promising line, but black can maintain
a draw; better for (human) chess anyway, i suppose..
:)
PS2 Sanny i would kindly suggest you set up your own forum
for technical getclub chesssprogram discussions, or get some
more advice at the programmers forum at icdchess.com;
it would probably benefit your progress, and leave some
room for other topics on the chess usenet groups..
> Bye
> Sanny
>
> Play Chess at: http://www.GetClub.com/Chess.html
>
>
>
|
| |
Date: 12 Sep 2008 21:42:34
From: Simon Krahnke
Subject: Re: Another game to analyze.
|
* help bot <[email protected] > (20:32) schrieb:
> The Java platform seems to have slowed
> Sanny's program down quite a bit, and his
> "Bugster" programming language may not
> be ideal for chess.
The code written is not ideal for the Java programming language.
mfg, simon .... l
|
| |
Date: 12 Sep 2008 11:32:43
From: help bot
Subject: Re: Another game to analyze.
|
On Sep 12, 1:36=A0pm, Martin Brown <
|
| |
Date:
From: Martin Brown
Subject: Re: Another game to analyze.
|
|
| |
Date: 12 Sep 2008 07:09:20
From: Sanny
Subject: Re: Another game to analyze.
|
After 2 month today GetClub Chess was made Twice Stronger by removing
a Buggy code with better code.
Now, GetClub will play much stronger moves !!!.
Last 2 months there was no improvement, But today somehow GetClub was
made stronger.
Now very tough games at GetClub.
Beginner: 2200+
Easy: 2300+
Normal: 2400+
Play Chess at: http://www.GetClub.com/Chess.html
A few advices of Help Bot were used to improve the game. So now you
will see a real tough game at GetClub.
Now, you can play against Human Opponents also. Just choose your
Opponent and start playing with him. But you need to wait for some
opponent to arrive.
Bye
Sanny.
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 11 Sep 2008 23:52:35
From: help bot
Subject: Re: How about your this game
|
On Sep 12, 1:28=A0am, Sanny <[email protected] > wrote:
> This game was played after the Bug was removed still Help Bot easily
> pin the Rook of Easy Level and win the game.
> Please let me know the weakneses in Easy level game.
> White -- Black
> (easy) -- (help bot)
>
> 1. d2-d4{2} Ng8-f6{6}
> 2. c2-c4{0} c7-c5{6}
> 3. d4-d5{0} g7-g6{10}
> 4. Ng1-f3{98} Bf8-g7{20}
> 5. Nf3-e5{94}
One obvious weakness is the program's
decided tendency to leap Knights forward,
to no effect (apart from a loss of time).
> ... Ke8-g8{22}
That's me, castling to get my King out
of the center-- take note that this occurs
at only move five.
> 6. a2-a4{272} d7-d6{54}
> 7. Ne5-d3{40} e7-e6{32}
> 8. d5-e6{22} Bc8-e6{76}
> 9. Qd1-c2{44} Nb8-c6{24}
> 10. e2-e3{36}
Okay, here we are at move ten-- double
the move number where I got my King
out of harm's way. Q: where is GetClub's
King? A: still in the center!
> ... Rf8-e8{92}
The heavy artillery is rolled into position,
and the half-ton shells are being loaded.
> 11. f2-f3{94} b7-b6{78}
> 12. g2-g4{50}
Move twelve, and what is White doing?
Still messing around, waiting for the axe
to fall, for the chickens to come home to
roost, for bell-bottoms to come back in
style.
> ... d6-d5{76}
> 13. c4-d5{0} Be6-d5{22}
> 14. Nb1-d2{278} Nc6-d4{56}
> 15. Qc2-b1{30} Nf6-g4{66}
> 16. f3-g4{172} Bd5-h1{36}
The game is already decided by this
point, but what the heck-- let's see how
much longer GetClub can expose itself,
er, I mean expose its King to attack.
> 17. h2-h3{146} Bh1-d5{92}
> 18. Ke1-d1{20} Bd5-b3{80}
> 19. Nd2-b3{0} Nd4-b3{4}
> 20. Ra1-a3{32} c5-c4{18}
> 21. Ra3-b3{58} c4-b3{10}
> 22. Kd1-e2{52} Ra8-c8{20}
> 23. Bc1-d2{34} Rc8-c2{34}
> 24. Qb1-d1{70} Bg7-d4{122}
> 25. e3-e4{210} Re8-e4{26}
> 26. Ke2-f3{0} Qd8-d5{30}
> 27. Kf3-g3{24} Bd4-e5{46}
> 28. Kg3-f2{52} Be5-d4{66}
Uh oh-- it looks like I'm gonna need
more men for this job. Maybe if I try
zapping a key defender?
