On using chess engines in CC

Given that we all know that chess engines are forbidden by the rules of CC, I would like to ask a couple of hypothetical questions that have been on my mind recently.

-First scenario:
As you know all chess enginess today can do a deep evaluation and (amognst other things) output a score of the current position. Positive if white is ahead or negative for black. Now, assume that one does not look at the evaluated positions but only at the current position score given by the engine (e.g. as a graph plot). So for example, I am white and it is my turn to play and I ask the engine to give me an evaluation of the position. Say the engines tell me that the score is +2. Great!. I make my move (evaluated by myself) and continue. Then perhaps I ask the engine to re-evaluate the position after my move. It gives me -10. Crap!. That was a blunder for sure.

I can imagine that such a use might affect psychology and nothing else? I can think however of an example that might be considered unfair advantage. Say you are again white and down by a rook. You think that your position is not very good. You ask the engine for an evaluation and it gives you +2. Interesting. There must be a very good move that will get me from say -5 to +2. I should spend some more time finding it (and perhaps I do). Of course one should always assume there is a strong move (perhaps not of the above magnitude) and try to find it, but sometimes we do get sloppy.

What do you guys thing about this type of engine use? Hypothetically speaking.


-Second scenario:
Say I am decided to play openning XX. Before I even begin to play I ask from the engine to evaluate the best N-moves in 14-ply search. Just an exemplar number. Doesn't really matter. All the evaluated permutations are then saved in an openning database. Much more deep and dense than what is available online or in books for the same openning XX. Now, when I finally start my game, is it allowed to consult the created database?

Imagine for a second that before the game, during offline I studied the same openning XX. From books, online databases and without the help of the engine I studied some new lines from the opening. Now I guess there is nothing wrong with that. Even players in OTB chess do that. They even do analysis with an engine. I guess the difference is that with CC we can store the opening databases whereas with OTB you must memorise.

Originally posted by vzografos
Given that we all know that chess engines are forbidden by the rules of CC, I would like to ask a couple of hypothetical questions that have been on my mind recently.

-First scenario:
As you know all chess enginess today can do a deep evaluation and (amognst other things) output a score of the current position. Positive if white is ahead or negative fo ...[text shortened]... fference is that with CC we can store the opening databases whereas with OTB you must memorise.

Scenarios:

1. Cheating. You discover you have a better move by consulting a machine during the game. It gives you information based on the actual position, that you don't otherwise have. If it isn't cheating you could legally and blindly test every alternative move, until you hit on the +2.

2. I think this is OK if developed entirely in advance without knowledge of your opponent's actual moves. You could in principle be consulting a dusty old volume. The problem you will have is that the 'book' of best N moves for you as W (say) are predicated on some specific sequence of moves by B. Presumably the machine will optimize B's moves according to its algorithms, but B might have other ideas.

3. I think you've answered this.

Of course the arbiters of the competition may think differently than I do, and I am open to other ideas.

Hi and thanks for the reply. I am interested to what people have to say on this matter.

First of 2) and 3) are essentially the same thing. I was just giving my opinion on scenario 2 i.e. the database building. I guess ample computing power and storage makes database generation an easy task and one can consult them during game play. I guess the emphasis is on generating them BEFORE the game as in other way that would be considered cheating. Interesting if one thinks about it, whether in the future it would be possible to store many millions of permutations which would amount to almost like using the engine during gameplay (at least in the first N moves).
I think what makes it different from conventional openning books and online databases is the density one can achieve using and engine. It is probably possible to lead the game into an openning you have "researched" upon and any deviation by say Black from the lines would probably indicate a weakening of his/her position. Then it is up to you to exploit that.
This requires further discussion for sure


Scenario 1). Slight missunderstanding there. I did not mean try many moves while consulting the engine score after every move as this would be outright cheating in my opinion.

I was more thiking along the lines of:

Your turn to move. Consult engine on the position score. Get some number. Turn of engine. Play and commit your move (that you have thought). Your opponent makes his/her move. etc. Perhaps re-consult the engine afterwards.

So in a few words, use the engine only for positional evaluation BEFORE you make your move. Would that be considered cheating in your opinions? Or is it somewhere in the borders? I mean considered not from the rules or anything but from a hypothetical point of view.

Originally posted by vzografos


So in a few words, use the engine only for positional evaluation BEFORE you make your move. Would that be considered cheating in your opinions? Or is it somewhere in the borders? I mean considered not from the rules or anything but from a hypothetical point of view.

I guess the line is drawn as follows:

- before any game: you can do whatever you want --> this is preparation
- during any game: don't dare to ask a computer what he/she* would do or even think about the position. Think about it as if you're playing OTB, supplemented with all available material that existed before the position arose (e.g. opening theory and endgames).

*: is a computer male or female?

Originally posted by vzografos
Given that we all know that chess engines are forbidden by the rules of CC, I would like to ask a couple of hypothetical questions that have been on my mind recently.

Unfortunately this simply is not true. Engine use is banned by some CC organisations and sites but not all. ICCF and IECG (now defunct but replaced by LSS) do not ban engine use for example. Others ban but do not enforce for various reasons, while some (RHP is an example) attempt to enforce a ban to some extent.

Originally posted by vzografos
Scenario 1). Slight missunderstanding there. I did not mean try many moves while consulting the engine score after every move as this would be outright cheating in my opinion.

I was more thiking along the lines of:

Your turn to move. Consult engine on the position score. Get some number. Turn of engine. Play and commit your move (that you have thought). Your opponent makes his/her move. etc. Perhaps re-consult the engine afterwards.

