Computer and Human Evaluations

This forum is for general discussions and questions, including Collectors Corner and anything to do with Computer chess.

Moderators: Harvey Williamson, Steve B, Watchman

Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
Volodymyr
Member
Posts: 141
Joined: Sat Apr 08, 2017 1:03 pm
Location: Ukraine,Radyvyliv

Post by Volodymyr »

Nick,what insults? You do not play chess.
I played 5 times a month. Sometimes I do not play for half a year.

Continue farther. What's the difference for me,I do not pay money. A suspicious activity.
Very different manipulations and results are possible. Continue!

There could be no horror for me.This is 5 minutes blitz.
The horror is your collection and understanding of the game. Very big difference.

Already now all this is not interesting for me.
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Volodymyr wrote:Nick,what insults? You do not play chess.
I played 5 times a month. Sometimes I do not play for half a year.

Continue farther. What's the difference for me,I do not pay money. A suspicious activity.
Very different manipulations and results are possible. Continue!

There could be no horror for me.This is 5 minutes blitz.
The horror is your collection and understanding of the game. Very big difference.

Already now all this is not interesting for me.
Volodymyr,

I am frustrated with your comments because you keep challenging me about my chess ability which is something you just don't know. I don't know how old you are maybe you are just a young child, but I can tell you that regardless of how often you play, once you have played chess, it is like riding a bicycle, you do not forget how to ride it. You might wobble when you get started again but then after a little time you can ride the bicycle.

Also just because you do not play actively against humans anymore, you still enjoy those chess books you have and still in that same spare time study the game or even play against a chess computer.

I don't play humans because I am scared of the commitment. It is easy for me to play chess everyday in the way I do it because I can talk to my girlfriend, answer emails and problems on my work computer and watch tv or browse the internet all at the same time. In other words multitask. So I am multitasking, making work happy, girlfriend happy and myself happy as I am doing my hobby every day. You know it is not all about chess computers. I have studied chess books before I ever had my first chess computer or chess program. Over the years I have collected a quite large collection of chess books and played through them. You name the player and if he has a book and I tell you I have it and probably played through his games including a few Russian books. This is how the interest evolved for me.

Over the years I have learned to admire people who are not Worldchampions even more so than Worldchampions. And you might ask why? Well I tell you why, its because they have talents outside chess yet still managed to play at the same time at the highest level. Milan Vidmar for example, not only does he go down in history for being one of the top players of his time but he was also an engineer and inventor, committed to another life for which he also goes down in history. Philidor for operas etc etc so many examples of genius.

As I said before you don't need to tell me about how good or bad I am at chess because you don't know.

I learned to play when I was 7 or 8, by the time I was 12/13/14 I had already memorized chess books by heart played top board for my school against other regional schools, played top board for my chess club against other regional chess clubs and so on. Up to the age of 16 or 17 I was obsessed with chess, all night studies, memorizing everything I got my hands on, replaying adjourned games through your head through the night. Thinking back it was a terrible time really because my schoolwork suffered, football and some other sports that I was really good at back then suffered. I would turn up late to catch a team bus on a Saturday morning because I had overslept and get my ass kicked on Monday morning for not turning up. Especially if the team lost. It was so bad that I missed a regional match against another region in England in football one time. Got kicked off the team and did not get back on it again.

So what I am telling you is that chess to play seriously has a commitment and a unique obsession that I don't want to go through again ever.

I think fortunately for me it was girls that settled it, hormones kicked in better not to be a nerd and bookworm and life became good again :)

Only a couple of years ago I bought a chess computer from someone who lives locally from Ebay who figured out who I was from posts here. He talked me into going to the city's chess club with him which I did. Sure I was rusty playing humans again but as I said its like riding a bicycle. It took two or three visit before I was being asked to play for their team. And I knew this would happen once I started again. Well its a commitment that I don't want anymore so I stopped going the moment I was asked.

You can play online as much as you want its your enjoyment not mine. But it doesn't change what I know about chess and what you know about chess. As I told you before you know nothing about me and what I can do and can't do. And yes I do know pretty quickly what a good move is or a bad move and I don't need a computer to help me with that. You don't ever forget. You do become rusty and you do become older but you don't forget!

Since I seem to post a lot and play a lot of games it seems to provide an assumption that I have a lot of time for this. The truth is I do not have any more time for this than anyone else here. But I have taken up computer chess on as a hobby that keeps me involved most days.

What I have learned over the years through many many mistakes and obsessions is that you have to have a balance to make things work.

For me it is easy to relax a couple of hours at night multitasking a computer chess game, talking to girlfriend, with TV and Internet in background, work computer on desk to answer and stay on top of work emails and I relax (hobby) while I talk and work and keep up to date with news at the same time. :) That is how I do it. That is how I am able to talk soccer, football, news etc at the same time as chess. It's a multitasking skill that I have learned over the years that works for me.

So how do you expect me to manage all that successfully if I commit to a chess club or online games? Tell my girlfriend to shut up because I am busy playing someone? You try and tell her to shut up!! :)

Anyway my point is you don't know me, you don't know what I know about chess and you don't know how good I can play.