> 29. Kf2-g3{58} Rc2-d2{118}
> 30. Qd1-d2{42} Re4-e3{18}
> 31. Kg3-h2{50} Qd5-e4{132}
> 32. Qd2-d1{138} Kg8-g7{158}
> 33. a4-a5{96} b6-a5{22}
> 34. Qd1-c1{230} Re3-d3{62}
> 35. Bf1-d3{22} Qe4-d3{18}
> 36. Qc1-e1{26} a5-a4{30}
> 37. h3-h4{28} a4-a3{22}
> 38. b2-a3{44} b3-b2{16}
> 39. Kh2-g2{196} Rb2-b1{R}{64}
That's pawn-to-b1, promotes to Rook
for you folks who can decipher Bacchi.
> 40. Qe1-b1{32} Qd3-b1{26}
> 41. h4-h5{34} Qb1-e4{24}
> 42. Kg2-g3{32} g6-g5{32}
> 43. a3-a4{10}
> Was there any way to save its Rook.
Wrong question. The issue is: was there
any way to save its King? All the other men
are expendable in chess, while the very
object of the game revolves around check-
mating the opponent's King.
This brings us back to the problem of
leaving one's King in the center of the
board, where it is often subject to attack.
---------------------------------------------------
I (somewhat) enjoy playing against this
unconventional style, where I am not
faced with rote developing moves that
can lead to relatively sterile positions.
The bizarre Knight hops, the refusal to
castle or develop "normally", can lead
to very interesting chess, even if it is a
bit one-sided.
-- help bot
|
| |
Date: 11 Sep 2008 22:28:51
From: Sanny
Subject: How about your this game
|
Game between Help Bot and Easy Level.
This game was played after the Bug was removed still Help Bot easily
pin the Rook of Easy Level and win the game.
Please let me know the weakneses in Easy level game.
Game Played between help bot and easy at GetClub.com
--------------------------------------------------------------------------------
help bot: (Black)
easy: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM26689&game=Chess
--------------------------------------------------------------------------------
White -- Black
(easy) -- (help bot)
1. d2-d4{2} Ng8-f6{6}
2. c2-c4{0} c7-c5{6}
3. d4-d5{0} g7-g6{10}
4. Ng1-f3{98} Bf8-g7{20}
5. Nf3-e5{94} Ke8-g8{22}
6. a2-a4{272} d7-d6{54}
7. Ne5-d3{40} e7-e6{32}
8. d5-e6{22} Bc8-e6{76}
9. Qd1-c2{44} Nb8-c6{24}
10. e2-e3{36} Rf8-e8{92}
11. f2-f3{94} b7-b6{78}
12. g2-g4{50} d6-d5{76}
13. c4-d5{0} Be6-d5{22}
14. Nb1-d2{278} Nc6-d4{56}
15. Qc2-b1{30} Nf6-g4{66}
16. f3-g4{172} Bd5-h1{36}
17. h2-h3{146} Bh1-d5{92}
18. Ke1-d1{20} Bd5-b3{80}
19. Nd2-b3{0} Nd4-b3{4}
20. Ra1-a3{32} c5-c4{18}
21. Ra3-b3{58} c4-b3{10}
22. Kd1-e2{52} Ra8-c8{20}
23. Bc1-d2{34} Rc8-c2{34}
24. Qb1-d1{70} Bg7-d4{122}
25. e3-e4{210} Re8-e4{26}
26. Ke2-f3{0} Qd8-d5{30}
27. Kf3-g3{24} Bd4-e5{46}
28. Kg3-f2{52} Be5-d4{66}
29. Kf2-g3{58} Rc2-d2{118}
30. Qd1-d2{42} Re4-e3{18}
31. Kg3-h2{50} Qd5-e4{132}
32. Qd2-d1{138} Kg8-g7{158}
33. a4-a5{96} b6-a5{22}
34. Qd1-c1{230} Re3-d3{62}
35. Bf1-d3{22} Qe4-d3{18}
36. Qc1-e1{26} a5-a4{30}
37. h3-h4{28} a4-a3{22}
38. b2-a3{44} b3-b2{16}
39. Kh2-g2{196} Rb2-b1{R}{64}
40. Qe1-b1{32} Qd3-b1{26}
41. h4-h5{34} Qb1-e4{24}
42. Kg2-g3{32} g6-g5{32}
43. a3-a4{10}
--------------------------------------------------------------------------------
help bot: (Black)
easy: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM26689&game=Chess
Was there any way to save its Rook.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 11 Sep 2008 15:48:01
From: help bot
Subject: Re: Another game to analyze.
