![[ru]](/mwf/v12/flags/ru.png "Russian Federation")
Another ways for 1) are
a) Select Home-View tab
b) Click Copy as EPD
or
a) Right-click the mouse in game notation.
b) Select Copy/ Copy position as EPD
a) Select Home-View tab
b) Click Copy as EPD
or
a) Right-click the mouse in game notation.
b) Select Copy/ Copy position as EPD
Thanks to Dadi and 7-men TB for the explanation of how to generate an EPD file, which I sent to 7-men TB. He promptly gave me the result of the TB.
His work is an amazing and pioneering accomplishment that advances our knowledge of chess endings.
Will it be a long time before all (or most of) the 7-men TBs are generated?
Perhaps Chess OK and/or Shredder will agree to financially support the work in return for using the output as it becomes available?
His work is an amazing and pioneering accomplishment that advances our knowledge of chess endings.
Will it be a long time before all (or most of) the 7-men TBs are generated?
Perhaps Chess OK and/or Shredder will agree to financially support the work in return for using the output as it becomes available?
Interesting project! Here is another example:
8/2k4p/5R2/5p2/6P1/6K1/8/7r w - - 0 68
That's draw. 1.Kf3
Thank you so much! :-)
![[pr]](/mwf/v12/flags/pr.png "Puerto Rico")
What about this one :
8/6k1/8/1p6/r7/4R1PP/8/7K b - - 0 54
{DRAW} 1... Kf7
![[us]](/mwf/v12/flags/us.png "United States")
This position is interesting in that white to move could win.
The position below occurred in Anand-Shirov, Wijk aan Zee 2004. Karsten Mueller (in his book "How to play Chess Endgames") and V. Khuzman in CBM both mistakenly conclude the position is a draw. Anand played 1.f5? and eventually won due to further mistakes by Shirov. Khuzman only considered 1.Rxa7? and 1.Rg5? leading to draws. What is the winning move?
6r1/p7/3k4/R7/5P2/8/P6K/8 w - - 0 1
![[ie]](/mwf/v12/flags/ie.png "Ireland")
> What is the winning move?
Ra5-e5
![[de]](/mwf/v12/flags/de.png "Germany")
I agree.
Thanks for this interesting position. I beleive that engines are able to find the win without tables.
[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "Evaluation: 87 half-moves to mate"]
[Black "?"]
[Result "*"]
[FEN "
[SetUp "1"]
1. Re5 $1 1... Rg4 2. Re4 $1 2... a5
3. a4 Rg6 4. Kh3 Rg8 5. Re5 Rh8+ 6. Kg4 Rg8+ 7. Kf5 Ra8 8. Ke4 Ra7 9. Kd4 Rb7
10. Rxa5 Rb4+ {[%t bLon] Mate in 34:} 11. Ke3 $1 11... Ke6 12. Ra6+ Kf5 13.
Ra8 Rb3+ 14. Kd4 Rb4+ 15. Kc5 Rb3 16. a5 Rc3+ 17. Kb6 Rb3+ 18. Ka7 Kxf4 19. a6
Rb2 20. Rb8 Rg2 21. Rb5 Rg3 22. Rb4+ Ke5 23. Kb6 Rg6+ 24. Kb5 Rg7 25. Ra4 Rg8
26. a7 Ra8 27. Kc6 Rc8+ 28. Kb7 Rd8 29. a8=Q Rxa8 30. Rxa8 Kd4 31. Ra4+ Kc5
32. Kc7 Kd5 33. Kd7 Ke5 34. Kc6 Kf6 35. Ra5 Ke7 36. Re5+ Kf6 37. Kd6 Kf7 38.
Re6 Kf8 39. Re7 Kg8 40. Rd7 Kh8 41. Ke5 Kg8 42. Kf6 Kh8 43. Kg6 Kg8 44. Rd8# *
[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "Evaluation: 87 half-moves to mate"]
[Black "?"]
[Result "*"]
[FEN "
6r1/p7/3k4/R7/5P2/8/P6K/8 w - - 0 1"][SetUp "1"]
1. Re5 $1 1... Rg4 2. Re4 $1 2... a5
3. a4 Rg6 4. Kh3 Rg8 5. Re5 Rh8+ 6. Kg4 Rg8+ 7. Kf5 Ra8 8. Ke4 Ra7 9. Kd4 Rb7
10. Rxa5 Rb4+ {[%t bLon] Mate in 34:} 11. Ke3 $1 11... Ke6 12. Ra6+ Kf5 13.
Ra8 Rb3+ 14. Kd4 Rb4+ 15. Kc5 Rb3 16. a5 Rc3+ 17. Kb6 Rb3+ 18. Ka7 Kxf4 19. a6
Rb2 20. Rb8 Rg2 21. Rb5 Rg3 22. Rb4+ Ke5 23. Kb6 Rg6+ 24. Kb5 Rg7 25. Ra4 Rg8
26. a7 Ra8 27. Kc6 Rc8+ 28. Kb7 Rd8 29. a8=Q Rxa8 30. Rxa8 Kd4 31. Ra4+ Kc5
32. Kc7 Kd5 33. Kd7 Ke5 34. Kc6 Kf6 35. Ra5 Ke7 36. Re5+ Kf6 37. Kd6 Kf7 38.
Re6 Kf8 39. Re7 Kg8 40. Rd7 Kh8 41. Ke5 Kg8 42. Kf6 Kh8 43. Kg6 Kg8 44. Rd8# *
Sorry, I will not be available next two days. So it will be no answers from me here.