Therefore your persistent attacks on me on not knowing chess are annoying because you just don't know. In fact I find your conclusions naïve. You have not lived my life and neither have you lived my chess life.

But I know! and that is where our opinions differ!

PS. ask around here and you will be surprised at how many people here multitask in order to balance their home and work llife with their passion. It doesn't have to be chess, it can be some other interest. Now try and insult them by telling them that they don't know anything about their life interest and see how they react to your ignorance.

Best regards
Nick
Volodymyr
Member
Posts: 141
Joined: Sat Apr 08, 2017 1:03 pm
Location: Ukraine,Radyvyliv

Post by Volodymyr »

spacious_mind wrote: Volodymyr,

I am frustrated with your comments because you keep challenging me about my chess ability which is something you just don't know. I don't know how old you are maybe you are just a young child, but I can tell you that regardless of how often you play, once you have played chess, it is like riding a bicycle, you do not forget how to ride it. You might wobble when you get started again but then after a little time you can ride the bicycle.

Also just because you do not play actively against humans anymore, you still enjoy those chess books you have and still in that same spare time study the game or even play against a chess computer.

I don't play humans because I am scared of the commitment. It is easy for me to play chess everyday in the way I do it because I can talk to my girlfriend, answer emails and problems on my work computer and watch tv or browse the internet all at the same time. In other words multitask. So I am multitasking, making work happy, girlfriend happy and myself happy as I am doing my hobby every day. You know it is not all about chess computers. I have studied chess books before I ever had my first chess computer or chess program. Over the years I have collected a quite large collection of chess books and played through them. You name the player and if he has a book and I tell you I have it and probably played through his games including a few Russian books. This is how the interest evolved for me.

Over the years I have learned to admire people who are not Worldchampions even more so than Worldchampions. And you might ask why? Well I tell you why, its because they have talents outside chess yet still managed to play at the same time at the highest level. Milan Vidmar for example, not only does he go down in history for being one of the top players of his time but he was also an engineer and inventor, committed to another life for which he also goes down in history. Philidor for operas etc etc so many examples of genius.

As I said before you don't need to tell me about how good or bad I am at chess because you don't know.

I learned to play when I was 7 or 8, by the time I was 12/13/14 I had already memorized chess books by heart played top board for my school against other regional schools, played top board for my chess club against other regional chess clubs and so on. Up to the age of 16 or 17 I was obsessed with chess, all night studies, memorizing everything I got my hands on, replaying adjourned games through your head through the night. Thinking back it was a terrible time really because my schoolwork suffered, football and some other sports that I was really good at back then suffered. I would turn up late to catch a team bus on a Saturday morning because I had overslept and get my ass kicked on Monday morning for not turning up. Especially if the team lost. It was so bad that I missed a regional match against another region in England in football one time. Got kicked off the team and did not get back on it again.

So what I am telling you is that chess to play seriously has a commitment and a unique obsession that I don't want to go through again ever.

I think fortunately for me it was girls that settled it, hormones kicked in better not to be a nerd and bookworm and life became good again :)

Only a couple of years ago I bought a chess computer from someone who lives locally from Ebay who figured out who I was from posts here. He talked me into going to the city's chess club with him which I did. Sure I was rusty playing humans again but as I said its like riding a bicycle. It took two or three visit before I was being asked to play for their team. And I knew this would happen once I started again. Well its a commitment that I don't want anymore so I stopped going the moment I was asked.

You can play online as much as you want its your enjoyment not mine. But it doesn't change what I know about chess and what you know about chess. As I told you before you know nothing about me and what I can do and can't do. And yes I do know pretty quickly what a good move is or a bad move and I don't need a computer to help me with that. You don't ever forget. You do become rusty and you do become older but you don't forget!

Since I seem to post a lot and play a lot of games it seems to provide an assumption that I have a lot of time for this. The truth is I do not have any more time for this than anyone else here. But I have taken up computer chess on as a hobby that keeps me involved most days.

What I have learned over the years through many many mistakes and obsessions is that you have to have a balance to make things work.

For me it is easy to relax a couple of hours at night multitasking a computer chess game, talking to girlfriend, with TV and Internet in background, work computer on desk to answer and stay on top of work emails and I relax (hobby) while I talk and work and keep up to date with news at the same time. :) That is how I do it. That is how I am able to talk soccer, football, news etc at the same time as chess. It's a multitasking skill that I have learned over the years that works for me.

So how do you expect me to manage all that successfully if I commit to a chess club or online games? Tell my girlfriend to shut up because I am busy playing someone? You try and tell her to shut up!! :)

Anyway my point is you don't know me, you don't know what I know about chess and you don't know how good I can play.

Therefore your persistent attacks on me on not knowing chess are annoying because you just don't know. In fact I find your conclusions naпve. You have not lived my life and neither have you lived my chess life.

But I know! and that is where our opinions differ!