|
On Sep 11, 5:45=A0am, Martin Brown <
|
| |
Date: 11 Sep 2008 13:06:43
From: Taylor Kingston
Subject: Re: "Charlie" Bug found & Removed.
|
On Sep 11, 3:52=A0pm, Mike Murray <[email protected] > wrote:
> On Thu, 11 Sep 2008 10:40:50 -0700 (PDT), SBD <[email protected]>
> wrote:
>
> >On Sep 11, 12:22=A0pm, Sanny <[email protected]> wrote:
>
> >> What do you say Chalie and Martin Brown?
>
> >He a clown, that Chalie Brown.
>
> But will he get caught?
>
> We'll have to wait and see.
Why's evuhbody always pickin' on San-nee?
http://www.youtube.com/watch?v=3Dd2hoRQDXGVo
|
| | |
Date: 11 Sep 2008 13:43:40
From: Frisco Del Rosario
Subject: Re: "Charlie" Bug found & Removed.
|
In article
<[email protected] >,
Taylor Kingston <[email protected] > wrote:
> > >He a clown, that Chalie Brown.
> >
> > But will he get caught?
> >
> > We'll have to wait and see.
>
> Why's evuhbody always pickin' on San-nee?
Click through the software nice and slow
Watch it freezing up NO NO NO!!!
GetClub now
Improved wow!
GetClub now
Improved wow!
It's gonna beat Rybka
Just you wait and see
Why is everybody always pickin' on Sanny?
|
| |
Date: 11 Sep 2008 10:40:50
From: SBD
Subject: Re: "Charlie" Bug found & Removed.
|
On Sep 11, 12:22=A0pm, Sanny <[email protected] > wrote:
> What do you say Chalie and Martin Brown?
He a clown, that Chalie Brown.
|
| | |
Date: 11 Sep 2008 12:52:13
From: Mike Murray
Subject: Re: "Charlie" Bug found & Removed.
|
On Thu, 11 Sep 2008 10:40:50 -0700 (PDT), SBD <[email protected] >
wrote:
>On Sep 11, 12:22�pm, Sanny <[email protected]> wrote:
>
>> What do you say Chalie and Martin Brown?
>
>
>He a clown, that Chalie Brown.
But will he get caught?
We'll have to wait and see.
|
| |
Date: 11 Sep 2008 10:26:16
From: Sanny
Subject: Re: Another game to analyze.
|
> Hmm, looks like I've been chummed. =A0I assumed the OP was a genuine
> request for assistance. =A0Ah well, fool me once. =A0And oh by the way, I
> think this Sanny is trying to tell me that the "beginner" level is of
> 2200 quality. =A0Well, if "beginner" is his program playing the engine
> "Jester" in the OP'd game, I'd give it about 800, no more. =A0The Jester
> side played at about 1200. =A0Food for thought Sanny, but don't expect
> any more replies from me.
There was a bug when that game was played today Beginner beat the
Jester.
Beginner : 15 sec / move
Jester: 3 sec / move
Is Jester 1200? at above time control?
