World Cup Recap: Day 2

40 players were eliminated from the World Cup today, as the second and final classical games of round one were completed. With today’s results there have officially been eight first-round upsets booked, and more could occur tomorrow as the remaining 24 matches will be settled with rapid (and if necessary blitz and armegeddon) tie breaks. The eight players who were eliminated by lower seeds without any need for tie breaks were:

Seed Player
30  Ni Hua (CHN)
36  Igor Kovalenko (LAT)
39  Ivan Cheparinov (BUL)
41  Ray Robson (USA)
47  Gata Kamsky (USA)
51  Sanan Sjugirov (RUS)
55  Ivan Saric (CRO)
62  Evgeniy Najer (RUS)

These, of course, are mostly not particularly exciting upsets. Najer was knocked out by Rauf Mamedov (AZE), whom he only had a one point rating edge on before the event began. Ni Hua did have a 105 ELO edge on Sandro Mareco (ARG), in the biggest upset so far, and Kamsky is a big name, but only outrated Hrant Melkumyan (AZE) by 58 points. For those rooting for carnage, the top 29 seeds remain alive for now, so it hasn’t been a particularly satisfying first round yet. Eight of those top seeds are among the players who face tie breaks tomorrow, though, so there’s still chance of a shocker! We estimate a 38% chance that all eight will advance, cementing the #99 seed Mareco as the biggest upset of the event, but that means nearly a two in three chance of at least one triple digit seed emerging victorious tomorrow!

So what were the biggest surprises of the day? Well, first of all today’s results highlighted how this format (with two-game mini-matches) creates a difference between a surprising result in an individual game and a result that creates “surprise” by drastically shifting the odds of the eventual result.

For example: Ding Liren had the white pieces against Tomas Krnan today, 336 points his inferior by rating. In such a game, white should normally win 82% of the time, and Ding merely drew. Is that an upset? Normally we would be impressed by Krnan achieving the half point, but since Ding had already won yesterday this draw clinched his berth in round 2, and had approximately zero impact on our projected odds (Ding was considered a 99.95% favorite to advance prior to today’s game). Not an upset in this context.

The biggest upset in a decisive game was Milos Perunovic (SRB) knocking off Wang Hao, a result that should only happen 15% of the time for a 106 point ratings underdog, but this was only the 11th most impactful game from an odds perspective. Before the game Perunovic had a 3.4% chance of advancing to round two, because Wang had won the first game. Now they will go to tie breaks, but Wang is a strong favorite there. Perunovic’s odds improved, but he still has just a 25.7% chance of actually winning the round. His odds improved a solid 22 percentage points with the victory, but 10 other players saw their odds improve more.

Here are the top 10 gainers, in terms of how much today’s result improved their odds of reaching the second round in comparison to where they stood after yesterday’s game:

Player Seed R2 Odds Gain Opponent NEW Odds
 Alexander Ipatov (TUR) 90 62%  Ivan Cheparinov (BUL) 100.0%
 Maxim Matlakov (RUS) 40 41%  Gadir Guseinov (AZE) 64.9%
 Alexander Moiseenko (UKR) 32 40%  Lu Shanglei (CHN) 73.3%
 Sandro Mareco (ARG) 99 37%  Ni Hua (CHN) 100.0%
 Leinier Dominguez Perez (CUB) 18 35%  Federico Perez Ponsa (ARG) 85.8%
 Maxim Rodshtein (ISR) 42 33%  Eduardo Iturrizaga (VEN) 100.0%
 Viktor Laznicka (CZE) 50 33%  Varuzhan Akobian (USA) 100.0%
 Le Quang Liem (VIE) 37 33%  Vasif Durarbayli (AZE) 100.0%
 Julio Granda (PER) 52 33%  Alexandr Fier (BRA) 100.0%
 Wen Yang (CHN) 93 30%  Igor Kovalenko (LAT) 100.0%

Seven of the top ten gainers are favorites (top 64 seeds); overall today was a pretty good day for the favorites collectively. Before the tournament began the top 64 seeds had a combined 4877% chance of advancing to round 2, which is to say there was an implied expectation of 48.8 favorites reaching the second round. Adding the probabilities isn’t quite statistically valid, but remember we said the most likely number of upsets was 15, and this implies 15.2, so it’s pretty close. Game one was good news for the underdogs, collectively, reducing this number to 47.4 (the day one results implied an expectation of about 1 1/2 more round one upsets than we initially expected). However today’s game two results erased that edge. The favorites now can expect to advance in 48.9 spots, right about back to the original expectations, and a strong recovery from a slightly bad first day.

