# Elo disadvantage due to one pawn move in Indian Chess

In Indian chess, the pawn doesn't move two squares, however in the western chess, pawn can move two squares. Now, since the pawn can't move 2 in the Indian style, the fianchetto is a natural square to develop your bishops.

So for a player like Sultan Khan playing against other chess players from West, what kind of an Elo disadvantage would that put him? How does one estimate such a number?

Suppose one puts a constraint like this in Stockfish and lets it play against another Stockfish and observe how it does in a million games - would that be an acceptable way to estimate disadvantage?

• Sultan Khan did push pawns two squares Commented Jan 18, 2022 at 12:29
• When Indian chess has different rules for pawn movement, then it isn't really the same game. That makes ELO score not really comparable. That would be like comparing American Football players with Rugby players. Sure, they are similar and some skills are transferable, but they are still two different games. Which means people who practice both games would need a separate score for each. Commented Jan 18, 2022 at 13:38
• @Philipp I think the question is phrased a bit awkwardly, but is really "If a chess player self-imposed the handicap to never move a pawn two squares, how much ELO would that player lose?"
– Stef
Commented Jan 18, 2022 at 18:11
• ELO is a rock band. Elo is the name of the man who devised a rating system.
– bof
Commented Jan 19, 2022 at 2:03
• @stef well the thought came like that only - if sultan khan had always been playing indian version - how much of an opening disadvantage he have against western players. Commented Jan 19, 2022 at 3:25

Revised stockfish to skip searching double pawn push. Each engine is using 2 threads at TC 1m+1s. It got 4 draws in 20 games.

### Test 1

`sf_no_pawnpush4` is not aware that its opponent can play a double step pawn push and `sf_normal` is not aware that its opponent can only play a single step pawn push.

``````Score of sf_no_pawnpush4 vs sf_normal: 0 - 16 - 4 [0.100]
...      sf_no_pawnpush4 playing White: 0 - 7 - 3  [0.150] 10
...      sf_no_pawnpush4 playing Black: 0 - 9 - 1  [0.050] 10
...      White vs Black: 9 - 7 - 4  [0.550] 20
Elo difference: -381.7 +/- 252.4, LOS: 0.0 %, DrawRatio: 20.0 %
20 of 20 games finished.
``````

### Test 2

Each engine is aware of its opponent's capabilities.

``````Score of sf_no_pawnpush4_aware vs sf_normal_aware: 0 - 19 - 1 [0.025]
...      sf_no_pawnpush4_aware playing White: 0 - 10 - 0  [0.000] 10
...      sf_no_pawnpush4_aware playing Black: 0 - 9 - 1  [0.050] 10
...      White vs Black: 9 - 10 - 1  [0.475] 20
Elo difference: -636.4 +/- nan, LOS: 0.0 %, DrawRatio: 5.0 %
20 of 20 games finished.
``````
• Although this is perhaps one of the most direct way of calculating the difference, I don't think this is a very good estimate. Chess engines study openings, and limiting pawn movement simply changes the whole thing. The most accurate comparison would be training an engine with only 1 square pawn movement, and the same engine with 2 squares pawn movement, then compare the results. Commented Jan 18, 2022 at 7:40
• I imagine a Indian chess player would understand that the opponent may play e2-e4 even if they don't play it themselves. What happens if Stockfish-Indian is aware that the opponent may play double pawn moves? Commented Jan 18, 2022 at 12:37
• I interpret the question as "you can't move pawns 2 squares, but you know your opponent can". Not searching double pawn push is a significantly stronger handicap. Commented Jan 18, 2022 at 12:53
• @MateenUlhaq In addition, it may be added that Stockfish-Indian is aware that the opponent is aware that Stockfish-Indian is not able to play double pawn move. Commented Jan 18, 2022 at 13:17
• @Allure You're presumably right. What I wonder is whether the opponent (Stockfish pure) also has an asymmetric search i.e. knows that handicapped Stockfish cannot play double pawn pushes. Commented Jan 18, 2022 at 20:08