Game Played between Jester and beginner at GetClub.com
---------------------------------------------------------------------------=
-----
Jester : (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=3DDM26662&game=
=3DChess
---------------------------------------------------------------------------=
-----
White -- Black
(Jester ) -- (beginner)
1. e2-e4{4} c7-c5{0}
2. Ng1-f3{4} g7-g6{0}
3. d2-d4{6} c5-d4{0}
4. Nf3-d4{8} Bf8-g7{0}
5. Nb1-c3{10} h7-h5{48}
6. Bc1-e3{10} a7-a6{10}
7. f2-f4{12} d7-d6{28}
8. Bf1-e2{10} Ng8-f6{26}
9. f4-f5{10} e7-e5{6}
10. f5-e6{10} f7-e6{6}
11. Ke1-g1{12} Ke8-g8{12}
12. Be2-c4{12} d6-d5{8}
13. e4-d5{10} e6-d5{12}
14. Bc4-b3{10} Nb8-c6{18}
15. Nd4-e2{12} Nc6-a5{34}
16. Bb3-d5{12} Nf6-d5{10}
17. Rf1-f8{14} Kg8-f8{10}
18. Ne2-f4{14} Na5-c4{14}
19. Nc3-d5{12} Bg7-b2{6}
20. Nf4-g6{14} Kf8-f7{10}
21. Ng6-e5{10} Bb2-e5{10}
22. Qd1-h5{10} Kf7-g7{18}
23. Be3-g5{10} Qd8-d5{6}
24. Ra1-f1{12} Be5-d4{8}
25. Kg1-h1{10} Bc8-e6{22}
26. a2-a4{14} Ra8-h8{26}
27. Qh5-e2{10} Nc4-b6{30}
28. Bg5-e3{12} Bd4-e3{8}
29. Qe2-e3{10} Nb6-d7{6}
30. Qe3-g3{12} Kg7-h6{10}
31. Qg3-h4{12} Qd5-h5{14}
32. Qh4-d4{14} Rh8-e8{34}
33. Qd4-e4{12} b7-b6{22}
34. Rf1-b1{14} Be6-d5{6}
35. Qe4-f4{10} Qh5-g5{26}
36. Qf4-f2{10} Bd5-c6{10}
37. Rb1-d1{10} Re8-f8{12}
38. Qf2-e2{10} Bc6-a4{8}
39. Qe2-a6{10} Ba4-c2{6}
40. Rd1-d6{12} Rf8-f6{12}
41. Qa6-a3{16} Bc2-f5{16}
42. Rd6-f6{10} Nd7-f6{26}
43. Qa3-a1{12} Bf5-e4{6}
44. Qa1-b2{12} Qg5-g7{10}
45. Qb2-f2{12} b6-b5{12}
46. Kh1-g1{12} Qg7-g6{10}
47. h2-h3{10} b5-b4{10}
48. g2-g3{10} Kh6-g7{12}
49. Qf2-a7{14} Qg6-f7{10}
50. Qa7-a5{10} Qf7-b3{12}
51. Qa5-g5{12} Kg7-f7{6}
52. Kg1-f1{10} Qb3-f3{6}
53. Kf1-e1{14} Qf3-h1{0}
54. Ke1-e2{14} Qh1-h3{8}
55. Qg5-e5{10} Qh3-h2{10}
56. Ke2-d1{10} Qh2-c2{8}
57. Kd1-e1{8} Qc2-c3{0}
58. Qe5-c3{12} b4-c3{8}
59. g3-g4{10} c3-c2{44}
60. Ke1-d2{10} Nf6-g4{6}
61. Kd2-c1{12} Ng4-e5{6}
62. Kc1-d2{14} Ne5-c4{8}
---------------------------------------------------------------------------=
-----
Jester : (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=3DDM26662&game=
=3DChess
So what do you say about this victory of Beginner Level.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 11 Sep 2008 10:22:00
From: Sanny
Subject: "Charlie" Bug found & Removed.
|
This game had a bug when playing that was removed today.
Today after the Bug was removed GetClub Beginner Level defeated Jester
and Beginner win the game.
Here is the game where Beginner win against the Jester.