Here's another example, away from theoretical endgames. What have you for a result?
Thank you.
5r2/1P6/8/4k2P/4B3/4K1P1/8/8 b - - 0 61
Thank you.
Sorry, the tables for this ending are planned later this year. Curently KBPP-KBP is under generation.
![[in]](/mwf/v12/flags/in.png "India")
How big is the 7 man TB for KPP-KP?
The size of RPP-RP is relatively small 111 GB (119312209006 bytes). But to build it you need to generate more than 9TB (packed size) of other tables.
Need to say that for practical usage the sizes of the tables possibly will increase (more than 10%). Current pages inside the tables are too large, so random access to positions is not fast.
111 GB is not a good size for a usual person possible to operate. May be there is sense to provide bit-bases. This makes practical usage a reality.
Need to say that for practical usage the sizes of the tables possibly will increase (more than 10%). Current pages inside the tables are too large, so random access to positions is not fast.
111 GB is not a good size for a usual person possible to operate. May be there is sense to provide bit-bases. This makes practical usage a reality.
For comparison, our DTC RPP-RP tables, with 128k block size and 7zip compression, are about 62 GB. DTZ would be a bit smaller still, confirming that DTM vs DTZ will have about a factor of two space penalty.
62 GB is interesting. I also have impression of about 2x ratio of DTM to DTC size.
It will be interesting to generate our RPP-RP in DTC metrics (it is an easy change) and to look at the size. We also use LZMA compression, but not directly from 7-zip (I feel that 7-zip implementation of LZMA for low-density data is better).
128k block size is still not good for practical purposes but much closer than our current 2MB block size. Nalimov used 8k size. Possibly 32kb will be reasonable compromise between size and speed.
Did you implement some cache algorithm for reading data in random order?
It will be interesting to generate our RPP-RP in DTC metrics (it is an easy change) and to look at the size. We also use LZMA compression, but not directly from 7-zip (I feel that 7-zip implementation of LZMA for low-density data is better).
128k block size is still not good for practical purposes but much closer than our current 2MB block size. Nalimov used 8k size. Possibly 32kb will be reasonable compromise between size and speed.
Did you implement some cache algorithm for reading data in random order?
For basic analysis purposes a large block size is fine. I actually use 8MB for this ending. There were fewer than 1.5 million krppkrp positions collected from about 6 million games, so even with a large block size (and reasonable caching) this was scanned pretty quickly in a few hours.
But the large size is the main reason Yakov and I have not made these tablebases available to a larger audience.
But the large size is the main reason Yakov and I have not made these tablebases available to a larger audience.
... okay, thank you!
Black loses according to the database Yakov Konoval and I created for this ending, although in this case the win is so straightforward the database is not really needed.
... thanks for replying. What a database you created?
For this ending, created BPPP-R were only queen promotions are considered. This ending is relatively rare, so creating all the possble promotions would be a huge space waste without creating much additional accuracy. Obviously if White can win with queen promotions only, as he can in your position, he could certainly also win with all promotions considered.
My analysis of Anand-Gelfand (game 3) has reached this ending:
The verdict?
Thanks,
Charles Sullivan
5R2/8/8/K4R2/3r4/1P6/r5k1/8 w - - 0 58
The verdict?
Thanks,
Charles Sullivan
[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "mate in 103 halfmoves"]
[Black "?"]
[Result "1-0"]
[SetUp "1"]
[FEN "
1.Kb5 Rg4 2.Rf4 Rg5+ 3.Kc4 Rc2+ 4.Kd3 Ra2 5.b4 Rg3+ 6.Kd4 Rb2 7.Kc5 Rc3+ 8.Kb6
Rcb3 9.Rg8+ Kh3 10.Kb5 Rb1 11.Ka4 R3b2 12.Rf3+ Kh4 13.Rb8 Re2 14.Rb3 Rf1 15.
Rh8+ Kg4 16.Rg8+ Kh4 17.b5 Re4+ 18.Ka3 Ra1+ 19.Kb2 Ree1 20.Rb8 Rab1+ 21.Ka3
Ra1+ 22.Kb4 Re4+ 23.Kc3 Re3+ 24.Kb2 Ree1 25.b6 Rab1+ 26.Ka3 Ra1+ 27.Kb4 Re7 28.
b7 Rh7 29.Kc3 Rc1+ 30.Kb2 Rcc7 31.Ka2 Rcd7 32.Rb1 Rd2+ 33.Ka1 Rxb7 34.R8xb7
Kg5 35.Rf1 Rh2 36.Rg7+ Kh6 37.Rfg1 Rh3 38.R1g6+ Kh5 39.Rg8 Ra3+ 40.Kb1 Ra8 41.
Rg5+ Kh4 42.Rg2 Rxg8 43.Rxg8 Kh5 44.Kc1 Kh4 45.Kd2 Kh3 46.Ke3 Kh4 47.Rg1 Kh3
48.Kf4 Kh2 49.Rg3 Kh1 50.Kf3 Kh2 51.Kf2 Kh1 52.Rh3# 1-0
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "mate in 103 halfmoves"]
[Black "?"]
[Result "1-0"]
[SetUp "1"]
[FEN "
5R2/8/8/K4R2/3r4/1P6/r5k1/8 w - -"]1.Kb5 Rg4 2.Rf4 Rg5+ 3.Kc4 Rc2+ 4.Kd3 Ra2 5.b4 Rg3+ 6.Kd4 Rb2 7.Kc5 Rc3+ 8.Kb6
Rcb3 9.Rg8+ Kh3 10.Kb5 Rb1 11.Ka4 R3b2 12.Rf3+ Kh4 13.Rb8 Re2 14.Rb3 Rf1 15.