PS. ask around here and you will be surprised at how many people here multitask in order to balance their home and work llife with their passion. It doesn't have to be chess, it can be some other interest. Now try and insult them by telling them that they don't know anything about their life interest and see how they react to your ignorance.

Best regards
Well done Nick!
Why do I need this? I do not shoot TV shows.

Two games Level (1400ELO)(5 min/game) and level (2700ELO)(120 min/game). There are magic numbers - there is no difference. What discussions?
There is no discussion!
It's all!

In this topic I am no longer writing.

Thanks.
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Volodymyr wrote: My impression is you do not play chess. But you know the rules of the game of chess.:D
Volodymyr wrote: I repeat,Nick you do not play chess.:D
Sometimes a dubious move - is the way to victory.
This is the way to confuse the opponent.

Whoever does not play against a chessman will not understand this either.Then play online.
The strategy of the game against the computer and the chess player is different.
Volodymyr wrote:
Nick,I'm worried because I saw other of your messages.

There is no evolution of the brain - there can be no progress in the game of chess players.
Experience and memory are important. Now other possibilities.
One good book can give 100-150 Elo. If you master practically.
Base and analysis of games, online coach lessons, more than 1000 books on chess are available, chess engines for sparring, etc.
Nick wrote - no brain evolution.There is no evolution - no progress.


My rating is the most correct.
There are many questionable positions.
Where the archive with games?

Nick,this is your idea. Is this the evidence base?
There is no progress and my rating is the most correct one.

Nick, but you do not play chess.
What questions can there be? Or the correct ratings.

That's Nick. It's the Internet. :)
Volodymyr wrote:Nick,you found! But you watched the game, and looked at the analysis.
Your rating is the most correct!
But where is the archive with the games?
Volodymyr wrote:Nick,what insults? You do not play chess.
I played 5 times a month. Sometimes I do not play for half a year.

Continue farther. What's the difference for me,I do not pay money. A suspicious activity.
Very different manipulations and results are possible. Continue!
Volodymyr wrote:
Well done Nick!
Why do I need this? I do not shoot TV shows.
So let me understand this you don't think that you have barraged me with a continuous flow of insults. Every single post has one. How do you expect me to figure out what you are trying to say when every second word is an insult?

You accuse me of what?

1) I don't know chess
2) I don't know how to play chess.
3) Some quote you tell me I quoted. Please share the link so that we can all read it together
4) I do not play chess
5) Suspicous Activity? I do this for money?
6) Manipulations
7) TV Show actor?

What I have not answered is that all these games come from Chessbase. I have shared the results but I am not going to share the games. As I am not going to share the player names involved. The method I have shared in my first post. These games come from real over the table matches. German Bundesliga Games, Official Tournaments, French, Polish, English leagues and championships. Buy Chessbase and test it for yourself. These are not Online games and neither are they some 5 minute online games that you are trying to base your counter argument on.

You might not like my results, you might not like my approaches, you might not like my conclusions but I can assure you that you don't have to question my data source or accuracy.

If you really want to discuss your concern about a 5 min game scoring high which results in your eyes that all the other results are bad, then just ask yourself first how and why this could happen in a game?

1) Is it one of the 2% examples I showed you?
2) A game with very little complexity?
3) What else?

So if you want to test online games, then why don't you pull 100 of these games and see how they work out and lets compare official games from Chessbase to online games. Now that would also be an interesting exercise to compare real documented tournament games against online play.

Best regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Continuing with Human and Computer comparisons, I have recently spent a lot more time on this subject and the evaluations of human and computer games.

1400 ELO PLAYERS

Image

I analyzed a total of 1188 Chessbase Tournament games between the years of 1988 and 2013. I took all the games that Chessbase had between the ratings of 1376 ELO and 1425 ELO with games that had 35 or more White moves.

There were a total of 792 male evaluations and 278 female evaluations from official tournaments played in 27 countries.

I used Stockfish 18 PLY for the evaluations. Stockfish 18 Ply performs at over 3050 ELO and therefore is a pretty good tool to evaluate humans and dedicated computers.

The average ELO over the 1188 humans was 1401 ELO.

There were also included 118 games from a Blind Olympiad where every player had an arbitrary rating of 1400. I separated these out and the average score for these 118 games was 60.90 which is better than the average 1400 ELO male player who scored 66.33 ELO.

This Blind Olympiad Tournament was held in South Africa with representatives from Belgium, Brazil, Czech Republic, Denmark, Finland, Greece, Hungary, Italy, Kazakhstan, Moldavia, Montenegro, Netherlands, Norway, Slovakia, South Africa, Sweden, Switzerland, Turkey and Venezuela.

Blind Players from 19 countries were evaluated. These players of course must have been their best blind players in their respective countries.

The above table shows a complete breakdown of these 1188 games.

Super Grandmaster 2800+ ELO

Next, I decided to also evaluate all the games played between Super Grandmasters with ratings between 2776 ELO to max. Therefore I narrowed the search to all games between 2776 and 2900 ELO in order to find them all. I found a total of 274 games (548 evaluations) between the years 2008 and 2013. Therefore I evaluated them all. I omitted as usual any games where the White player played less than 35 moves.