Game Played between sanjay11 and beginner at GetClub.com
--------------------------------------------------------------------------------
sanjay11: (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM26662&game=Chess
--------------------------------------------------------------------------------
White -- Black
(sanjay11) -- (beginner)
1. e2-e4{4} c7-c5{0}
2. Ng1-f3{4} g7-g6{0}
3. d2-d4{6} c5-d4{0}
4. Nf3-d4{8} Bf8-g7{0}
5. Nb1-c3{10} h7-h5{48}
6. Bc1-e3{10} a7-a6{10}
7. f2-f4{12} d7-d6{28}
8. Bf1-e2{10} Ng8-f6{26}
9. f4-f5{10} e7-e5{6}
10. f5-e6{10} f7-e6{6}
11. Ke1-g1{12} Ke8-g8{12}
12. Be2-c4{12} d6-d5{8}
13. e4-d5{10} e6-d5{12}
14. Bc4-b3{10} Nb8-c6{18}
15. Nd4-e2{12} Nc6-a5{34}
16. Bb3-d5{12} Nf6-d5{10}
17. Rf1-f8{14} Kg8-f8{10}
18. Ne2-f4{14} Na5-c4{14}
19. Nc3-d5{12} Bg7-b2{6}
20. Nf4-g6{14} Kf8-f7{10}
21. Ng6-e5{10} Bb2-e5{10}
22. Qd1-h5{10} Kf7-g7{18}
23. Be3-g5{10} Qd8-d5{6}
24. Ra1-f1{12} Be5-d4{8}
25. Kg1-h1{10} Bc8-e6{22}
26. a2-a4{14} Ra8-h8{26}
27. Qh5-e2{10} Nc4-b6{30}
28. Bg5-e3{12} Bd4-e3{8}
29. Qe2-e3{10} Nb6-d7{6}
30. Qe3-g3{12} Kg7-h6{10}
31. Qg3-h4{12} Qd5-h5{14}
32. Qh4-d4{14} Rh8-e8{34}
33. Qd4-e4{12} b7-b6{22}
34. Rf1-b1{14} Be6-d5{6}
35. Qe4-f4{10} Qh5-g5{26}
36. Qf4-f2{10} Bd5-c6{10}
37. Rb1-d1{10} Re8-f8{12}
38. Qf2-e2{10} Bc6-a4{8}
39. Qe2-a6{10} Ba4-c2{6}
40. Rd1-d6{12} Rf8-f6{12}
41. Qa6-a3{16} Bc2-f5{16}
42. Rd6-f6{10} Nd7-f6{26}
43. Qa3-a1{12} Bf5-e4{6}
44. Qa1-b2{12} Qg5-g7{10}
45. Qb2-f2{12} b6-b5{12}
46. Kh1-g1{12} Qg7-g6{10}
47. h2-h3{10} b5-b4{10}
48. g2-g3{10} Kh6-g7{12}
49. Qf2-a7{14} Qg6-f7{10}
50. Qa7-a5{10} Qf7-b3{12}
51. Qa5-g5{12} Kg7-f7{6}
52. Kg1-f1{10} Qb3-f3{6}
53. Kf1-e1{14} Qf3-h1{0}
54. Ke1-e2{14} Qh1-h3{8}
55. Qg5-e5{10} Qh3-h2{10}
56. Ke2-d1{10} Qh2-c2{8}
57. Kd1-e1{8} Qc2-c3{0}
58. Qe5-c3{12} b4-c3{8}
59. g3-g4{10} c3-c2{44}
60. Ke1-d2{10} Nf6-g4{6}
61. Kd2-c1{12} Ng4-e5{6}
62. Kc1-d2{14} Ne5-c4{8}
--------------------------------------------------------------------------------
sanjay11: (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM26662&game=Chess
So you see how Beginner won the game. Was there any better way of
winning?
What do you say Chalie and Martin Brown?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date:
From: Martin Brown
Subject: Re: Another game to analyze.
|
|
| | |
Date: 11 Sep 2008 06:45:52
From: Charlie Siegrist
Subject: Re: Another game to analyze.
|
On Thu, 11 Sep 2008 02:45:35 -0700 (PDT), Martin Brown
<
|
| | | |
Date: 11 Sep 2008 15:33:18
From: David Richerby
Subject: Re: Another game to analyze.
|
Charlie Siegrist <[email protected] > wrote:
> Hmm, looks like I've been chummed.
Yeah, sorry.
> I assumed the OP was a genuine request for assistance. Ah well,
> fool me once. And oh by the way, I think this Sanny is trying to
> tell me that the "beginner" level is of 2200 quality.
He is, yes. Draw your own conclusions.
> Well, if "beginner" is his program playing the engine "Jester" in
> the OP'd game
It was, yes. Jester's actually a pretty decent program so I guess it
was playing at a very restricted time control and/or on a resource-
starved machine (perhaps starved by Sanny's program running at the
same time).
Dave.
--
David Richerby Incredible Solar-Powered Gnome (TM):
www.chiark.greenend.org.uk/~davidr/ it's like a smiling garden ornament
but it doesn't work in the dark and
it'll blow your mind!
|
| |
Date: 10 Sep 2008 23:07:29
From: Sanny
Subject: Another game to analyze.
|
Help Bot easily took White pieces due to think Normal Level was
playing wrong moves?
