What will happen during Ding Liren's 13 games in the 2024 Tata Steel Tournament? weird promotions, pawn cube? MegaMarket
resolved May 23
Knight in a corner
Pawn Cube
Connect 5
Connect 6
Quadrupled Pawns
Center Pawn Cube
Castle with Check
Opposite side Castling game
10% for each win
10% for each draw

Will the things listed show up in any of the 14 games played by Ding Liren in the Tata steel 2024 tournament, excluding tiebreaks https://en.wikipedia.org/wiki/Tata_Steel_Chess_Tournament_2024 It runs Jan 12-28 2024

We include just Ding's games, but moves by either player are both counted.

Only the 1st 14 (or fewer) games, to make it easier to estimate. I wish we were guaranteed 14 total games; I'd rather run the risk of it ending early than prepare to do some difficult interpolation, so we'll just use as many of the non tiebreak 14 games as they actually play.

For the same reason we do not count tiebreak games. That would distort estimation too much until the exact time.

If he plays at least 9 games to a normal result, we will not NA. If he drops out before finishing 9 games, I will NA and we will be sad, but we will try again, since globally this type of dropout is very rare. If Ding tries to play but other people NA and Ding ends up not playing 9 games, we also will NA.

Rules and explanations for the in-game terms

  • By "points" I mean the traditional (~inaccurate) point heuristic: 1 point per pawn, 3 points per bishop or knight, 5 for a rook, 9 for a queen. Because of complexity, I do not currently use anything related to engine's "points of advantage estimation" in this market.

  • Bishop and knight means this endgame on board: b+k+possibly pawns vs n+k+possibly pawns, with at least one pawn

  • There is no "Ding POV"; so if the claim is "10% for each win" it does NOT mean "for each Ding win" - it means a win outcome for the game.

  • "Piece" means any chess piece. I don't use the meaning "a knight or a bishop" here, for simplicity

  • For questions which mention Pawns dying, they are considered to die when promoting

  • For questions which mention "survival" we do mean linear personal survival. i.e. if a queen is captured and then the player gains another via promotion, then the game ends, you would not say that the queen survived. It was captured.

  • Knights on the rim means a knight on the a or h file.

  • Strictness: if the claim is we get to a "mate with bishop and knight" it means we really need that on the board. If it's B+N+1 pawn+K vs K, and the person with B+N just has to take the pawn, but the opponent resigns, that's gotta be a NO since it didn't appear on board. This isn't ideal, but in general I'm attempting to avoid "implied" positions here to make judging possible.

  • I know it's very annoying, but if they shake hands on a draw 1 move before an official stalemate happens, that means the stalemate didn't happen. I wish we could resolve this "ends early" ambiguity but I don't see a way now. Open to suggestions.

  • "self-capture chess" aka "capture anything". https://chess.stackexchange.com/questions/28487/is-there-a-name-for-a-chess-variant-where-you-are-allowed-to-capture-your-own-pi Note that if the game mode switches, the defender can also capture anything to get out of check. So be careful when evaluating this.

This market is an early, test version of a similar megamarket for the 2024 world chess championships, which is here:


Get Ṁ600 play money

🏅 Top traders

#NameTotal profit
Sort by:

in Max Warmerdam-Ding Liren 2024.01.27 looks like the game ended with B +10 on board, so this is 100%

I'm making progress w/the code and lots of these have full results on the repo. Will gradually resolve them as they go. Thanks for your work submitting / trading.

22 short castles, 2 long castles by my count.

so if you just check a few of em you'll see it clearly resolves 100%

@Bayesian unfortunately it looks like most of these are much quicker to check manually than with code (some aren't though). I could tell you the results I found if you trust my observations but otherwise yeah it takes a while

Thanks for playing. Not all of the evaluators are done - here is the WIP code. I added a bunch today.

My next steps: figure out the right answers by either writing code to do it, or doing it manually :)


I'll try to get your mana back to you as soon as possible. Thanks!

Pawn Cube

does a pawn cube require the 4 pawns to be of 1 color or can they be either color? @Ernie

@Bayesian same question for connect 5

@Bayesian same for the center pawn cube

@Bayesian based on the rosen score, I believe these all require the same color.

I'm coding this one now, and realized it's a bit underspecified. If w has +1 material at one game (an extra pawn on board) and b has +1, does that count as 0 total, or 2? I think zero makes sense, since otherwise at scale this number would keep going up and up since it'd always be either zero or more.

In this view, the way it would be interpreted would be "over the games, will W end up having had a significant material advantage on-board?"

Are you all okay with this?

@Ernie I took it to mean the max of any one game, counted in the most naive possible way so I was thinking it would be based on game 12 where black has an “extra” queen.

@NateWatson I also took it to be this fwiw

@Ernie Same for “20% for each extra pawn held by the side with the greatest pawn advantage, in any position in any game”. I suggest saying “any one” game to make it clearer though.

This was so close to happening both in game 5 and game 12, but alas it never did.

Largest empty square in any final position seems to be 4x4, although you could fit three of them in the final position against Pragg

Game 7 against Wei Yi

7 pieces each, 14 in total

On the road... Will review latest comments and update the markets today

Since Ding draw the game in round 11 it is not possible anymore for him to have three decisive games in a row

Ah you're right - 14 players, 13 rounds. So I've been wrong about the number of games the whole time. Fixing that in the description, and you are right about your point also so resolving that too.

This one is computed after each game finishes (cumulatively) or at the end of tournament?

@Weezing I believe this means that if you think about game result points (0 for loss, 1/2 for draw, 1 for win) the B player will have ever been ahead at any point. Does that sound right? I think I need to rewrite it to clarify that these points are related to "tournament standing" points not to "points of material on the board."

@Ernie Yea I though it refererred to material. This resolves yes then, Ding won game 3 with 2 draws in previous games.

@Weezing okay, super busy today, I'll try to resolve tonight/tomorrow am

@Ernie no rush, thanks for being so responsive

@Weezing thanks for your patience, this is YES now.