Image

14 Grandmasters fell into this category where they played each other in games. The above table shows who these Grandmasters were, the amount of games each was evaluated on, and, their average rating over the amount of games they played. This was a total of 274 games and 548 evaluations. The average ELO for these games was 2799 ELO.

These Super Grandmaster games are split into 4 categories. Standard Tournament matches, Rapid Tournament matches, Blitz Tournament matches and Blindfold matches (which probably were mostly friendly exhibition games as almost all of them ended in a draw).

For standard tournament games these Super 2800 ELO Grandmasters averaged a score of 18.69 and 25.50 for Rapid Tournaments. Which means that these GM's play 36.4% better at Standard over Rapid. With the Blitz score of 40.17 these Super 2800 ELO Grandmasters play 115% better at Standard than Blitz. I would without having evaluated all the ELO categories yet, guess that these Super GM's play Blitz at the same level as Human players playing standard tournament chess between ELO 2300 to 2500.

Image

The above chart is another way of sorting and looking at the 2800 ELO Super Grandmaster statistics.

At Standard Tournament play, these Super Grandmaster play almost 80% of their games scoring 29.99 points or less. At Rapid Tournament play, 80% of their games score 39.99 points or less. At Blitz Tournament play, 77.6% of their games score 59.99 points or less. At Exhibition Blindfold games, 76.7% of their games score 49.99 points or less.

Performance Comparison between other ELO Reference Points

Image

I also spent some time on 250 evaluations between 1776 and 1825 ELO = 1801 average ELO. And also 250 evaluations between 1976 and 2025 ELO = 2001 ELO. Adult male players were only evaluated for both grade categories in order to keep the comparisons the same.

As you can see from the above chart 1800 ELO finished with an average score of 59.87 which is slightly better than the average score from the Blind International players that competed in the Blind Olympiad in South Africa in 2008.

2000 ELO finished with an average score of 50.70.

All these 1800 and 2000 games were played in 2013 and therefore it is quite an accurate comparison of how chess players play today and not throughout history. Since 250 games is still a little too few games, I will probably at some point continue to evaluate more until I get around a 1000 for each category. Same rule applied in that all the games were 35 White moves or more.

I think you have probably spotted by now that this was good time to introduce a dedicated chess computer into these comparisons. I used Novag Super Constellation as there are quite a few games available for both standard and rapid levels. Also through history, probably more was written about this dedicated computer than any other, having also received an Official USCF rating for official over the board Tournament play of ELO 2018.

In order to evaluate Novag Super Constellation and to keep the comparisons as equal as possible, I also used only games where 35 moves or higher were played by White. I also excluded all games where the opponent was 100+ ELO higher or lower rated than Novag Super Constellation. (Remember with humans I only used +/- 25 points, therefore a small bias still favoring humans). This resulted in 107 Standard games (2hrs/40) and 41 Rapid games (30 seconds per move average time).

Novag Super Constellation scored 51.83 for standard which was just slightly below the average 2000 ELO human player and 54.30 for rapid play (30 seconds per move average).

Comparing 1400 to 2000 ELO Players

Image

This is the same chart I used for the 2800+ Super Grandmasters. Here you can see that the average International Blind 2008 Olympiad player who played in South Africa is about the same strength as the average 1800 ELO male human player. Novag Super Constellation may not play as many almost perfect games as the average 2000 ELO human player with 5.6% compared to 10.8% but certainly Super Constellation seems to make less major mistakes than the average 2000 ELO human player with only 9.3% compared to 12.4% (80+ scores).

After completing the Novag Super Constellation evaluations, I started to wondering about what is right and what is wrong. How for example Larry Kaufmann a Grandmaster in his own esteemed rights can rate Novag Super Constellation at 2051 ELO with the CRA Rating list (Computer Ratings Agency) and the US Chess Federation provide Novag Super Constellation an official Tournament Rating against human players of 2018 ELO? Whereas computer rating lists show Novag Super Constellation as much lower nowadays.

Comparing different ratings for Novag Super Constellation

Image

Well I guess it all depends on how you look at it. If you want a rating on how chess computers compare against other chess computers, then both SSDF and Schachcomputer.Info's lists are perfectly adequate to show that.

If however you want to compare how a chess computer compares against humans, then it would be wise to read the old reports from Larry Kaufmann and Eric Hallsworth, USCF etc more seriously and really compare the ratings against humans.

For Novag Super Constellation it certainly seems that USCF was correct with their 2018 rating. Remember the above games I evaluated for 2000 ELO and 1800 ELO are all 2013 games and not games played 1986. Therefore nothing much has really changed between a 2000 player in 1986 that played Novag Super Constellation or what the evaluations showed for a 2013 human player.