Game Played between help bot and normal at GetClub.com
--------------------------------------------------------------------------------
help bot: (Black)
normal: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM26644&game=Chess
--------------------------------------------------------------------------------
White -- Black
(normal) -- (help bot)
1. e2-e4{2} c7-c5{6}
2. Nb1-c3{224} g7-g6{26}
3. Ng1-f3{444} Bf8-g7{20}
4. Bf1-c4{492} Nb8-c6{24}
5. b2-b3{500} Ng8-f6{42}
6. Ra1-b1{254} Ke8-g8{30}
7. Ke1-g1{298} e7-e6{88}
8. Nf3-g5{198} d7-d5{32}
9. e4-d5{160} e6-d5{272}
10. Bc4-e2{98} h7-h6{104}
11. Ng5-f3{106} d5-d4{48}
12. Nc3-a4{94} b7-b6{158}
13. Bc1-b2{462} Rf8-e8{138}
14. Be2-c4{208} a7-a6{86}
15. Rf1-e1{90} Re8-e1{48}
16. Qd1-e1{82} b6-b5{70}
17. Bc4-f7{106} Kg8-f7{70}
18. Na4-c5{108} Qd8-d5{74}
19. Nc5-d3{118} Nf6-h5{44}
20. Rb1-c1{176} b5-b4{54}
21. Kg1-h1{94} Bc8-g4{78}
22. Qe1-e2{98} Ra8-e8{18}
23. c2-c4{428} d4-c3{36}
24. Qe2-f1{370} Bg4-f3{54}
25. d2-c3{144} Bf3-e2{48}
26. Qf1-e1{152} Be2-d3{34}
27. Qe1-g1{208} Re8-e2{76}
28. c3-c4{214} Qd5-e6{70}
29. Bb2-a1{354} Bg7-a1{116}
30. Rc1-a1{84} Nh5-f6{70}
31. f2-f3{116} Nc6-e7{50}
32. Qg1-d4{354} Ne7-f5{118}
33. Qd4-c5{122} Nf6-d7{76}
34. Qc5-d5{458} Qe6-d5{38}
35. c4-d5{204} Nf5-e3{46}
36. d5-d6{88} Ne3-g2{54}
37. Ra1-f1{340} Ng2-e3{72}
38. Rf1-a1{294} Nd7-e5{72}
39. d6-d7{4}
--------------------------------------------------------------------------------
help bot: (Black)
normal: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM26644&game=Chess
What were the major problems in Normal Level game.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 10 Sep 2008 21:21:42
From: Charlie Siegrist
Subject: Re: Pls analyze this game of Beginner Level
|
On Sat, 6 Sep 2008 00:57:46 -0700 (PDT), Sanny <[email protected] >
wrote:
>What do you say, What were the wrong moves by Beginner that it should
>have avoided?
Ok, I gave it a go, shooting from the hip. I checked a couple
tactical shots afterward with Ruffian. Basically, black's choice of
opening is not so good for a beginner, and the play is poor
throughout. The *very* basic opening guidelines - develop minor
pieces, occupy the center with pawns, don't move the queen early,
prepare to castle - are *all* violated. Hope this helps, read a book!
>Game Played between Jester and beginner at GetClub.com
>
>--------------------------------------------------------------------------------
>Jester : (White)
>beginner: (Black)
>Game Played at: http://www.GetClub.com/Chess.html
>View Recorded Game: http://www.getclub.com/playgame.php?id=DM26050&game=Chess
>--------------------------------------------------------------------------------
>
>White -- Black
>(Jester ) -- (beginner)
>
>1. e2-e4{8} d7-d5{0}
>2. e4-d5{6} Qd8-d5{0}
>3. Nb1-c3{6} Qd5-d8{0}
>4. Bf1-b5{10}
I see this as almost provoking a strengthening of black's position,
besides wasting time.
> ... c7-c6{18}
>5. Bb5-c4{10} Ng8-f6{8}
>6. d2-d3{18} Bc8-g4{82}
>7. f2-f3{12}
Ack. Play Nf3.
> ... Bg4-e6{6}
Double ack. If it isn't obvious, this wrecks black's center pawns, a
caution every beginner should know. The game is about lost here, but
for the fact that white continues to play poorly for a while.
>8. Bc4-e6{12} f7-e6{8}
>9. Qd1-e2{10}
Steals the knight's developing square, and moves the queen early.
While black's early queen move is part and parcel of the opening
theory, which many believe only marginally playable, but this move is
just wrongly timed.
So far, white has made several beginner mistakes, only to be answered
by worse mistakes by black.