Rh8+ Kg4 16.Rg8+ Kh4 17.b5 Re4+ 18.Ka3 Ra1+ 19.Kb2 Ree1 20.Rb8 Rab1+ 21.Ka3
Ra1+ 22.Kb4 Re4+ 23.Kc3 Re3+ 24.Kb2 Ree1 25.b6 Rab1+ 26.Ka3 Ra1+ 27.Kb4 Re7 28.
b7 Rh7 29.Kc3 Rc1+ 30.Kb2 Rcc7 31.Ka2 Rcd7 32.Rb1 Rd2+ 33.Ka1 Rxb7 34.R8xb7
Kg5 35.Rf1 Rh2 36.Rg7+ Kh6 37.Rfg1 Rh3 38.R1g6+ Kh5 39.Rg8 Ra3+ 40.Kb1 Ra8 41.
Rg5+ Kh4 42.Rg2 Rxg8 43.Rxg8 Kh5 44.Kc1 Kh4 45.Kd2 Kh3 46.Ke3 Kh4 47.Rg1 Kh3
48.Kf4 Kh2 49.Rg3 Kh1 50.Kf3 Kh2 51.Kf2 Kh1 52.Rh3# 1-0
Thank you very much. It seems that 34.d7 would have been a winner for Anand.
Charles Sullivan
Charles Sullivan
Hello,
sorry, I don't know if this has already been asked but, has it been calculated how much space will the 7 pieces tablebases occupy?
Thanks. Very interesting.
sorry, I don't know if this has already been asked but, has it been calculated how much space will the 7 pieces tablebases occupy?
Thanks. Very interesting.
Average size of one table is 130 GB. As there are 1001 7-man endings the whole size will be 130 TB. Really for some reasons we expect greater number. Operating with this size of data is a real pain.
Size of really useful tables is much less - 20-30 TB. Reducing tables to bit-bases it will be possible to place most essential info on one 4 TB disk. But we need to generate and to check everything first before making reductions.
I remember this screen was published last month for giving common impression.
The sizes of RPP-RP ending and all dependent endings: http://dl.dropbox.com/u/72038782/krpp-krp_final.png
The largest size currently has kRBPkBN table - 540 GB.
Size of really useful tables is much less - 20-30 TB. Reducing tables to bit-bases it will be possible to place most essential info on one 4 TB disk. But we need to generate and to check everything first before making reductions.
I remember this screen was published last month for giving common impression.
The sizes of RPP-RP ending and all dependent endings: http://dl.dropbox.com/u/72038782/krpp-krp_final.png
The largest size currently has kRBPkBN table - 540 GB.
Thanks, really fascinating!
I wonder how many years it will take to generate the full set or at least the most useful tables.
I wonder how many years it will take to generate the full set or at least the most useful tables.
To generate most useful you need to generate plenty of minor endings including many useless endings. So the question is how long to generate the whole set of 4+3 endings and how long to generate the whole set of 5+2 endings.
We are able to finish 4+3 endings in 2012. But who knows - huge resources are essential to support the process.
So 5+2 endings (that are much less important ) are for 2013.
We are able to finish 4+3 endings in 2012. But who knows - huge resources are essential to support the process.
So 5+2 endings (that are much less important ) are for 2013.
Are you a human? Mr 7-man TB!
His mother must have really hated him to give him this name! 
But who am I to bring this up?
But who am I to bring this up?
Why not?
Possibly the reason of the question is to know the names of the developers. That's easy:
1. Development: Vladimir Makhnychev,
2. Project management: Victor Zakharov
3. Other people also contributed their efforts into the project. Mostly all of them are employees of Convekta Ltd and Lomonosov Moscow State University. Both organizations provide invaluable technical and financial support for the project. All this will be described after official presentation of the project. Unfortunately there is too much work for support of generation process, so public actions are still delayed.
Possibly the reason of the question is to know the names of the developers. That's easy:
1. Development: Vladimir Makhnychev,
2. Project management: Victor Zakharov
3. Other people also contributed their efforts into the project. Mostly all of them are employees of Convekta Ltd and Lomonosov Moscow State University. Both organizations provide invaluable technical and financial support for the project. All this will be described after official presentation of the project. Unfortunately there is too much work for support of generation process, so public actions are still delayed.
So you're many humans it seems...
I wish you guys would return to the Development Area -at some point-soon!
Yes, the range of interesting 5-2 endgames is much smaller. Most of the difficult endings involve one side just having a queen, which does not occur that often in practice. One of those endings that does occur occasionally is krbppkq, where the queen should usually be able to draw, according to our database. A recent example of this ending is Grischuk - Gashimov (Amber-rapid 2011), where Black defended very well except for one slip, which however White did not take advantage of:
Black played 124...Qc1, after which White could have won with 125.Be6!
Almost any move that prevents 125.Be6 would have been ok for Black, such as Qe1,Qe3,Qa5.
Some of the endings that occur more often in practice are krkpppp and krkbppp, although in many of those cases engines without tablebases quickly find mistakes made by humans. One such example is Timman - Deze, Sombor 1974, which was analyzed in detail by Peter Griffiths in his book "Exploring the endgame":
[Event "Parcetic Memorial 6th"]
[Site "Sombor"]
[Date "1974.??.??"]