Based on the 51.83 score for Novag Super Constellation it was quite easy to come up with a Novag Super Constellation rating of 1975 for standard play and 1921 for rapid play by taking the difference between 1800 ELO human and 2000 ELO human and calculating it out and converting it to the 2028 USCF in order to compare to the official 2018 USCF rating. The 1921 rapid rating is based on how it would perform playing against a 1975 player playing tournament level and is not a new rating for the computer. It is just an idea.

Obviously with more 1800 and 2000 games, as well as adding 1900 ELO games, the accuracy would considerably increase. Hopefully with time and patience, someday I will be able evaluate 1000 games under each category in order to obtain this further accuracy.

Therefore in summary, none of the ratings are wrong, it all depends on what you want to compare. Computers versus Computers or Humans versus Computers?

You can of course make a computer look bad by identifying its faults and weaknesses and constantly praying on those to prove a point in forum posts or a human superiority. But where does that lead to? It proves nothing, it devalues the true abilities of the computer and ultimately weakens the player who repeatedly plays the same way over and over again thereby never really improving himself or adding to his width of knowledge.

Here is a nice Larry Kaufmann quote:

"The way to beat your computer is to get to know it, and find its weaknesses. But if you want to learn from your machines, play to its strength instead, and get beaten."

In other words allow the computer to play its full opening repertoire, don't lead it over and over again into repetitive methods. Let it play and broaden your skills and knowledge at the same time. It's ok to get beaten! Enjoy it!

So anyway I really did find it surprising at how close these evaluation ended for Novag Super Constellation in comparison to its official 2018 USCF rating.

Yep, I know it is a lot to read and work through, but hopefully not too boring for everyone :)

Best regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

I don't normally show games that I play against chess computers, but I figured I would show one that I just played against Novag Constellation Forte. Last game I played was about 1 month ago while I was on a flight to Germany to visit family in North Germany who I have not seen in years. As it happened Michael (Blaubaer) was at that time also visiting family and was not too far away from where I was. So I drove up to meet him and we spent an afternoon and evening having dinner and talking a bit about chess and a bit about other stuff. Michael is a nice guy and I was glad that I had a chance to meet him.

Anyway on the plane you are able to play Kasparov Gambit on the little TV which you have in front of you on international flights. So it was fun to play a couple of games against this old program.

Since I returned from Vacation in early August I have been really too busy catching up with work and doing these human game evaluations in my spare time.

So finally after finishing my posts today I really did feel like a game of chess. Since I had just written about Novag Super Constellation, it seemed appropriate to play a game against a Novag computer. When I looked for one to play against I immediately saw Novag Constellation Forte staring at me. This apparently is a slightly improved version of Super Constellation with 5 Mhz instead of Super Constellation's 4 MHz. In the 1986 CRA report it was shown at 2091 USCF ELO. Schachcomputer.Info shows it as 1885 ELO which is the rating I am using. Forte is playing at level 3 which is 30 seconds per move average time. I tried to play at about the same time but tend to move a little quicker unless a game situation is very complicated. Complicated situations is something I try to really avoid if I can with chess computers.

[Event "?"]
[Site "?"]
[Date "2017.09.04"]
[Round "?"]
[White "Spacious Mind, 30s"]
[Black "Novag Constellation Forte, LV 3."]
[Result "1-0"]
[BlackElo "1885"]
[PlyCount "97"]

1. e4 e5 2. Nf3 Nc6 3. Bb5 d6 4. h3 {Novag Constellation Forte out of book}

[fen]r1bqkbnr/ppp2ppp/2np4/1B2p3/4P3/5N1P/PPPP1PP1/RNBQK2R w KQkq - 0 4[/fen]

I have not really practiced chess openings in 20 plus years therefore whatever I do nowadays tends to be from my failing old memory or what I remember from dedicated computer matches. So when it comes down to it, it is not very much really and I get thrown out of my memory book pretty easily as with example the move 3. ... d6 in this game when I had been expecting 3. ... a6 and forthcoming Spanish variations.

Anyway in a situation like this, I tend to retaliate immediately by playing a move that throws the chess computer also out of the book, thereby hopefully giving me some even chances in the opening. After I had played 4. h3 Forte was also out of book which of course puts me back into a place where I feel more comfortable.

Nf6 5. Nc3 Bd7 6. O-O Be7 7. d4 O-O 8. dxe5 Nxe5 9. Nxe5 dxe5 10. Bxd7 Qxd7 11. Qxd7 Nxd7 12. Nd5 Bd6 13. Bg5

[fen]r4rk1/pppn1ppp/3b4/3Np1B1/4P3/7P/PPP2PP1/R4RK1 w - - 0 13[/fen]

Playing 13. Bg5 seems kind of like a beginners move, but over the years I had learned two things that generally helped me to beat chess computers. One is to quickly take control of an open file and the second is to be seemingly aggressive with moves that make the dedicate computer feel uncomfortable where they will have the tendency to start moving pawns forward. This is a trick that typically helps me later during the endgame phase that I can exploit as I know that this is where I tend to be better than most dedicated chess computers and gives me the best chance to win the game. In other words bluff them from time to time. As expected Forte did not really like Bg5 and played f3 to move my Bishop back, which I did to e3.

