Ways to avoid "Draw Death" in Computer Chess

[Laskos]

One way to avoid "Draw Death" would be the appearance of one single engine that plays significantly stronger than the current top engines, say 100 or 200 Elo points. As of today this seems to be unlikely, but can we really exclude it? In the past there were already times of stagnation, then suddenly a heavy improvement came up when nobody thought it were possible.

I think today we are far away yet from knowing what the best openings really are, and there are also some endgame types that are not always evaluated correctly by top engines. Think about fortresses for instance, or about complex endgames with rooks, minor pieces and pawns for which we still lack any proven theoretical knowledge due to lack of EGTBs for more than 6 or 7 pieces. For me these are two good reasons for not believing that we are already close to "perfect play" in computer chess. It might be "perfect-looking play" only. All we know, in my opinion, is that we have a couple of very strong engines that are really hard to beat with today's chess programming knowledge and hardware.

Once that new mega engine appears (and as a programmer I hope this will happen one day, although it will probably not be my own engine ...) I expect that parts of these mathematical models may become outdated, or will have to be reviewed at least.

I am more of opinion of Greg, that Chess is too easy for computers. You list deficiencies of chess engines, which are true deficiencies, but the bulk is that they are extremely strong.

What does "extremely strong" really mean, other than "much stronger than almost all other (human or computer) chess players"? There is no absolute playing strength, only a comparison.

I took very balanced endgame positions (0cp-5cp unbalance) from real games and made the following experiments (no TBs):

Code: Select all

Score of Stockfish dev vs Zurichess_00: 297 - 1 - 702  [0.648] 1000
ELO difference: 106.01 +/- 10.81
Finished match

Stockfish dev is a recent version of Stockfish. Zurichess_00 is the first version of Zurichess I have, and it is about 1800 ELO human level (or pretty strong amateur). 30% of these ultra-balanced (endgame) openings are still playable against strong amateurs. Then I took the same Stockfish dev and Stockfish 7 separated by not that small 120 ELO points (on purpose not that small), and got:

Code: Select all

Score of Stockfish dev vs Stockfish 7: 8 - 4 - 988  [0.502] 1000
ELO difference: 1.39 +/- 2.35
Finished match

Now, only 1.2% of endgames are playable. So, 96% of endgames playable against strong amateurs are dead draws now.

These results are not surprising at all. But my explanation is different than yours: the strength difference between the two Stockfish versions for the "sub-discipline" of playing balanced chess endgame positions is obviously much smaller than their Elo rating difference for playing whole games, perhaps because of their endgame evaluations being more similar than their evaluation functions as a whole. You won't get a 99% draw outcome between these two SFs when playing 1000 normal games since the expected result is something about 660:340 so you should never get significantly more than 68% draws.

And it is not strength difference (a respectable 120 ELO points), but strength itself which gives more draws and less sensitivity.

Here I disagree. The draw rate between two players A and B must also be a function of the strength difference between A and B. See my explanation regarding endgames above.

The paradigm of Computer Chess is unchanged since Knuth of 40 years ago. The same alpha-beta, which reduces effective branching factor to 4-5, then ever more pruning and reductions to EBF 1.5 of today. According to this paradigm, the Computer Chess is capped to 400-800 more ELO points compared to today's top engines. The most comprehensive induction is from Andreas Strangmüller results here (and the discussion):
http://www.talkchess.com/forum/viewtopic.php?t=61784 ,
which points to lower values.

This "capping" theory seems to be based on the assumption that letting an engine search with "infinite speed", or "infinitely long", would (in theory) lead to solving chess by searching deep enough to get a perfect analysis of a given position, and therefore may serve as a justification for stating that "infinite speed" defines an upper bound for Elo ratings. However, I am not convinced of it. Today's heavy pruning and reductions often cause even a 50-ply search to be incomplete and imperfect. We only know that searching 50 plies deep is almost always better than searching only 40 plies and usually leads to a higher rating, but we still don't know whether the resulting analysis is "perfect". Now we could think of switching off all the pruning and reduction stuff, and repeat the test - but then I guess we will get very different results for the same test that was done by Andreas, i.e. there might be a much smaller effect of diminishing returns.

That a new, unchanged for 40 years, paradigm will appear, which will make a revolution in these conditions is almost as unlikely in the near future as a solution to Chess. It will be done, but in pretty far future. I believe in 10-20 years we will still be talking in the same terms and paradigm, but with 95%+ draw rate among top engines even in fast tests (with balanced positions). Sure, it might happen that an engine will appear which will be superior significantly, but this superiority diminishes _objectively_ (as separation power) with general strength, longer TC and more hardware.

Here I mostly agree, although nobody really knows what will happen during the next 10-20 years.