[Round "12"]
[White "Timman, Jan H"]
[Black "Deze, Anton"]
[Result "1-0"]
[SetUp "1"]
[FEN "1k6/8/8/3P4/PP4r1/2P1p2B/3K4/8 w - - 0 66"]
66. Kxe3 $5 {Praised by Griffiths as "four advanced pawns with their
king ideally placed easily outrun a rook, even in the presence of the enemy
king." However, the ending is actually drawn.} ({However, the other natural
winning attempt does not work either:} 66. Kd3 $5 Rg3 67. Bd7 e2+ 68. Kxe2 Rxc3
{is also drawn according to our database.}) 66... Rg3+ 67. Kd4 Rxh3 68. c4 Kc7
69. Kc5 Rh8 70. b5 Rc8 $2 {Not commented on by Griffiths, although engines
quickly find this is the decisive mistake.} ({Engines quickly find the drawing
move} 70... Kb7 $11 {because now} 71. b6 {can be met by} Ka6) 71. b6+ Kb8+ 72.
Kb5 Rh8 73. a5 Rh1 74. a6 Rb1+ 75. Kc6 Ra1 76. d6 1-0
5k2/3R4/8/5BK1/5PP1/8/3q4/8 b - - 0 124
Black played 124...Qc1, after which White could have won with 125.Be6!
Almost any move that prevents 125.Be6 would have been ok for Black, such as Qe1,Qe3,Qa5.
Some of the endings that occur more often in practice are krkpppp and krkbppp, although in many of those cases engines without tablebases quickly find mistakes made by humans. One such example is Timman - Deze, Sombor 1974, which was analyzed in detail by Peter Griffiths in his book "Exploring the endgame":
[Event "Parcetic Memorial 6th"]
[Site "Sombor"]
[Date "1974.??.??"]
[Round "12"]
[White "Timman, Jan H"]
[Black "Deze, Anton"]
[Result "1-0"]
[SetUp "1"]
[FEN "1k6/8/8/3P4/PP4r1/2P1p2B/3K4/8 w - - 0 66"]
66. Kxe3 $5 {Praised by Griffiths as "four advanced pawns with their
king ideally placed easily outrun a rook, even in the presence of the enemy
king." However, the ending is actually drawn.} ({However, the other natural
winning attempt does not work either:} 66. Kd3 $5 Rg3 67. Bd7 e2+ 68. Kxe2 Rxc3
{is also drawn according to our database.}) 66... Rg3+ 67. Kd4 Rxh3 68. c4 Kc7
69. Kc5 Rh8 70. b5 Rc8 $2 {Not commented on by Griffiths, although engines
quickly find this is the decisive mistake.} ({Engines quickly find the drawing
move} 70... Kb7 $11 {because now} 71. b6 {can be met by} Ka6) 71. b6+ Kb8+ 72.
Kb5 Rh8 73. a5 Rh1 74. a6 Rb1+ 75. Kc6 Ra1 76. d6 1-0
![[be]](/mwf/v12/flags/be.png "Belgium")
> Yes, the range of interesting 5-2 endgames is much smaller.
We could make the list :
Q vs all natural combinations with no queen
R vs 4p
R vs 3p+(R or B or N)
Minor vs 3p+minor (very few draws, I suppose)
Else others ?
Has knpppkn been completed and if so, what size is it? I had a position come up like that where I had to spend a while working it out to a mate in 15, but a 7-piece file on that would have given me the answer instantly.
Sorry. It seems that this ending has the least priority in our plans.
I can't promise something firm for it except that it will be certainly available not later than in 2013.
I can't promise something firm for it except that it will be certainly available not later than in 2013.
Could you provide the statistics, like no of Wins for all White to move positions, etc?
Ok, will be done today's evening for krpp-krp.
Looks like collecting statistics will take more time. It is single threaded process on Windows computer.
For now statistics is ready only for rank 2, i.e. both white pawns are on rank 2 (it is named krpp2krp). This slice contains less positions as compared to others.
The worst slice is krpp7krp (at least one pawn is on rank 7), other white pawn can be on any rank. This rank7-slice contains more positions than any other. Its statistics will be ready only tomorrow. When statisitcs will be ready I will attach the files for every slice to this message (and a bit later total statistics - it will be the sum of statistics in 6 files).
Fluent look at statistics shows that Draws field was not displayed - some silly bug. Output will be fixed for the next endings.
For now statistics is ready only for rank 2, i.e. both white pawns are on rank 2 (it is named krpp2krp). This slice contains less positions as compared to others.
The worst slice is krpp7krp (at least one pawn is on rank 7), other white pawn can be on any rank. This rank7-slice contains more positions than any other. Its statistics will be ready only tomorrow. When statisitcs will be ready I will attach the files for every slice to this message (and a bit later total statistics - it will be the sum of statistics in 6 files).
Fluent look at statistics shows that Draws field was not displayed - some silly bug. Output will be fixed for the next endings.
Attachment: krpp2krp.tbs (18k)
Attachment: krpp3krp.tbs (18k)
Attachment: krpp4krp.tbs (18k)
Attachment: krpp5krp.tbs (18k)
Attachment: krpp6krp.tbs (20k)
Attachment: krpp7krp.tbs (22k)
Attachment: krppkrp.tbs (23k)
How about the KRNPKRB position below, which is so far the deepest DTZ position for 7-man endings with pawns we have found so far:
4r3/1P6/6R1/5k2/1K5b/8/8/1N6 w - - 0 1
Mark, thanks for this position and the real test!
270 moves till conversion.
Only Mark & Yakov team knows such tricky positions.
[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "601 half-moves to mate"]
[Black "?"]