f6 14. Be3 Rfd8 15. Rad1 Nb6 16. b3 Nxd5 17. Rxd5 c6 18. Rd3 Bc7 19. Rfd1 Rxd3 20. Rxd3 Bb6 21. Bxb6 axb6 22. a4 Rf8 23. Kf1 g6 24. Ke2 Kg7

[fen]5r2/1p4kp/1pp2pp1/4p3/P3P3/1P1R3P/2P1KPP1/8 w - - 0 25[/fen]

So by now I have simplified the game through exchanges and controlling the open file, centralizing the King, where I knew it would now be almost impossible for me to lose this game. The very least a draw and with the confidence that I have in endgames I felt good about my winning chances. Therefore to ensure that I can't lose this game, I decided to exchange the Rooks and make it a Kings and Pawns endgame.

25. Rd7+ Rf7 26. Rxf7+ Kxf7 27. Kd3 Ke6 28. Kc4 f5 29. f3 fxe4 30. fxe4

[fen]8/1p5p/1pp1k1p1/4p3/P1K1P3/1P5P/2P3P1/8 w - - 0 30[/fen]

Now this is the critical position. I had advanced my King to c4 threatening to breakthrough on the Queenside. It is in positions like this where dedicated computers tend to make a mistake as they often cannot search deeply enough to find the right alternatives. In the above position Forte sees an advanced white King and sees its own double pawns as a weakness for sure. Therefore it struggles to see that it only has one good move which is to counterattack with his King by moving to Kf6 followed by Kg5 threatening Kf4 and its own King pawn advances. In short after 30. Kf6 which was the only non losing move in the above position I probably would have had to retreat my King to d3, protect my Kingside and hope for later opportunities to launch something on the Queenside. Anyway Forte played 30. ... h6? and the win now became easy.

h6? 31. b4 Kd6 32. c3 Kc7 33. b5 Kd7 34. bxc6+ bxc6 35. Kb4 Kd6 36. c4 g5 37. g4 Kc7 38. c5 Kb7 39. cxb6 Kxb6 40. a5+

[fen]8/8/1kp4p/P3p1p1/1K2P1P1/7P/8/8 w - - 0 40[/fen]

Now its all over.

Ka7 41. Kc5 Ka6 42. Kxc6 Kxa5 43. Kd5 Kb6 44. Kxe5 Kc7 45. Kf6 Kd7 46. e5 Ke8 47. Ke6 h5 48. gxh5 Kf8 49. Kd7 1-0

Well that's it. The next time I post a game might be in 10 years or so :)

Oh btw... I also ran the game through Lichess. It looks like I did not do too bad :) But then again it is definitely easier to play at home and under no pressure :)

Spacious Mind, 30s
0 Inaccuracies
0 Mistakes
0 Blunders
23 Average centipawn loss

Novag Constellation Forte, LV 3. (1885)
2 Inaccuracies
4 Mistakes
2 Blunders
45 Average centipawn loss

Best regards
Nick
Reinfeld
Member
Posts: 486
Joined: Thu Feb 17, 2011 3:54 am
Location: Tacoma, WA

Post by Reinfeld »

Nice game, Nick!

- R.
"You have, let us say, a promising politician, a rising artist that you wish to destroy. Dagger or bomb are archaic and unreliable - but teach him, inoculate him with chess."
– H.G. Wells
User avatar
Dave C
Member
Posts: 460
Joined: Fri Jul 17, 2015 1:50 am
Location: SoCal, US

Lots to read but very interesting

Post by Dave C »

Hi Nick,

Excellent description of your analysis of computers and humans game strength.

And, well played against the Constellation Forte.

Thanks,
Dave
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

I just completed 250 evaluations between 1876 to 1925 ELO which averaged out exactly at 1900 ELO. All the evaluations are adult male players.

Image

1900 ELO came out somewhere in between 1800 and 2000 ELO as you can see above. Also it performed slightly better than Novag Super Constellation playing rapid chess (30 seconds per move average time).

Image

Since I now have this 1900 value, I recalculated Novag Super Constellation Rapid by taking the difference between 1800 and 1900 Human ELO scores.

I then did the same to recalculate Novag Super Constellation Standard (2hrs/40) by taking the difference between 1900 and 2000 ELO.

It is really uncanny that when I took this difference for Novag Super Constellation Standard, its rating came out exactly the same as its official USCF rating of 2018 ELO!

This seems to be too much of a coincidence therefore I am probably going to have to increase the 1800, 1900 and 2000 evaluations to 1000 instead of the current 250 sample size, and see where this comes out.

Amazed regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

I spent some more time on Human evaluations. Below is the performance summary list.

Performance Summary

Image

I have added 2100, 2200, 2300, 2400 and 2500 ELO. Using only 2013 players and games in sequential order with a search of +/- 25 ELO with a move range of 35 - 99 moves. 250 evaluations for each of the strength categories. I still have 2600 and 2700 ELO to complete and also below 1800 ELO.