Among these players who gained the most ground today, the gains came in a variety of ways. Three of these players were favorites that had lost yesterday, and today’s win was necessary just to force a tie break, even if it wasn’t entirely shocking as a result. That would be Dominguez, Moiseenko, and Matlakov who so recovered. They aren’t entirely out of the woods yet, but they are now favorites in tomorrow’s tie break rounds. Five others, Ipatov, Rodshtein, Laznicka, Liem, and Granda had drawn their first games yesterday, but won today’s game to advance. Notice that Ipatov’s win was over a higher rated opponent, making it the most significant result of the day (his odds improved by 62 percentage points, to 100% now, from only 38% after yesterday’s game), while the other four won as favorites and only improved their chances 33 percentage points in clinching the round. Finally, the last two players, Mareco and Wen, had achieved their surprising upset wins yesterday and “merely” drew today, but that was enough to clinch a victory that was not yet certain. They were already favorites to advance to round 2, thanks to their 1-0 leads, but with today’s draws they became certain victors.

Here are the remaining 54 match ups, and how much ground was gained by the player for whom today’s result was more favorable:

Player Seed R2 Odds Gain Opponent NEW Odds
 Milos Perunovic (SRB) 96 22%  Wang Hao (CHN) 25.7%
 Sam Shankland (USA) 63 16%  Ivan Popov (RUS) 100.0%
 Michael Adams (ENG) 15 14%  Mariya Muzychuk (UKR) 100.0%
 Nguyen Ngoc Truong Son (VIE) 53 12%  Robert Kempinski (POL) 100.0%
 Hrant Melkumyan (ARM) 82 12%  Gata Kamsky (USA) 100.0%
 Radoslaw Wojtaszek (POL) 20 12%  M.R. Lalith Babu (IND) 100.0%
 Rauf Mamedov (AZE) 67 12%  Evgeniy Najer (RUS) 100.0%
 Yuri Vovk (UKR) 88 11%  Ray Robson (USA) 100.0%
 Bassem Amin (EGY) 74 10%  Ivan Saric (CRO) 100.0%
 S.P. Sethuraman (IND) 78 8%  Sanan Sjugirov (RUS) 100.0%
 Vladislav Artemiev (RUS) 45 8%  Surya Shekhar Ganguly (IND) 100.0%
 Andrei Volokitin (UKR) 75 6%  Alexander Onischuk (USA) 43.6%
 Alexander Areshchenko (UKR) 56 6%  Denis Khismatullin (RUS) 100.0%
 Eltaj Safarli (AZE) 69 6%  Csaba Balogh (HUN) 49.4%
 Sergei Zhigalko (BLR) 65 6%  Ivan Bukavshin (RUS) 49.7%
 Alexander Motylev (RUS) 61 6%  Boris Grachev (RUS) 50.0%
 Ilia Smirin (ISR) 48 6%  Romain Édouard (FRA) 100.0%
 Lázaro Bruzón Batista (CUB) 59 6%  Vidit Santosh Gujrathi (IND) 52.5%
 Liviu-Dieter Nisipeanu (GER) 49 6%  David Anton Guijarro (ESP) 100.0%
 Rafael Leitao (BRA) 83 6%  Hou Yifan (CHN) 38.8%
 Baskaran Adhiban (IND) 71 5%  Vladimir Fedoseev (RUS) 54.6%
Ernesto Inarkiev (RUS) 57 5%  Yuniesky Quezada Pérez (CUB) 55.2%
 Mateusz Bartel (POL) 85 5%  Gabriel Sargissian (ARM) 34.8%
 Igor Lysyj (RUS) 43 4%  Constantin Lupulescu (ROU) 62.8%
 Zhao Jun (CHN) 95 4%  Ian Nepomniachtchi (RUS) 28.2%
 Teimour Radjabov (AZE) 17 4%  Samuel Sevian (USA) 87.1%
 Boris Gelfand (ISR) 13 4%  Cristobal Villagra Henriquez  IM (CHI) 91.8%
 Anton Korobov (UKR) 38 4%  Dragan Solak (TUR) 100.0%
 David Navara (CZE) 25 3%  Tamir Nabaty (ISR) 81.2%
 Rustam Kasimdzhanov (UZB) 35 3%  Anton Kovalyov (CAN) 72.6%
 Alexander Grischuk (RUS) 7 3%  Yusup Atabayev IM (TKM) 97.6%
 Peter Leko (HUN) 29 3%  Aleksey Goganov (RUS) 100.0%
 Dmitry Andreikin (RUS) 27 3%  Zhou Jianchao (CHN) 100.0%
 Laurent Fressinet (FRA) 31 3%  Ante Brkic (CRO) 75.7%
 Vassily Ivanchuk (UKR) 23 2%  Ahmed Adly (EGY) 100.0%
 Pavel Eljanov (UKR) 26 2%  Rinat Jumabayev (KAZ) 100.0%
 Samvel Ter-Sahakyan (ARM) 101 2%  Nikita Vitiugov (RUS) 20.8%
 Wei Yi (CHN) 24 2%  Salem Saleh (UAE) 100.0%
 Yu Yangyi (CHN) 22 2%  Viorel Iordachescu (MDA) 100.0%
 Shakhriyar Mamedyarov (AZE) 19 2%  Pouya Idani (IRI) 100.0%
 Maxime Vachier-Lagrave (FRA) 21 1%  Isan Reynaldo Ortiz Suárez (CUB) 100.0%
 Anish Giri (NED) 4 1%  Arthur Ssegwanyi IM (UGA) 100.0%
 Peter Svidler (RUS) 16 1%  Emre Can (TUR) 100.0%
 Pendyala Harikrishna (IND) 14 0%  Max Illingworth  IM (AUS) 100.0%
 Evgeny Tomashevsky (RUS) 12 0%  Ziaur Rahman (BAN) 93.9%
 Sergey Karjakin (RUS) 11 0%  Ermes Espinoza Veloz FM (CUB) 100.0%
 Dmitry Jakovenko (RUS) 10 0%  Ilia Iljiushenok IM (RUS) 93.8%
 Wesley So (USA) 5 0%  Parham Maghsoodloo untitled (IRI) 100.0%
 Levon Aronian (ARM) 9 0%  Michael Wiedenkeller IM (LUX) 100.0%
 Ding Liren (CHN) 8 0%  Tomas Krnan IM (CAN) 100.0%
 Vladimir Kramnik (RUS) 6 0%  Deysi Cori T. WGM (PER) 100.0%
 Fabiano Caruana (USA) 3 0%  Amir Zaibi FM (TUN) 100.0%
 Hikaru Nakamura (USA) 2 0%  Richmond Phiri IM (ZAM) 100.0%
 Veselin Topalov (BUL) 1 0%  Oladapo Adu IM (NGR) 100.0%