[Result "*"]
[FEN "
[SetUp "1"]
1. Rh6 $1 1... Re2 2. Rh5+ $1 2...
Bg5 3. Kb3 Re8 4. Kc4 $1 4... Kg6 5. Rh3 $1 5... Bf4 6. Rb3 $1 6... Re4+ 7.
Kc5 Re5+ 8. Kb6 Re6+ 9. Ka7 Bh2 10. Rh3 $1 10... Bg1+ 11. Kb8 $1 11... Be3
12. Rf3 Bd4 13. Rd3 $1 13... Be5+ 14. Ka7 $1 14... Bh2 15. Rd1 Bf4 16. Rg1+
Kh6 17. Rf1 Be3+ 18. Kb8 $1 18... Kg5 19. Kc8 $1 19... Rc6+ 20. Kd8 Rb6 21.
Kc7 $1 21... Rb3 22. Re1 Bf4+ 23. Kc6 $1 23... Kg6 24. Rg1+ Kh6 25. Rd1 Re3
26. Rc1 Re7 27. Rc3 Kg6 28. Kb6 Be3+ 29. Ka6 Re6+ 30. Ka5 $1 30... Re5+ 31.
Ka4 Ba7 32. Rc7 $1 32... Rf5 33. Rc6+ $1 33... Kg7 34. Nc3 $1 34... Rf7 35.
Rb6 $1 35... Rf6 36. Rb5 $1 36... Kf7 37. Ka5 $1 37... Ke8 38. Rd5 $1
38... Rf1 39. Na4 Bb8 40. Rd3 Rh1 41. Ka6 Ra1 42. Kb5 $1 42... Rh1 43. Nc5
Ba7 44. Rc3 Ke7 45. Nb3 Bb8 46. Rc8 Be5 47. Kc6 Rh6+ 48. Kc5 Rh3 49. Na5 Bd6+
50. Kc6 $1 50... Bf4 51. Rg8 $1 51... Rh6+ 52. Kc5 Rh5+ 53. Kb6 $1 53...
Be3+ 54. Ka6 $1 54... Rh6+ 55. Kb5 $1 55... Rb6+ 56. Kc4 $1 56... Bf4 57.
Kc5 $1 57... Be3+ 58. Kd5 Bf4 59. Rg4 Bb8 60. Rg7+ Kf6 61. Rg8 $1 61... Bc7
62. Rf8+ Kg7 63. Rf3 Bb8 64. Kc5 Ba7 65. Kc4 $1 65... Bb8 66. Rh3 Kg6 67. Rh8
Be5 68. Re8 Kf6 69. Kc5 Rb1 70. Kc6 Rc1+ 71. Kb5 Rb1+ 72. Ka6 Bc7 73. Nc4 $1
73... Kf7 74. Re4 Rb3 75. Re2 Kf6 76. Re8 Kf7 77. Rc8 Bf4 78. Na5 $1 78...
Rb1 79. Rc3 Ke6 80. Rd3 Bb8 81. Nb3 $1 81... Ke7 82. Nc5 $1 82... Bd6 83.
Rd5 Rb2 84. Na4 Rb1 85. Ka7 Ke6 86. Rd3 Ke7 87. Rc3 Rb5 88. Rc4 Kd7 89. Nb6+
Kd8 90. Ka6 Rb1 91. Rd4 Kc7 92. Ka7 Ra1+ 93. Na4 $1 93... Rb1 94. Rc4+ Kd8
95. Nc3 Rb2 96. Nd5 Rb1 97. Nb4 Bh2 98. Rd4+ Ke7 99. Rh4 Bg3 100. Rg4 Bh2 101.
Rg7+ Kf6 102. Rg2 Bd6 103. Nd5+ Kf7 104. Nc3 Rb4 105. Ka6 Ke6 106. Rg1 Bb8
107. Rd1 Rc4 108. Rd8 $1 108... Bh2 109. Rh8 Bc7 110. Rh6+ $1 110... Ke5
111. Rh5+ $1 111... Ke6 112. Nd5 Bg3 113. Rh3 Bb8 114. Rh8 Bg3 115. Nb6 Rc1
116. Re8+ $1 116... Kf7 117. Re4 $1 117... Ra1+ 118. Na4 Bb8 119. Rh4 Ke7
120. Rh8 $1 120... Be5 121. Rh7+ $1 121... Kf6 122. Rh4 Bg3 123. Re4 Kf7
124. Kb6 Rb1+ 125. Kc6 Rc1+ 126. Nc5 Bb8 127. Ra4 Ke7 128. Ra2 Kf7 129. Ra8
Bf4 130. Rh8 Bg3 131. Rh3 Bb8 132. Kd5 Rd1+ 133. Nd3 Ke7 134. Kc6 Rb1 135.
Rh7+ Kf6 136. Rh8 Bg3 137. Rg8 Bd6 138. Nf2 Ke7 139. Rg7+ Ke6 140. Nd3 Kf6
141. Rg8 Bh2 142. Nc5 Rc1 143. Kb6 Rb1+ 144. Ka7 Ke7 145. Rg7+ Kd6 146. Rg2
Be5 147. Na6 Kd7 148. Rd2+ Ke7 149. Re2 Kf6 150. Nc5 Bc7 151. Rc2 Ke7 152. Rc3
Bd6 153. Re3+ Kf8 154. Nb3 Rb2 155. Ka6 Bb8 156. Nd4 Rb1 157. Re4 Kf7 158. Nc6
Bh2 159. Nb4 Bd6 160. Nd5 Rb2 161. Re1 Rd2 162. Nb6 Bc7 163. Rc1 Bb8 164. Na4
Ke7 165. Rc8 Bf4 166. Nc5 Ra2+ 167. Kb6 Rb2+ 168. Kc6 Rc2 169. Rh8 Bg3 170.