All these evaluations seem to fit perfectly and exactly where they belong. Therefore 250 evaluations seems so be a pretty decent sample size.

I also added as you can see from the above list two more dedicated chess computers who both have sufficient Tournament Level (2hrs/40) games to allow for a good evaluation and comparison.

From the above chart you can see that Mephisto 68030 London (Genius London) finished with an evaluation between 2300 and 2400 ELO. Mephisto Nigel Short finished with an evaluation between 2100 and 2200 ELO.

Evaluated Dedicated Chess Computers at Tournament Level

Image

Since all the evaluated dedicated chess computer ratings all seem to be finishing higher than dedicated chess computer rating lists, I also went back and searched ratings from older publications to compare as well as shown in the above chart.

In all the examples the ratings shown in the past at Selective Search and at CCR match almost exactly to the evaluated ratings of the above 3 dedicated chess computers.

When I have time I will do about 2 or 3 more dedicated computer evaluations in order to confirm the above indications.

A Different Comparison

Image

The above is a different look which is based on total performance loss (mistakes). The lower the number the better the game performance. The dedicated chess computers that I have evaluated again show in the above chart that they fit in exactly where they should fit based on their game performance.

Calibrated Rating List based on Human and Dedicated Chess Computer Evaluations

Image

The above rating list I created as an approximate guideline based on the 3 completed dedicated chess computer evaluations (computers marked in brown). I used Schachcomputer.Info's tournament list and calculated it based on the proportionate difference between Mephisto 68030 London, Mephisto Nigel Short and Novag Super Constellation.

Based on the evaluations, I really do believe these computers in over the table games against humans would perform pretty closely to the same equivalent rated humans with both playing at Tournament Level. But then again that is just my opinion.

Anyway it has been a fun exercise.

Best regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Last Thursday night after having stuffed myself with Thanksgiving turkey, cakes and desert I had the urge to play a game against Millennium Chess Genius while my girlfriend was watching a movie on TV (was feeling bored with watching a movie on TV).

Well this was the first time that I had actually played against Millennium Chess Genius myself even though I have operated it many times in computer matches.

So as usual when playing an unknown good opponent for the first time I was full of anticipation and a little bit anxious with how I would do against it.

Playing time was 30 seconds per move rapid chess.

Here is the game.

https://lichess.org/63lhNWVH

[Event "Computer Test Match"]
[Site "Alabama"]
[Date "2017.11.24"]
[Round "1"]
[White "SpaciousMind, 30S."]
[Black "Millennium Chess Genius, 30S."]
[Result "*"]
[ECO "B19"]
[BlackElo "2224"]
[PlyCount "119"]
[EventDate "3017.11.24"]
[EventType "rapid"]

1. e4 c6 2. d4 d5 3. Nd2 dxe4 4. Nxe4 Bf5 5. Ng3 Bg6 6. Nf3 Nd7 7. h4 h6 8. h5 Bh7 9. Bd3 Bxd3 10. Qxd3 Ngf6 11. Bf4 e6 12. O-O-O Be7 13. Ne5 Nxe5 14. dxe5 Nd5 15. Bd2 Qc7 16. Qe4 Bc5 17. f4 O-O 18. f5 Rfe8 19. Rhf1 a5 20. c4 Nb4 21. a3 Na6 22. f6 Bf8

[fen]r3rbk1/1pq2pp1/n1p1pP1p/p3P2P/2P1Q3/P5N1/1P1B2P1/2KR1R2 w - - 0 23[/fen]

After 22 moves I had what I thought was a great position, but I started to second guess myself. For a long time I looked at playing 23. Qg4 but in the end I chickened out as after 23. ... Qxe5 I began to worry that my position might be destroyed with Black following Qxe5 with Red8 and developing a strong attack through the center while my position collapsed. Anyway I worried too much about MCG's game instead of focusing on what appeared to be a great position. Perhaps I ate too much turkey or I had too much respect for Millennium Chess Genius who is after all rated at 2218 ELO in Schachcomputer.info's active chess list. So anyway I played 23. Be3 instead and the game continued. Maybe its old age because I know in my younger years I would have just gone and risked it.

23. Be3 Red8 24. Rxd8 Rxd8 25. Bd4 Nc5 26. Bxc5 Bxc5 27. fxg7
Be7 28. Ne2 Bg5+ 29. Kb1 Kxg7 30. Nf4 Qd7 31. Nh3 Qd2 32. Nxg5 Qxg5 33. Qf3 Rd7 34. Qf6+

[fen]8/1p1r1pk1/2p1pQ1p/p3P1qP/2P5/P7/1P4P1/1K3R2 w - - 0 34[/fen]

Well. I had made a few more mistakes with the advantage going to Millennium Chess Genius when finally I decided my best chance was to try and force an exchange of Queens.

In the above position today's analysis shows that Black would retain an advantage with the move 34. ... Kg8. Black however played as I had hoped at the time 34. ... Qxf6.