As a general remark, what I am really missing is a test where not time, speed, number of threads, or any similar resource is doubled subsequently but knowledge ... I hope someone can do that at some point in the future ?!

I disagree with you on many points. "Extremely Strong" has a clear meaning in my terminology and understanding. And that "absolute strength" is to blame for 98.8% draw rate in drawish endgames and not "similar endgame evaluation" or "similar paradigm". It might be the paradigm, but the paradigm works. ELO (logistic or otherwise) is not that abstract and volatile as it might seem to you. The difference between a random player and non-losing player from standard opening position seems to be capped at about 4000 logistic ELO points. Top engines already cut 3500-3600 ELO points of this 4000 ELO span. A strong human amateur or Zurichess_00 have it only at 1900 ELO points, they are closer to random player than to non-losing player. This is even more dramatic in "Endgame Chess". The ELO span there from random to non-losing is on only about 2000 ELO points from drawn position, and Stockfish is already at 1950 or so mark. It plays drawn endgames almost perfectly (in non-losing from drawn position sense). A strong amateur or Zurichess are at 1600 mark or so. In this sense my "Extremely Strong" has a meaning, it is extremely close to non-losing player from a drawn position. That's why the _absolute_strength_ gives that 98.8% draw rate from drawish endgame positions between Stockfish dev and Stockfish 7. There is no place for them to play other than a draw, and it's not "similar evaluation" and "similar paradigm". As for full game of Chess, if the capping to 400-500 more ELO points is real, it means that top engines play about 90% of all moves as non-losing moves from standard opening position, there are whole sequences of non-losing moves, and about 10% of games of top engines at LTC are perfect games in the sense of non-losing from standard opening position.

And returning to the toy "Endgame Chess", I have large files of 5-men Draws and 5-men Wins occurring in games. Against a Syzygy enabled, Stockfish misses 2 draws out of 1000, Zurichess 72 out of 1000. Out of 1000 Wins, Stockfish misses 87 Wins, Zurichess 524 Wins. (On a side note, to observe that it's much easier for engines to miss a Win than to miss a Draw). These results show too the dramatic absolute strength of Stockfish in drawn endgames and generally in endgames. These sorts of results will occur in normal Chess in maybe 20 years, with 95%+ draw rates of top engines from balanced positions even in fast games, keeping the same paradigm, which seems to work, and which seems hard to improve. The parallel with Checkers seems hard to avoid.

As for your proposal for "doubling knowledge", I don't know how to measure "knowledge". Its efficiency ELO-wise? Its size? Its scaling with time control? The current paradigm is that progress in "knowledge" does not come uniformly with progress in ELO. Rybka 1.0 had less knowledge than Shredder 9, but was significantly ahead as strength goes and a real progress. Andscacs might have more "knowledge" than Stockfish, but it is behind. It seems, though, the more knowledgeable engines are scaling better to longer TC and hardware.

[Dann Corbit]

I would use statistics.

Choose game collections with at least 1000 games at the pivot position (book exit).

Choose openings from these games where:

(white_wins + black_wins) / (white_wins + black_wins + draws) > desired_ratio

And yet white_wins/black_wins between desired_min and desired_max

There are definitely such things as fair decisive openings.

[Laskos]

What's to avoid? It's pretty clear that chess (FIDE rules) is a draw from the initial position with optimal play from both sides.
Unless you rig the opening by having one side play vastly suboptimal, a draw is the expected outcome. If you do unbalance the opening like that, you're not measuring a result, just confirming your input.

Or am I missing something?

The current state of top engines is that for a class of opening positions you can have 50% draws, even if from the standard opening position you can have 95% draws. That 50% draws is the "sweet point" of maximum power of separation between strength of engines (at least theoretically shown for Bayeselo and Davidson draw models). This means, the least amount of games (therefore resources) to separate engines in testing.

Or you want to say that Chess is anyway a draw from standard opening position, and let's abandon real-life engines?

[Laskos]

I would use statistics.

Choose game collections with at least 1000 games at the pivot position (book exit).

Choose openings from these games where:

(white_wins + black_wins) / (white_wins + black_wins + draws) > desired_ratio

And yet white_wins/black_wins between desired_min and desired_max

There are definitely such things as fair decisive openings.

There is a theoretical model for this in both Bayeselo (Rao-Kupper) and Davidson draw models. The "sweet spot" for very high draw rates (say eloDraw above 400) from balanced positions is to choose unbalanced opening with Draw rate 50%, Win rate close to 50%, Loss rate close to 0%. Or viceversa, Draw rate 50%, Win rate close to 0%, Loss rate close to 50%.