Rh7+ Kf6 171. Rh1 Bb8 172. Kd5 Rd2+ 173. Kc4 Rh2 174. Rd1 Rc2+ 175. Kb4 Ba7
176. Ra1 Bb8 177. Ra8 Bg3 178. Ra3 Rb2+ 179. Kc4 Bd6 180. Ra7 Bh2 181. Ra6+
Kf7 182. Rh6 Bg3 183. Rh7+ Kg6 184. Re7 Kf6 185. Re3 Bb8 186. Re8 Bg3 187.
Rf8+ Ke7 188. Rg8 Bf4 189. Nd3 Kf7 190. Rh8 Rh2 191. Rc8 Bg3 192. Kb5 Rh5+
193. Kb6 Rh6+ 194. Ka7 Rh1 195. Rc3 Ra1+ 196. Kb6 $1 196... Rb1+ 197. Kc6 Bb8
198. Nc5 Rd1 199. Na6 Rd6+ 200. Kb5 $1 200... Ba7 201. Rc7+ Kf6 202. Rc6 Ke7
203. Nc5 Rd1 204. Rc7+ Kf6 205. Na4 Bb8 206. Rc8 Rb1+ 207. Ka6 Bh2 208. Rh8
Bc7 209. Rh6+ Kg7 210. Rc6 Bb8 211. Rc8 Bh2 212. Rc2 Bb8 213. Nc3 Rh1 214. Nb5
Kf7 215. Rc8 Bf4 216. Rd8 Ke7 217. Rd6 $1 217... Rh8 218. Rg6 Rh7 219. Kb6
Kd7 220. Nc3 Bc7+ 221. Ka7 Rh8 222. Ne4 Bb8+ 223. Kb6 $1 223... Ke7 224. Kc6
Rh4 225. Rg8 Rh6+ 226. Kd5 Bh2 227. Rg7+ Ke8 228. Nc5 Rh5+ 229. Kc6 Rh6+ 230.
Kb5 Kf8 231. Rd7 Ke8 232. Rd4 Bc7 233. Re4+ Kf7 234. Nd7 Rh5+ 235. Kc6 Bg3
236. Re3 Bh2 237. Re2 Bg3 238. Rg2 Rh6+ 239. Kb5 Bd6 240. Rg1 Rh2 241. Kc6 Rh6
242. Kd5 Bc7 243. Rc1 Rh5+ 244. Kc6 Bg3 245. Nc5 Rh6+ 246. Kb5 Ke7 247. Re1+
Kf7 248. Rd1 Bb8 249. Na6 Ba7 250. Rd7+ Ke8 251. Rc7 Rd6 252. Rg7 Kd8 253. Nb4
Rb6+ 254. Ka5 Re6 255. Nd5 Re1 256. Kb5 Rb1+ 257. Nb4 Ke8 258. Rh7 Bb8 259.
Kc6 Rc1+ 260. Kb6 Bd6 261. Kb5 Kd8 262. Rg7 Rc5+ 263. Kb6 Rc1 264. Nc6+ Ke8
265. Nd4 Be5 266. Nb5 $1 266... Kd8 267. Rg5 Bh2 268. Rg2 Bf4 269. Rd2+ Ke7
270. Rd6 Bxd6 271. Nxd6 {[%t bLon] Mate in 31:} 271... Kxd6 272. b8=Q+ $1
272... Kd5 273. Qd8+ Ke4 274. Qh4+ Kd3 275. Kb5 Rc2 276. Qg3+ Kd4 277. Qf4+
Kd3 278. Qf5+ Kd2 279. Kb4 Kd1 280. Kb3 Re2 281. Kc3 Ke1 282. Qg5 Rf2 283. Kd4
Rf3 284. Ke4 Kf2 285. Qh4+ Kg2 286. Qg4+ Rg3 287. Qe2+ Kg1 288. Kf4 Rh3 289.
Qd2 Kf1 290. Kg4 Rh8 291. Qc1+ Ke2 292. Qb2+ Kd3 293. Qxh8 Kc4 294. Qf6 Kd5
295. Qc3 Ke6 296. Qd4 Kf7 297. Qe5 Kg8 298. Qc7 Kh8 299. Kh5 Kg8 300. Kg6 Kh8
301. Qd8# *
270 moves till conversion.
Only Mark & Yakov team knows such tricky positions.
[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "601 half-moves to mate"]
[Black "?"]
[Result "*"]
[FEN "
4r3/1P6/6R1/5k2/1K5b/8/8/1N6 w - - 0 1"][SetUp "1"]
1. Rh6 $1 1... Re2 2. Rh5+ $1 2...
Bg5 3. Kb3 Re8 4. Kc4 $1 4... Kg6 5. Rh3 $1 5... Bf4 6. Rb3 $1 6... Re4+ 7.
Kc5 Re5+ 8. Kb6 Re6+ 9. Ka7 Bh2 10. Rh3 $1 10... Bg1+ 11. Kb8 $1 11... Be3
12. Rf3 Bd4 13. Rd3 $1 13... Be5+ 14. Ka7 $1 14... Bh2 15. Rd1 Bf4 16. Rg1+
Kh6 17. Rf1 Be3+ 18. Kb8 $1 18... Kg5 19. Kc8 $1 19... Rc6+ 20. Kd8 Rb6 21.