I can see a practical difficulty here. Usually the minimum information given includes not only score for the position but some kind of assessment of what the engine "thinks" is the best move in that position. Either way, using an engine in that way in a game is cheating on this site.

scenario 1 is actually quite clear. you are using an engine, even if only to plot a line. it does affect your game even, as you said.

so, in that case, one is consulting an engine during a game. on rhp, that is forbidden.

Interesting replies all of them especially for the first scenario.

If I may I want to continue a little bit on scenario 1 (playing devil's advocate if you like) because I think it is an interesting, academic problem.

We all agree that a position evaluation score from an engine somehow affects the behaviour of the player (pshychologically or otherwise) that might have an impact in the course of the game. Not sure about a clear advantage but perhaps an impact nevertheless. For example a bad score might cause a player to persist longer and try harder to find a strong move, or alternatively it might demoralise the player leading to an early loss. And vice versa of course.


Ok, so suppose now that we have the following situation. Instead of an engine evaluating the position, we have a random number generator (perhaps skewed so that it gives a slightly more positive result). The player does not know it is a random number generator (and he/she is not that good at evaluating the position either 🙂 ), and uses that "score" to either try harder or quit early (see above).


So in other words, the random number generator would have a similar behavioural or psychological response. Would that constitute cheating?
And how is it different from the engine above?

I would be interested in hearing some opinions on that.

(Think of ithis, outside the domain of RHP and in the realms of philosophy)

well, one of them is a chess engine, the other is a random number generator. one is forbidden, the other not. thats the difference.

and 'philosophically' speaking: same thing. it is very simple and very clear: a chess engine during a game is forbidden.

the question of how much it affects a player, when he believes he uses a chess engine and it is in reality just a random number generator, is not very interesting actually. he will behave the same way and his play may or not improve depending on how he draws conclusions from a line plot...

Originally posted by vzografos
Interesting replies all of them especially for the first scenario.

If I may I want to continue a little bit on scenario 1 (playing devil's advocate if you like) because I think it is an interesting, academic problem.

We all agree that a position evaluation score from an engine somehow affects the behaviour of the player (pshychologically or otherwise nions on that.

(Think of ithis, outside the domain of RHP and in the realms of philosophy)

The RNG would have no bearing on the outcome of the game. The player might think a killer move was there but the information would be false. He'd likely lose all his timebank looking for it. And vice-versa. In fact now I think about it false information is likely to make him play worse.

hmmm....I am not really convienced that it is so easy to draw conclusions on how an engine score vs an RNG would affect the player.

For example, a false score indicating the existence of a good move might cause the player to examine the position more thoroughly, pay more attention and generally have a better game. The only thing that the player (thinks he/she) knows is that there is or there isn't a good move. He /she can never be sure if he /she found the correct move even when an engine gives the correct score.
Unless of course he/she asks from the engine or RNG for an evaluation after every move, which defeats the purpose of the exercise somewhat because it becomes easy to infer what happened.
I am talking about sparse sampling of the game with the score.


Anyway, it is more of a behavioural exercise. I wish I could actually test this and see how an RNG actually affects a player. Food for thought.

Originally posted by tharkesh
well, one of them is a chess engine, the other is a random number generator. one is forbidden, the other not. thats the difference.

and 'philosophically' speaking: same thing. it is very simple and very clear: a chess engine during a game is forbidden.

the question of how much it affects a player, when he believes he uses a chess engine and it is in r ...[text shortened]... way and his play may or not improve depending on how he draws conclusions from a line plot...

well, one of them is a chess engine, the other is a random number generator. one is forbidden, the other not. thats the difference.

and 'philosophically' speaking: same thing. it is very simple and very clear: a chess engine during a game is forbidden.

Yeah ok. We know that engine game is forbidden. But that's not the point. The purpose of this question is to examine if this engine use is at all beneficial when used in that way. It's nothing wrong with discussing a few things even if they involve rules and policies.

he will behave the same way and his play may or not improve depending on how he draws conclusions from a line plot...

So if I understood correctly what you said, from the point of view of the user an engine and an RNG are one and the same and their outcome on the game is purely subjective and user specific?
Interesting, because if that is the case the argument of not using an engine (in that scope) as a tool to gain unfair advantage might be somewhat unsupported by evidence. For example, I would argue that having access to a compendum of opennings and endgames might give more of an unfair advantage to a user than an engine (again, when used in such as way as described above).

Originally posted by vzografos

So if I understood correctly what you said, from the point of view of the user an engine and an RNG are one and the same and their outcome on the game is purely subjective and user specific?

no, i didn't explain it well enough then.

on the psychology of the player, both will have the same effect. obviously, since the player doesn't know the difference.

the games will be affected differently of course. once, the numbers will indeed help, and once it will just push the player into confusion, because there is no relationship to the actual position.

as you said, a line plot of the position evaluation will affect the game, because it is additional knowledge. it still is crystal clear, that this is not allowed on this rhp site, on others it is allowed for sure. the construction of your 'thought experiment' has not much value, because it can actually never be done for real - the player will either believe, he/she is violating the tos (rng or not) or will know, it is a rng and thus the experiment falls apart...

Scemaro One: Gold digger one is searching for gold. He has a metal detector that bleeps when it detects gold. Gold digger two is searching for gold. He has a spade. Who finds the gold?

Your analogy assumes you are using the engine in the following brute-force way:

Offline:
check Score - Try a move - check again - try a different move - check again - etc etc

Online:
Commit the best move from the above search.


That's obviously cheating
I simply stated:

Offline:
Check the score

Online:
Commit a move

If you want to make a golddigging analogy think of the following:

Both players have a spade
One of them has a device that beeps if there is gold within a 10km radius.