[Uri Blass]

One way to avoid "Draw Death" would be the appearance of one single engine that plays significantly stronger than the current top engines, say 100 or 200 Elo points. As of today this seems to be unlikely, but can we really exclude it? In the past there were already times of stagnation, then suddenly a heavy improvement came up when nobody thought it were possible.

I think today we are far away yet from knowing what the best openings really are, and there are also some endgame types that are not always evaluated correctly by top engines. Think about fortresses for instance, or about complex endgames with rooks, minor pieces and pawns for which we still lack any proven theoretical knowledge due to lack of EGTBs for more than 6 or 7 pieces. For me these are two good reasons for not believing that we are already close to "perfect play" in computer chess. It might be "perfect-looking play" only. All we know, in my opinion, is that we have a couple of very strong engines that are really hard to beat with today's chess programming knowledge and hardware.

Once that new mega engine appears (and as a programmer I hope this will happen one day, although it will probably not be my own engine ...) I expect that parts of these mathematical models may become outdated, or will have to be reviewed at least.

I am more of opinion of Greg, that Chess is too easy for computers. You list deficiencies of chess engines, which are true deficiencies, but the bulk is that they are extremely strong.

What does "extremely strong" really mean, other than "much stronger than almost all other (human or computer) chess players"? There is no absolute playing strength, only a comparison.

I took very balanced endgame positions (0cp-5cp unbalance) from real games and made the following experiments (no TBs):

Code: Select all

Score of Stockfish dev vs Zurichess_00: 297 - 1 - 702  [0.648] 1000
ELO difference: 106.01 +/- 10.81
Finished match

Stockfish dev is a recent version of Stockfish. Zurichess_00 is the first version of Zurichess I have, and it is about 1800 ELO human level (or pretty strong amateur). 30% of these ultra-balanced (endgame) openings are still playable against strong amateurs. Then I took the same Stockfish dev and Stockfish 7 separated by not that small 120 ELO points (on purpose not that small), and got:

Code: Select all

Score of Stockfish dev vs Stockfish 7: 8 - 4 - 988  [0.502] 1000
ELO difference: 1.39 +/- 2.35
Finished match

Now, only 1.2% of endgames are playable. So, 96% of endgames playable against strong amateurs are dead draws now.

These results are not surprising at all. But my explanation is different than yours: the strength difference between the two Stockfish versions for the "sub-discipline" of playing balanced chess endgame positions is obviously much smaller than their Elo rating difference for playing whole games, perhaps because of their endgame evaluations being more similar than their evaluation functions as a whole. You won't get a 99% draw outcome between these two SFs when playing 1000 normal games since the expected result is something about 660:340 so you should never get significantly more than 68% draws.

And it is not strength difference (a respectable 120 ELO points), but strength itself which gives more draws and less sensitivity.

Here I disagree. The draw rate between two players A and B must also be a function of the strength difference between A and B. See my explanation regarding endgames above.

The paradigm of Computer Chess is unchanged since Knuth of 40 years ago. The same alpha-beta, which reduces effective branching factor to 4-5, then ever more pruning and reductions to EBF 1.5 of today. According to this paradigm, the Computer Chess is capped to 400-800 more ELO points compared to today's top engines. The most comprehensive induction is from Andreas Strangmüller results here (and the discussion):
http://www.talkchess.com/forum/viewtopic.php?t=61784 ,
which points to lower values.

This "capping" theory seems to be based on the assumption that letting an engine search with "infinite speed", or "infinitely long", would (in theory) lead to solving chess by searching deep enough to get a perfect analysis of a given position, and therefore may serve as a justification for stating that "infinite speed" defines an upper bound for Elo ratings. However, I am not convinced of it. Today's heavy pruning and reductions often cause even a 50-ply search to be incomplete and imperfect. We only know that searching 50 plies deep is almost always better than searching only 40 plies and usually leads to a higher rating, but we still don't know whether the resulting analysis is "perfect". Now we could think of switching off all the pruning and reduction stuff, and repeat the test - but then I guess we will get very different results for the same test that was done by Andreas, i.e. there might be a much smaller effect of diminishing returns.

That a new, unchanged for 40 years, paradigm will appear, which will make a revolution in these conditions is almost as unlikely in the near future as a solution to Chess. It will be done, but in pretty far future. I believe in 10-20 years we will still be talking in the same terms and paradigm, but with 95%+ draw rate among top engines even in fast tests (with balanced positions). Sure, it might happen that an engine will appear which will be superior significantly, but this superiority diminishes _objectively_ (as separation power) with general strength, longer TC and more hardware.

Here I mostly agree, although nobody really knows what will happen during the next 10-20 years.

As a general remark, what I am really missing is a test where not time, speed, number of threads, or any similar resource is doubled subsequently but knowledge ... I hope someone can do that at some point in the future ?!

I disagree with you on many points. "Extremely Strong" has a clear meaning in my terminology and understanding. And that "absolute strength" is to blame for 98.8% draw rate in drawish endgames and not "similar endgame evaluation" or "similar paradigm". It might be the paradigm, but the paradigm works. ELO (logistic or otherwise) is not that abstract and volatile as it might seem to you. The difference between a random player and non-losing player from standard opening position seems to be capped at about 4000 logistic ELO points. Top engines already cut 3500-3600 ELO points of this 4000 ELO span. A strong human amateur or Zurichess_00 have it only at 1900 ELO points, they are closer to random player than to non-losing player. This is even more dramatic in "Endgame Chess". The ELO span there from random to non-losing is on only about 2000 ELO points from drawn position, and Stockfish is already at 1950 or so mark. It plays drawn endgames almost perfectly (in non-losing from drawn position sense). A strong amateur or Zurichess are at 1600 mark or so. In this sense my "Extremely Strong" has a meaning, it is extremely close to non-losing player from a drawn position. That's why the _absolute_strength_ gives that 98.8% draw rate from drawish endgame positions between Stockfish dev and Stockfish 7. There is no place for them to play other than a draw, and it's not "similar evaluation" and "similar paradigm". As for full game of Chess, if the capping to 400-500 more ELO points is real, it means that top engines play about 90% of all moves as non-losing moves from standard opening position, there are whole sequences of non-losing moves, and about 10% of games of top engines at LTC are perfect games in the sense of non-losing from standard opening position.

And returning to the toy "Endgame Chess", I have large files of 5-men Draws and 5-men Wins occurring in games. Against a Syzygy enabled, Stockfish misses 2 draws out of 1000, Zurichess 72 out of 1000. Out of 1000 Wins, Stockfish misses 87 Wins, Zurichess 524 Wins. (On a side note, to observe that it's much easier for engines to miss a Win than to miss a Draw). These results show too the dramatic absolute strength of Stockfish in drawn endgames and generally in endgames. These sorts of results will occur in normal Chess in maybe 20 years, with 95%+ draw rates of top engines from balanced positions even in fast games, keeping the same paradigm, which seems to work, and which seems hard to improve. The parallel with Checkers seems hard to avoid.

As for your proposal for "doubling knowledge", I don't know how to measure "knowledge". Its efficiency ELO-wise? Its size? Its scaling with time control? The current paradigm is that progress in "knowledge" does not come uniformly with progress in ELO. Rybka 1.0 had less knowledge than Shredder 9, but was significantly ahead as strength goes and a real progress. Andscacs might have more "knowledge" than Stockfish, but it is behind. It seems, though, the more knowledgeable engines are scaling better to longer TC and hardware.

I think that you overestimate the random player.
I think that a player with rating 1900 is closer to perfect player relative to random player.

I do not know how you get 0 elo for random player and it seems to me high.
Maybe it is because some weak engines allow stalemates but
I believe that if you take non buggy engines that do not allow stalemates and play them at fixed depths then you will get more than 3600 elo difference between depth 1 and depth 20 when depth 1 is clearly more than 400 elo better than the random player and I believe more than 800 elo better than the random player.

[Laskos]

I think that you overestimate the random player.
I think that a player with rating 1900 is closer to perfect player relative to random player.

I do not know how you get 0 elo for random player and it seems to me high.
Maybe it is because some weak engines allow stalemates but
I believe that if you take non buggy engines that do not allow stalemates and play them at fixed depths then you will get more than 3600 elo difference between depth 1 and depth 20 when depth 1 is clearly more than 400 elo better than the random player and I believe more than 800 elo better than the random player.

No, I tested pretty thoroughly the random player to be at about -100 to -200 CCRL 40/40 ELO points according to Logistic (which is pretty firmly established for engine-engine matches on large ELO span). Look at this thread:
http://talkchess.com/forum/viewtopic.ph ... =0&t=62510
There I have a table:

Code: Select all

   # PLAYER         : RATING    POINTS  PLAYED    (%) 
   1 Random 0%      : 2697.0     935.0    1000   93.5% 
   2 Random 10%     : 2229.8    1033.0    2000   51.6% 
   3 Random 20%     : 1632.3     970.0    2000   48.5% 
   4 Random 30%     : 1156.3     582.0    2000   29.1% 
   5 Random 40%     : 1142.2    1217.0    2000   60.9% 
   6 Random 50%     :  961.6    1148.0    2000   57.4% 
   7 Random 60%     :  604.0     820.5    2000   41.0% 
   8 Random 70%     :  450.9    1097.5    2000   54.9% 
   9 Random 80%     :  204.7     872.0    2000   43.6% 
  10 Random 90%     :   76.6    1115.0    2000   55.8% 
  11 Random 100%    : -155.6     210.0    1000   21.0%

given in CCRL 40/40 ELO points. So, in my reply to Sven, I took -100 to -200 for random player, 1700-1800 for strong amateur and Zurichess_00, and 3800-3900 for non-losing from standard opening position player. These are all supported by empirical data.

[Uri Blass]

I think that you overestimate the random player.
I think that a player with rating 1900 is closer to perfect player relative to random player.

I do not know how you get 0 elo for random player and it seems to me high.
Maybe it is because some weak engines allow stalemates but
I believe that if you take non buggy engines that do not allow stalemates and play them at fixed depths then you will get more than 3600 elo difference between depth 1 and depth 20 when depth 1 is clearly more than 400 elo better than the random player and I believe more than 800 elo better than the random player.

No, I tested pretty thoroughly the random player to be at about -100 to -200 CCRL 40/40 ELO points according to Logistic (which is pretty firmly established for engine-engine matches on large ELO span). Look at this thread:
http://talkchess.com/forum/viewtopic.ph ... =0&t=62510
There I have a table:

Code: Select all

   # PLAYER         : RATING    POINTS  PLAYED    (%) 
   1 Random 0%      : 2697.0     935.0    1000   93.5% 
   2 Random 10%     : 2229.8    1033.0    2000   51.6% 
   3 Random 20%     : 1632.3     970.0    2000   48.5% 
   4 Random 30%     : 1156.3     582.0    2000   29.1% 
   5 Random 40%     : 1142.2    1217.0    2000   60.9% 
   6 Random 50%     :  961.6    1148.0    2000   57.4% 
   7 Random 60%     :  604.0     820.5    2000   41.0% 
   8 Random 70%     :  450.9    1097.5    2000   54.9% 
   9 Random 80%     :  204.7     872.0    2000   43.6% 
  10 Random 90%     :   76.6    1115.0    2000   55.8% 
  11 Random 100%    : -155.6     210.0    1000   21.0%

given in CCRL 40/40 ELO points. So, in my reply to Sven, I took -100 to -200 for random player, 1700-1800 for strong amateur and Zurichess_00, and 3800-3900 for non-losing from standard opening position player. These are all supported by empirical data.

It may be interesting to test not only against random players but against normal engines or humans.

I cannot believe that random 20% can achieve fide rating of 1600 against humans.

It seems to me an engine that I guess that I can easily win against it
at blitz(5 minutes per game) and when I am clearly better than fide rating 1600 I believe my level at blitz is lower than 1600 fide rating(at tournament time control)

[cdani]

I cannot believe that random 20% can achieve fide rating of 1600 against humans.

It seems to me an engine that I guess that I can easily win against it
at blitz(5 minutes per game) and when I am clearly better than fide rating 1600 I believe my level at blitz is lower than 1600 fide rating(at tournament time control)

If your are curious, here it is the version of Andscacs used by Kai to test the randomness:
www.andscacs.com/andscacs_r087007.zip

http://talkchess.com/forum/viewtopic.ph ... acs+random

[Uri Blass]

won easily with white against 20%

going to try with black against 10%

[Event "Computer chess game"]
[Site "URIBLASS-THINK"]
[Date "2017.07.28"]
[Round "?"]
[White "UriBlass"]
[Black "Andscacs_r087007"]
[Result "1-0"]
[BlackElo "2000"]
[ECO "C60"]
[Opening "Spanish (Ruy Lopez)"]
[Time "14:20:52"]
[Variation "1.e4 e5"]
[WhiteElo "2400"]
[TimeControl "120+6"]
[Termination "normal"]
[PlyCount "83"]
[WhiteType "human"]
[BlackType "program"]

1. e4 e5 {(1. ... e5 2.Nf3 Nc6 3.Bb5 a6 4.Bxc6 dxc6 5.0-0 Bd6 6.d4 exd4
7.Qxd4 f6 8.e5 fxe5 9.Nxe5 Ne7 10.Nc3 Be6 11.Qe4 Bf5 12.Qe2 Bxe5 13.Qxe5
0-0 14.Re1 Ng6) -0.18/21 7} 2. Nf3 Nc6 {(2. ... Nc6 3.Bb5 a6 4.Ba4 Nf6
5.0-0 d6 6.Re1 b5 7.Bb3 Be7 8.Nc3 0-0 9.h3 Be6 10.d3 Nd4 11.Nxd4 exd4
12.Nd5 c5 13.Bf4 Nxd5 14.exd5 Bf5) -0.16/21 6} 3. Bb5 Qh4 {(3. ... a6 4.Ba4
Nf6 5.0-0 d6 6.Nc3 Bg4 7.h3 Bxf3 8.Qxf3 Be7 9.d3 0-0 10.Ne2 Nd7 11.Bb3 a5
12.Be3 a4 13.Bd5 a3 14.bxa3 Rxa3 15.Rab1) -0.19/23 9} 4. Nxh4 a6 {(4. ...
a6 5.Bxc6 dxc6 6.0-0 Nf6 7.d3 Be6 8.Nc3 0-0-0 9.h3 Nd7 10.Be3 f6 11.a3 g6
12.Qe2 Kb8 13.Nf3 h5 14.Nd2 f5 15.f4 exf4 16.Bxf4 Bc5+ 17.Kh2) -9.83/21 11}
5. Bxc6 dxc6 {(5. ... dxc6 6.0-0 Nf6 7.Nc3 Bg4 8.Nf3 Bd6 9.d3 Be6 10.Be3
0-0-0 11.Qe2 h6 12.h3 Kb8 13.a3 Nd7 14.Nd2 f6 15.Rfe1 Rhe8) -9.85/21 6} 6.
Nf3 Bd6 {(6. ... Bd6 7.d4 Bg4 8.dxe5 Bxe5 9.h3 Bxf3 10.Qxf3 Nf6 11.Qf5 Nd7
12.Nc3 Rd8 13.Be3 g6 14.Qf3 0-0 15.0-0-0 Bd6 16.Bd4 Rfe8 17.Kb1 Re6 18.Ne2
Rde8 19.Nc3 Ne5 20.Qf4) -10.19/23 15} 7. O-O Bf5 {(7. ... Ne7 8.d3 f6
9.Nbd2 0-0 10.Nc4 Ng6 11.Nxd6 cxd6 12.d4 Bg4 13.dxe5 dxe5 14.Qe2 Rfe8
15.Be3 b5 16.Rad1 Nf4 17.Bxf4 exf4 18.h3 Be6 19.b3 Bf7) -9.91/23 14} 8.
exf5 O-O-O {(8. ... 0-0-0 9.Ng5 Rd7 10.d3 Nf6 11.Nc3 Kb8 12.Nce4 Nxe4
13.Nxe4 f6 14.Be3 b6 15.Qg4 h5 16.Qg6 Kb7 17.f4 h4 18.a3 h3 19.g4 Rhd8)
-13.58/20 13 Black resigns} 9. Re1 f6 {(9. ... f6 10.Qe2 c5 11.Nc3 Ne7
12.Qe4 Nc6 13.b3 Kb8 14.Ba3 Ne7 15.Na4 b6 16.d3 Ka7 17.Bb2 g6 18.Nh4 g5
19.Nf3 h5 20.Nd2 b5 21.Nc3) -13.44/24 12 Black resigns} 10. Nc3 Ne7 {(10.
... Ne7 11.Nh4 Nd5 12.d3 Kb8 13.Ne4 Bb4 14.c3 Be7 15.Be3 Bd6 16.Qg4 Bf8
17.a4 a5 18.Qg3 b6 19.Rad1 Rg8 20.d4 Bd6 21.dxe5 Bxe5 22.f4 Bd6 23.Nxd6
cxd6) -13.39/21 12 Black resigns} 11. d4 Nxf5 {(11. ... Nxf5 12.dxe5 Bxe5
13.Bd2 Bxc3 14.bxc3 Rd7 15.Qe2 Rhd8 16.Qe4 Nd6 17.Qxh7 Nc4 18.Bf4 Nb6
19.Qf5 Nd5 20.Bd2 Kb8 21.Rab1 c5 22.c4 Nb6 23.Qxc5 Rxd2 24.Nxd2) -13.15/20
11 Black resigns} 12. dxe5 fxe5 {(12. ... fxe5 13.Bd2 Rhf8 14.Nxe5 g6
15.Nd3 Rd7 16.Ne4 Be7 17.Bf4 Bb4 18.c3 Be7 19.Qe2 Rd5 20.c4 Rd4 21.b3 Rfd8
22.Rad1 Kb8) -13.68/20 11 Black resigns} 13. Bg5 Rde8 {(13. ... Rdf8
14.Nxe5 Rhg8 15.Qd3 h6 16.Bd2 Bxe5 17.Rxe5 g6 18.Rd1 Kb8 19.Qe2 h5 20.Bg5
Nd6 21.Ne4 Nf7 22.Re7 Nxg5 23.Nxg5 Rd8 24.Ne4) -14.49/21 10 Black resigns}
14. Ne4 h6 {(14. ... h6 15.Nh4 Nxh4 16.Bxh4 Be7 17.Qg4+ Kb8 18.Bg3 Rhg8
19.Bxe5 Bb4 20.Bc3 Bxc3 21.Nxc3 Rd8 22.Rad1 Rc8 23.Re7 g6 24.Qe6 h5 25.Qd7
Ka8 26.Ne4 Kb8 27.Qe6 Rgd8) -14.97/21 10 Black resigns} 15. Be3 Bb4 {(15.
... Bb4 16.Bd2 Rd8 17.Nxe5 Bxd2 18.Nxd2 Rhf8 19.Qe2 Kb8 20.Nb3 Ka7 21.Rad1
Rd6 22.c3 Rfd8 23.Nc5 Rxd1 24.Rxd1 Rxd1+ 25.Qxd1 Nd6) -14.09/19 10 Black
resigns} 16. c3 Nxe3 {(16. ... Nxe3 17.Rxe3 Ba5 18.Rd3 g5 19.h4 Kb8 20.Rd7
Ka7 21.hxg5 hxg5 22.Nfxg5 Ref8 23.a4 Rf4 24.Rd8 Rhf8 25.Rxf8 Rxf8 26.b4 Bb6
27.a5 Bxf2+ 28.Nxf2 Rf4) -15.35/19 9 Black resigns} 17. Rxe3 Ba5 {(17. ...
Ba5 18.Rd3 Bb6 19.Nh4 a5 20.Ng6 Rhg8 21.Qg4+ Kb8 22.Rad1 Ka7 23.Rd7 Ka6
24.Qe2+ Ka7 25.Qh5 Ka8 26.Ne7 Rgf8 27.Qxe5 g5 28.Nf6 Rb8 29.Ng6 Rfc8 30.Qf5
Ka7) -15.77/22 9 Black resigns} 18. b4 Rd8 {(18. ... Bb6 19.Rd3 g5 20.Rd7
Rhf8 21.Qd3 Kb8 22.Rd1 Ka8 23.Rh7 Rf4 24.Rxh6 g4 25.Nfg5 Ref8 26.Ne6 Rxe4
27.Qxe4 Rxf2 28.Kh1 Rxa2 29.Qxg4 Ka7 30.h4) -15.08/22 9 Black resigns} 19.
bxa5 Rxd1+ {(19. ... Rxd1+ 20.Rxd1 Re8 21.Nc5 b6 22.Nd7 g5 23.axb6 cxb6
24.Nxb6+ Kc7 25.Nxe5 Rd8 26.Rxd8 Kxd8 27.Rd3+ Ke7 28.Nxc6+ Kf7 29.Nc4 Ke8
30.N6e5 h5 31.Rd6 a5 32.Re6+ Kf8 33.Nxa5 g4 34.Nac4 Kg7 35.a4) -16.65/21 8
Black resigns} 20. Rxd1 Re8 {(20. ... Re8 21.Nc5 b6 22.Nd7 e4 23.Nfe5 h5
24.axb6 cxb6 25.Nxb6+ Kc7 26.Rxe4 a5 27.Nbc4 Rd8 28.Rxd8 Kxd8 29.Nxc6+ Kd7
30.N6xa5 g6 31.a4 Kd8 32.Re5 Kd7 33.Rg5 Kc7 34.Rxg6) -17.02/21 8 Black
resigns} 21. Nc5 b6 {(21. ... b6 22.Nd7 e4 23.Nfe5 Kb7 24.Rb1 Re7 25.axb6
Rxe5 26.Nxe5 cxb6 27.Rxe4 c5 28.Rd1 Kc8 29.Rd7 g5 30.Rd6 h5 31.Rxb6 Kd8
32.Rxa6 Ke8 33.Nd3+ Kf7 34.Nxc5) -18.06/22 8 Black resigns} 22. axb6 cxb6
{(22. ... cxb6 23.Nd7 b5 24.Nfxe5 Kc7 25.Nc5 Rc8 26.Ne6+ Kb6 27.Nxg7 a5
28.Rd6 Rc7 29.Ne6 Rc8 30.Nd7+ Ka7 31.Nd4 b4 32.Nxc6+ Kb7 33.Nde5 bxc3
34.Rxc3 a4 35.Rd7+ Ka6 36.Ra7+ Kb5 37.Rc4) -16.68/20 8 Black resigns} 23.
Nd7 b5 {(23. ... b5 24.Nfxe5 Kc7 25.Nc5 Rc8 26.Ne6+ Kb6 27.Nxg7 Kb7 28.Rd6
h5 29.Nxh5 c5 30.h4 b4 31.g4 bxc3 32.Rxc3 c4 33.Nf6 Rh8 34.h5 Ka7 35.g5 Ka8
36.h6 Rxh6 37.Rxa6+ Kb7) -17.31/23 8 Black resigns} 24. Nfxe5 Kc7 {(24. ...
Kc7 25.Nc5 Rc8 26.Ne6+ Kb6 27.Nxg7 h5 28.Nxh5 c5 29.g4 a5 30.h4 c4 31.Rd6+