Kc7 $1 21... Rb3 22. Re1 Bf4+ 23. Kc6 $1 23... Kg6 24. Rg1+ Kh6 25. Rd1 Re3
26. Rc1 Re7 27. Rc3 Kg6 28. Kb6 Be3+ 29. Ka6 Re6+ 30. Ka5 $1 30... Re5+ 31.
Ka4 Ba7 32. Rc7 $1 32... Rf5 33. Rc6+ $1 33... Kg7 34. Nc3 $1 34... Rf7 35.
Rb6 $1 35... Rf6 36. Rb5 $1 36... Kf7 37. Ka5 $1 37... Ke8 38. Rd5 $1
38... Rf1 39. Na4 Bb8 40. Rd3 Rh1 41. Ka6 Ra1 42. Kb5 $1 42... Rh1 43. Nc5
Ba7 44. Rc3 Ke7 45. Nb3 Bb8 46. Rc8 Be5 47. Kc6 Rh6+ 48. Kc5 Rh3 49. Na5 Bd6+
50. Kc6 $1 50... Bf4 51. Rg8 $1 51... Rh6+ 52. Kc5 Rh5+ 53. Kb6 $1 53...
Be3+ 54. Ka6 $1 54... Rh6+ 55. Kb5 $1 55... Rb6+ 56. Kc4 $1 56... Bf4 57.
Kc5 $1 57... Be3+ 58. Kd5 Bf4 59. Rg4 Bb8 60. Rg7+ Kf6 61. Rg8 $1 61... Bc7
62. Rf8+ Kg7 63. Rf3 Bb8 64. Kc5 Ba7 65. Kc4 $1 65... Bb8 66. Rh3 Kg6 67. Rh8
Be5 68. Re8 Kf6 69. Kc5 Rb1 70. Kc6 Rc1+ 71. Kb5 Rb1+ 72. Ka6 Bc7 73. Nc4 $1
73... Kf7 74. Re4 Rb3 75. Re2 Kf6 76. Re8 Kf7 77. Rc8 Bf4 78. Na5 $1 78...
Rb1 79. Rc3 Ke6 80. Rd3 Bb8 81. Nb3 $1 81... Ke7 82. Nc5 $1 82... Bd6 83.
Rd5 Rb2 84. Na4 Rb1 85. Ka7 Ke6 86. Rd3 Ke7 87. Rc3 Rb5 88. Rc4 Kd7 89. Nb6+
Kd8 90. Ka6 Rb1 91. Rd4 Kc7 92. Ka7 Ra1+ 93. Na4 $1 93... Rb1 94. Rc4+ Kd8
95. Nc3 Rb2 96. Nd5 Rb1 97. Nb4 Bh2 98. Rd4+ Ke7 99. Rh4 Bg3 100. Rg4 Bh2 101.
Rg7+ Kf6 102. Rg2 Bd6 103. Nd5+ Kf7 104. Nc3 Rb4 105. Ka6 Ke6 106. Rg1 Bb8
107. Rd1 Rc4 108. Rd8 $1 108... Bh2 109. Rh8 Bc7 110. Rh6+ $1 110... Ke5
111. Rh5+ $1 111... Ke6 112. Nd5 Bg3 113. Rh3 Bb8 114. Rh8 Bg3 115. Nb6 Rc1
116. Re8+ $1 116... Kf7 117. Re4 $1 117... Ra1+ 118. Na4 Bb8 119. Rh4 Ke7
120. Rh8 $1 120... Be5 121. Rh7+ $1 121... Kf6 122. Rh4 Bg3 123. Re4 Kf7
124. Kb6 Rb1+ 125. Kc6 Rc1+ 126. Nc5 Bb8 127. Ra4 Ke7 128. Ra2 Kf7 129. Ra8
Bf4 130. Rh8 Bg3 131. Rh3 Bb8 132. Kd5 Rd1+ 133. Nd3 Ke7 134. Kc6 Rb1 135.
Rh7+ Kf6 136. Rh8 Bg3 137. Rg8 Bd6 138. Nf2 Ke7 139. Rg7+ Ke6 140. Nd3 Kf6
141. Rg8 Bh2 142. Nc5 Rc1 143. Kb6 Rb1+ 144. Ka7 Ke7 145. Rg7+ Kd6 146. Rg2
Be5 147. Na6 Kd7 148. Rd2+ Ke7 149. Re2 Kf6 150. Nc5 Bc7 151. Rc2 Ke7 152. Rc3
Bd6 153. Re3+ Kf8 154. Nb3 Rb2 155. Ka6 Bb8 156. Nd4 Rb1 157. Re4 Kf7 158. Nc6
Bh2 159. Nb4 Bd6 160. Nd5 Rb2 161. Re1 Rd2 162. Nb6 Bc7 163. Rc1 Bb8 164. Na4
Ke7 165. Rc8 Bf4 166. Nc5 Ra2+ 167. Kb6 Rb2+ 168. Kc6 Rc2 169. Rh8 Bg3 170.
Rh7+ Kf6 171. Rh1 Bb8 172. Kd5 Rd2+ 173. Kc4 Rh2 174. Rd1 Rc2+ 175. Kb4 Ba7
176. Ra1 Bb8 177. Ra8 Bg3 178. Ra3 Rb2+ 179. Kc4 Bd6 180. Ra7 Bh2 181. Ra6+
Kf7 182. Rh6 Bg3 183. Rh7+ Kg6 184. Re7 Kf6 185. Re3 Bb8 186. Re8 Bg3 187.