> ... Qd8-d6{8}
Like I just said. Let's take a look at what's going on here. After
nine moves, neither white nor black is prepared to castle, and won't
be for several more moves. Black's pawns are hopeless, as is the king
bishop. Only one minor piece is out for each player. This game so
far has violated almost every dictum in chapter one of every beginner
chess book.
>10. Ng1-h3{10}
The Knight's only square, and it stinks.
>...Nb8-d7{8}
>11. Ke1-g1{12} g7-g6{8}
Better would have been e5 or even h6, trying to keep white's bad
knight bad, and will somebody please start to put pressure on the
center soon?
>12. Nh3-g5{16} e6-e5{6}
>13. Kg1-h1{12}
Um, why?
> ... Bf8-g7{6}
>14. Nc3-e4{12} Nf6-e4{8}
>15. Ng5-e4{10}
Probably should develop the queen bishop before entertaining an
attack.
>... Qd6-d4{6}
Probably would be more comfortable on c7.
>16. Bc1-e3{12} Qd4-d5{10}
>17. Qe2-f2{12} b7-b6{8}
>18. Qf2-h4{10}
A lot of time wasting going on here, by both players, moving queens to
and fro, aimlessly trying to find a strong square.
>...Nd7-f6{20}
>19. Be3-d2{12} Nf6-e4{6}
Really, really bad. Removes a well posted piece from the board and
strengthens white's center pawns while opening a file for the white
rook on f1. Is white moving black's pieces here? (I have shut down
Arena, but IIRC, this piece is supported by the backward e pawn, so
the hope remains that white may foolishly take this piece and allow
black to repair his pawns - another reason to keep it where it is, if
one were needed. The final reason to keep it here is that its
obstructive effect will allow black to castle eventually. This was
black's best minor piece, and it is cast to the wind for less than
nothing.)
>20. f3-e4{10} Qd5-d4{8}
Since white has not been taking any initiative at all in attacking the
weak black center pawns, black decides to provoke the white attack.
>21. Bd2-c3{14} Qd4-e3{12}
>22. Qh4-g4{12} Bg7-f6{24}
Why not Rf8? Hmm, ok, Ruffian says only Qh6 (followed by g5) can take
the sting out of the coming Qe6. At this point, though, having failed
to castle and having voluntarily ruined its pawn structure, black is
hopeless.
>23. Qg4-e6{12} Qe3-g5{0}
>24. Rf1-f2{10} Ke8-d8{6}
>25. Qe6-c6{12} Ra8-b8{8}
>26. Ra1-f1{12} Rh8-g8{14}
Black is finding the most ineffectual squares possible for the rooks,
and I wonder if white could have achieved this position but for the
provocation of black to move the pieces where they have ended up?
>27. Qc6-d5{12} Kd8-c8{28}
Why does this move not surprise me?
>28. Qd5-g8{10} Kc8-b7{10}
>29. Qg8-d5{16} Kb7-c8{12}
>30. a2-a4{14} h7-h5{14}
>31. a4-a5{28} Rb8-b7{20}
>32. a5-b6{14} a7-b6{0}
>33. Qd5-c6{10} Rb7-c7{6}
>34. Qc6-b6{10} Rc7-d7{8}
>35. Rf1-a1{12} Rd7-c7{10}
>36. Qb6-b5{10} Rc7-a7{18}
If not for the unnecessary white move #13, this swindle would not be
playable. I'm guessing "Jester" is the engine playing white, so it's
unlikely that even a weak engine will fall for a mate-in-one swindle.
I haven't seen that from a computer since the seventies.
>37. Qb5-e8{14} Kc8-b7{6}
>38. Qe8-d7{14} Kb7-b8{0}
>39. Qd7-a7{14} Kb8-c8{0}
>40. Qa7-a8{10} Kc8-c7{2}
>--------------------------------------------------------------------------------
> Jester : (White)
>beginner: (Black)
>Game Played at: http://www.GetClub.com/Chess.html
>View Recorded Game: http://www.getclub.com/playgame.php?id=DM26050&game=Chess
>
>Please let me know your opinion about the above moves.
>
>Bye
>Sanny
>
>Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date:
From: Martin Brown
Subject: Re: Pls analyze this game of Beginner Level
|
|
| | |
Date: 11 Sep 2008 14:43:56
From: David Richerby
Subject: Re: Pls analyze this game of Beginner Level
|
[Egregious cross-post trimmed.]
Martin Brown <
|
|