Qxf6 35. exf6+ Kf8?

[fen]5k2/1p1r1p2/2p1pP1p/p6P/2P5/P7/1P4P1/1K3R2 w - - 0 36[/fen]

Kf8? was a further mistake Millennium Chess Genius played and in the above position I have a won game. Kg8 or Kh7 were better with decent chances for a draw. The key after the exchange of Queens is that white's Rook when it reaches d1 and d7 gives the advantage to White and exchanging the rooks results in a White win as black's King cannot defend white's queenside pawn advance while at the same time defend the g4 followed by g5 which results in a passed White h-pawn.

36. Kc2 c5 37. Rd1

[fen]5k2/1p1r1p2/4pP1p/p1p4P/2P5/P7/1PK3P1/3R4 w - - 0 37[/fen]

And there it is MCG's worst nightmare the Rook has reached d1 and game lost for Black.

Rd4 38. Rxd4 cxd4 39. Kd3 e5 40. b4 axb4 41. axb4 Ke8 42. g4

[fen]4k3/1p3p2/5P1p/4p2P/1PPp2P1/3K4/8/8 w - - 0 42[/fen]

Well Millenium Chess Genius realized it has lost as it now plays the not knowing what to do move 42. ... Kf8? Not that it mattered at this stage.

Kf8 43. c5 Ke8 44. b5 Kd7 45. g5 hxg5 46. h6 Ke6 47. h7 Kd5 48. c6 e4+ 49. Ke2 bxc6 50. b6 d3+ 51. Ke3 d2 52. Kxd2 Kd4 53. b7

[fen]8/5p1P/1Pp2P2/3k2p1/3pp3/8/4K3/8 w - - 0 50[/fen]

Two passed White pawns.

e3+ 54. Ke2 c5 55. b8=Q Kc3 56. h8=Q g4 57. Qhd8 c4 58. Qe5+ Kb3 59. Qb6+ Ka4 60. Qa1# 1-0

Final Position

[fen]8/5p2/1Q3P2/8/k1p3p1/4p3/4K3/Q7 w - - 0 60[/fen]

Lichess Analysis

SpaciousMind, 30S.
4 Inaccuracies
0 Mistakes
2 Blunders
24 Average centipawn loss

Millennium Chess Genius, 30S. (2224)
7 Inaccuracies
2 Mistakes
3 Blunders
40 Average centipawn loss

One less blunder than MCG! Well not exactly a grandmasterly win. But a win is a win! :)

Well not really. The game bugged me as I wanted to know what would have happened against MCG had I have played 23. Qg4. Therefore yesterday morning I got up at my usual time at 4.30 am, made a coffee and then grabbed MCG again. The nice thing about MCG is that when you turn it off, it keeps the game in memory. So when you turn it back on again you can even play back the whole game by using the back arrow key.

So anyway I back arrowed to the position 22. ... Bf8 and played 23. Qg4:

[fen]r3rbk1/1pq2pp1/n1p1pP1p/p3P2P/2P3Q1/P5N1/1P1B2P1/2KR1R2 w - - 0 23[/fen]

After playing 23. ... Qg4! and looking at the board from this position I realized duh! The threat that I was so worried about the night before 24. ... Qxe5 does not work for MCG because of the obvious White move 25. ... Rdf1! Worse still the move Bxh6! which was the whole point of me getting to this position in the first place absolutely works as well! Both would have been easy wins! :roll:

It's amazing what a night's sleep and a good cup of coffee in the morning does to your turkey stuffed mind from the night before!

Anyway MCG played the best move in the above lost position which was Kh8!

Kh8 24. Ne4 Red8? 25. Nd6 Rd7?

[fen]r4b1k/1pqr1pp1/n1pNpP1p/p3P2P/2P3Q1/P7/1P1B2P1/2KR1R2 w - - 0 26[/fen]

MCG was desperate its moves 24. ... Red8 and 25. ... Rd7 were both not that great, but it sets up for a nice finale! I played 26. Bxh6!

26. Bxh6 gxh6 27. Rd3 Rxd6 28. Rg3 Rd1+ 29. Rxd1 Kh7 30. Qg8# 1-0

Wow that would have been a checkmate in 30 moves against MCG!! :shock:

Final Position

[fen]r4bQ1/1pq2p1k/n1p1pP1p/p3P2P/2P5/P5R1/1P4P1/2KR4 w - - 0 30[/fen]

Lichess Analysis

https://lichess.org/swRMA2Ru

SpaciousMind, 30S.
2 Inaccuracies
0 Mistakes
2 Blunders
17 Average centipawn loss

Millennium Chess Genius, 30S. (2224)
4 Inaccuracies
1 Mistakes
2 Blunders
50 Average centipawn loss

So anyway what ended up a win is win against MCG might have been a crushing Grandmaster win against MCG :!:

So the moral of the story is, don't stuff yourself with Thanksgiving turkey before playing a game of chess!

Missed opportunity regards......
Nick
Post Reply