Rf8+ Ke7 188. Rg8 Bf4 189. Nd3 Kf7 190. Rh8 Rh2 191. Rc8 Bg3 192. Kb5 Rh5+
193. Kb6 Rh6+ 194. Ka7 Rh1 195. Rc3 Ra1+ 196. Kb6 $1 196... Rb1+ 197. Kc6 Bb8
198. Nc5 Rd1 199. Na6 Rd6+ 200. Kb5 $1 200... Ba7 201. Rc7+ Kf6 202. Rc6 Ke7
203. Nc5 Rd1 204. Rc7+ Kf6 205. Na4 Bb8 206. Rc8 Rb1+ 207. Ka6 Bh2 208. Rh8
Bc7 209. Rh6+ Kg7 210. Rc6 Bb8 211. Rc8 Bh2 212. Rc2 Bb8 213. Nc3 Rh1 214. Nb5
Kf7 215. Rc8 Bf4 216. Rd8 Ke7 217. Rd6 $1 217... Rh8 218. Rg6 Rh7 219. Kb6
Kd7 220. Nc3 Bc7+ 221. Ka7 Rh8 222. Ne4 Bb8+ 223. Kb6 $1 223... Ke7 224. Kc6
Rh4 225. Rg8 Rh6+ 226. Kd5 Bh2 227. Rg7+ Ke8 228. Nc5 Rh5+ 229. Kc6 Rh6+ 230.
Kb5 Kf8 231. Rd7 Ke8 232. Rd4 Bc7 233. Re4+ Kf7 234. Nd7 Rh5+ 235. Kc6 Bg3
236. Re3 Bh2 237. Re2 Bg3 238. Rg2 Rh6+ 239. Kb5 Bd6 240. Rg1 Rh2 241. Kc6 Rh6
242. Kd5 Bc7 243. Rc1 Rh5+ 244. Kc6 Bg3 245. Nc5 Rh6+ 246. Kb5 Ke7 247. Re1+
Kf7 248. Rd1 Bb8 249. Na6 Ba7 250. Rd7+ Ke8 251. Rc7 Rd6 252. Rg7 Kd8 253. Nb4
Rb6+ 254. Ka5 Re6 255. Nd5 Re1 256. Kb5 Rb1+ 257. Nb4 Ke8 258. Rh7 Bb8 259.
Kc6 Rc1+ 260. Kb6 Bd6 261. Kb5 Kd8 262. Rg7 Rc5+ 263. Kb6 Rc1 264. Nc6+ Ke8
265. Nd4 Be5 266. Nb5 $1 266... Kd8 267. Rg5 Bh2 268. Rg2 Bf4 269. Rd2+ Ke7
270. Rd6 Bxd6 271. Nxd6 {[%t bLon] Mate in 31:} 271... Kxd6 272. b8=Q+ $1
272... Kd5 273. Qd8+ Ke4 274. Qh4+ Kd3 275. Kb5 Rc2 276. Qg3+ Kd4 277. Qf4+
Kd3 278. Qf5+ Kd2 279. Kb4 Kd1 280. Kb3 Re2 281. Kc3 Ke1 282. Qg5 Rf2 283. Kd4
Rf3 284. Ke4 Kf2 285. Qh4+ Kg2 286. Qg4+ Rg3 287. Qe2+ Kg1 288. Kf4 Rh3 289.
Qd2 Kf1 290. Kg4 Rh8 291. Qc1+ Ke2 292. Qb2+ Kd3 293. Qxh8 Kc4 294. Qf6 Kd5
295. Qc3 Ke6 296. Qd4 Kf7 297. Qe5 Kg8 298. Qc7 Kh8 299. Kh5 Kg8 300. Kg6 Kh8
301. Qd8# *
What is your convention for assigning ! to a move? According to our database, 2.Rh5+, 4.Kc4, 6.Rb3, and a few others are also only moves in that they are the only moves that don't spoil the win.
Also, we think White can force a conversion in 260 moves, at the expense of possible a longer mate. With your pawn slice algorithm, you should be able to easily compute DTZ as well.
Also, we think White can force a conversion in 260 moves, at the expense of possible a longer mate. With your pawn slice algorithm, you should be able to easily compute DTZ as well.
Thanks, Mark for the careful consideration.
Really there is a problem with !. And there are two reasons for it.
1) The code used for output was not really designed for it. It was only for generaton, fortunately it mainly works ok for output. We are working for linking better code for output procedures that is used in Aquarium.
Some fixes were done for better output and I replaced the game.
2) If there is no some minor table this is also a problem. For now "!" is displayed although some additional moves exist in minor tables absent in the folder. So it is better not to trust all "!".
This will be the permanent problem due limit of disk space. Possibly we can produce bit-tables that help for !-output.
Really there is a problem with !. And there are two reasons for it.
1) The code used for output was not really designed for it. It was only for generaton, fortunately it mainly works ok for output. We are working for linking better code for output procedures that is used in Aquarium.
Some fixes were done for better output and I replaced the game.
2) If there is no some minor table this is also a problem. For now "!" is displayed although some additional moves exist in minor tables absent in the folder. So it is better not to trust all "!".
This will be the permanent problem due limit of disk space. Possibly we can produce bit-tables that help for !-output.
There is no a problem for us to generate DTZ. But we decided fully to solve more dificult problem - DTM. Solving two problems at one time demands too much efforts.
So your team can continue having glory of DTZ/DTC generation.
So your team can continue having glory of DTZ/DTC generation.
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill