Auto draw king-pawn fortresses

[johndoknjas]

Auto draw under these conditions:

• Variant is Standard, From Position, or Chess960.
• Only kings and pawns left.
• All the pawns are blocked by pawns.
• Both kings have no possible way to capture an enemy pawn, regardless of what the opponent does.

This kind of fortress isn't super common, but it does come up from time to time. E.g., I got it in two games from this month (rare):

• https://lichess.org/CxTaVjjI#87
• https://lichess.org/DbysyNO3#94

[lenguyenthanh]

hey, csv related notes:

• we can use a similar format of csv like this: https://github.com/lichess-org/scalachess/blob/master/test-kit/src/test/resources/horde_insufficient_material.csv. Basically, we can remove the double quote(for moves, we can just use space instead of comma to separate them) , to make it easier to parse.
• we have a tests related to csv: https://github.com/lichess-org/scalachess/blob/master/test-kit/src/test/scala/HordeInsufficientMaterialTest.scala. I think we should reuse some of that code here as well. I could help you if you're not familiar with the code.

[ornicar]

We don't throw exceptions in scala. If the function cannot compute its response type for every combination of its arguments, then it returns a Try[], an Either[] or an Option[].

[ornicar]

We almost never do procedural programming in scala. We almost never use var, while or return.
The only rare exception is for extreme performance optimization, after production profiling and benchmark testing.

[johndoknjas]

@lenguyenthanh Thanks for the feedback. I've addressed your first point, but for the second I've been having some trouble implementing it. If you'd be able to help here that would be appreciated.

[lenguyenthanh]

@lenguyenthanh Thanks for the feedback. I've addressed your first point, but for the second I've been having some trouble implementing it. If you'd be able to help here that would be appreciated.

I'll take a look 👍

[johndoknjas]

Hey @lenguyenthanh, just following up on this.

[lenguyenthanh]

Hey @lenguyenthanh, just following up on this.

I'm vacationing atm, will look in early February

[SergioGlorias]

I'm vacationing atm, will look in early February

@lenguyenthanh i think you forgot

[lenguyenthanh]

I'm vacationing atm, will look in early February

@lenguyenthanh i think you forgot

No, but I'm procastinating 🙏

[SergioGlorias]

• Only kings and pawns left.

maybe need include this

-- form https://lichess.org/forum/lichess-feedback/no-draw-3#2

[johndoknjas]

@SergioGlorias Should be pretty straightforward to handle other pieces. The bfs algorithm in core/src/main/scala/InsufficientMatingMaterial.scala can be applied to other pieces, just with the set of 'forbidden' squares being only the squares that friendly pawns are on (unless I'm missing some edge case). But to keep things simple maybe that should be tackled after/if this PR is merged.