Computer Analysis: Nunn's Convention

Nunn's convention is a systematic way of using chess notations such as ?,?!,!?,!,!!. For example, it defines ! as the only move that keeps the evaluation of the position, ? as any moves that worsens the position.

Nunn's convention is a harsh critic, but also a very principled one. I wish lichess computer can use to it annotate my games one day. Presumably it shouldn't be too hard to implement, right?

Sounds like a terrible idea to me.

@lguoterry
We could have it here but since we are used to the ones that engine is programmed to, it would become a little difficult and take more time to settle with it. So, it's unlikely to have it at Lichess. What's the problem with current conventions?

@Akbar2thegreat said in #3:

@lguoterry
We could have it here but since we are used to the ones that engine is programmed to, it would become a little difficult and take more time to settle with it. So, it's unlikely to have it at Lichess. What's the problem with current conventions?

Perhaps we can add Nunn's convention as an alternative option. It's like Zen mode, in the setting where you have to set it.

@Akbar2thegreat said in #3:

@lguoterry
We could have it here but since we are used to the ones that engine is programmed to, it would become a little difficult and take more time to settle with it. So, it's unlikely to have it at Lichess. What's the problem with current conventions?

There isn't any problems with the current conventions. It is just that (1) the current convention rarely gives ! to moves, even if it's the only winning move (2) also the current convention gives ?! to moves that somewhat changes the evaluation of the position. Nunn's convention instead gives ?! to any moves that make the opponent's life easier, in other words leading to fewer winning lines and more losing lines. I feel that Nunn's ?! is more informative than the current's ?!.

Have you ever seen an exclamation mark (!) in the lichess analysis?

Your proposal just works in the realms of a tablebase. You have to know exactly to which result every move leads to. Exactly!

@lguoterry said in #1:

Nunn's convention is a systematic way of using chess notations such as ?,?!,!?,!,!!. For example, it defines ! as the only move that keeps the evaluation of the position, ? as any moves that worsens the position.

Nunn's convention is a harsh critic, but also a very principled one. I wish lichess computer can use to it annotate my games one day. Presumably it shouldn't be too hard to implement, right?

If the position is already equal and cannot be won, does it mean that every move after that will be counted as '!' ?

This means Nunn's convention, too.

https://lichess.org/study/LS1vNRkj/yPlJYRf5

@MrPushwood said in #2:

Sounds like a terrible idea to me.
Nunn convention was at least in his endgame book I tried to study (way too difficult to be useful) . And I agree that in almost solvable endgame it makes sense. In middle game evaluation w is winning is quite often non objective and even more often objectively won position might be practically very hard to win. So then Nunn annotation symbols make no sense ans that is why it only used in end game books

@passionate_player said in #7:

If the position is already equal and cannot be won, does it mean that every move after that will be counted as '!' ?

Nope the definition is:
! The only move which maintains the current evaluation of the position: If the position is theoretically drawn, this is the only move which does not lose; if the position is theoretically won, this is the only move which secures the win. An "!" is used no matter how trivial the move in question; the only exception is if it is the only legal move.

In educational material saves adding to commentary sentence "the only move to draw"