Kb7 32.g5 b4 33.Nf6 Rc5 34.g6 Rc7 35.Nd5 Rg7 36.Nf7 Rg8 37.Re7+ Kb8 38.Rb6+
Ka8 39.Ra6+ Kb8 40.cxb4) -17.97/21 7 Black resigns} 25. Nxc6 b4 {(25. ...
Rxe3 26.fxe3 b4 27.Nxb4 a5 28.Na6+ Kd8 29.Ne5+ Ke7 30.Nc6+ Kf6 31.Nc5 g5
32.Rd6+ Kg7 33.Nxa5 h5 34.a4 g4 35.Nc4 Kh7 36.a5 h4 37.Rd7+ Kg6 38.a6 g3)
-17.83/21 7 Black resigns} 26. Rxe8 bxc3 {(26. ... bxc3 27.Ne7 c2 28.Rc8+
Kb7 29.Rxc2 g6 30.Rb1+ Ka7 31.Nc6+ Ka8) -M6/27 7 Black resigns} 27. Rc1
Kxd7 {(27. ... Kxc6 28.Re7 g6 29.Rxc3+ Kd6 30.Rce3 a5 31.R3e6+ Kd5 32.Rxg6
Kc4 33.Rxh6 Kc3 34.Re5 a4 35.Rh4 a3 36.Rh3+ Kb2 37.Re2+ Ka1 38.Rxa3 Kb1
39.Kf1 Ka1 40.Ke1) -77.86/21 3 Black resigns} 28. Re3 Kc7 {(28. ... Kxc6
29.Rexc3+ Kd5 30.Rd1+ Ke5 31.Rc5+ Ke6 32.Rc6+ Ke5 33.Rd7 g5 34.Rxa6 g4
35.Rxh6 Kf5 36.Rh4 g3 37.fxg3 Kg5 38.Kf2 Kf6 39.Kf3 Ke6 40.Ra7 Kd5 41.Rh6)
-74.14/22 7 Black resigns} 29. Rcxc3 h5 {(29. ... h5 30.Ne5+ Kd6 31.Ng6 Kd5
32.Red3+ Ke6 33.Rc6+ Kf5 34.Rd5+ Ke4 35.Ne7 g5) -M7/23 7 Black resigns} 30.
Na5+ Kd6 {(30. ... Kd6 31.Rcd3+ Kc5 32.Re5+ Kb4 33.Rb3+ Ka4 34.Nc4 g6)
-M5/27 3 Black resigns} 31. Rcd3+ Kc5 {(31. ... Kc5 32.Re5+ Kb4 33.Rb3+ Ka4
34.Nc6 g5) -M4/91 3 Black resigns} 32. Re4 Kb6 {(32. ... Kb6 33.Re5 Kc7
34.Re7+ Kc8 35.Nc6 a5) -M4/35 7 Black resigns} 33. Rb4+ Kc5 {(33. ... Kc5
34.Rc4+ Kb6 35.a3 g6 36.Rb3+ Kxa5 37.Rbb4 g5) -M5/34 7 Black resigns} 34.
Rc4+ Kb6 {(34. ... Kb6 35.Rb3+ Kxa5 36.a3 h4 37.Rbb4 h3) -M4/78 3 Black
resigns} 35. Rb3+ Kxa5 {(35. ... Kxa5 36.a3 h4 37.Rbb4 h3) -M3/127 0 Black
resigns} 36. Kf1 h4 {(36. ... h4 37.h3 g5 38.Rc5+ Ka4 39.Rxg5 a5) -M4/122 6
Black resigns} 37. Ke2 h3 {(37. ... h3 38.Rc5+ Ka4 39.gxh3 g5 40.Rxg5 a5)
-M4/116 8 Black resigns} 38. gxh3 g5 {(38. ... g5 39.Rc5+ Ka4 40.Rxg5 a5)
-M3/127 2 Black resigns} 39. Rc1 g4 {(39. ... g4 40.Rc5+ Ka4 41.hxg4 a5)
-M3/125 0 Black resigns} 40. hxg4 Ka4 {(40. ... Ka4) -11.96/1 0 Black
resigns} 41. Rc5 a5 {(41. ... a5) -M1/1 0 Black resigns} 42. Rc4# 1-0

[Ozymandias]

I believe in 10-20 years we will still be talking in the same terms and paradigm, but with 95%+ draw rate among top engines even in fast tests (with balanced positions).

Does that prediction include deep and wide books, or just balanced positions at shallow depths (ply 4-16)?