Here we see how little really changed today. A few players who were in danger of being eliminated today instead forced tie breaks to their benefit, but only Perunovic gained significant ground in this way. A few others were strong favorites, but not locks, to advance (like Shankland), and did what they needed to do to clinch a berth in round two (anyone now showing a 100% chance in the far right column). That said, only 19 players have odds of reaching round 2 that are 10 or more percentage points better than yesterday’s odds. The other 45 gains were pretty small, all things considered. In other words: not too much changed today.

What about the chances of reaching the finals, and earning a berth in the 2016 Candidates Tournament? There was minimal movement there as well. Here are the nine players whose odds went up or down by at least half a percentage point:

Player Seed R2 Odds Gain Finals Odds Gain
 Veselin Topalov (BUL) 1 0% 1.2%
 Leinier Dominguez Perez (CUB) 18 35% 0.8%
 Anish Giri (NED) 4 1% 0.6%
 Fabiano Caruana (USA) 3 0% 0.5%
 Alexander Grischuk (RUS) 7 3% -0.5%
 Boris Gelfand (ISR) 13 4% -0.5%
 Dmitry Jakovenko (RUS) 10 0% -0.5%
 Wang Hao (CHN) 33 -22% -0.5%
 Ding Liren (CHN) 8 0% -1.2%

Two of these shifts, Dominguez and Wang, are for obvious reason. A large gain, or drop, in the odds of reaching round two (for someone rated high enough to be a valid contender in later rounds) will result in a commensurate gain or drop in the odds of reaching the finals. The other eight, though, are more subtle. They were already basically locks to reach the second round, so they didn’t make gains in that manner. Rather the odds of upsets elsewhere in their brackets shifted, making their likely paths through future rounds either easier or harder.

Ding’s second round opponent will be either Ernesto Inarkiev or Yuniesky Quezada Perez. Yesterday we thought Inarkiev (the higher rated player) had just a 49.8% chance of advancing (as Perez had white today, and the ratings gap is small enough that we expect Perez to win more often than lose, when he has the advantage of the white pieces). However they drew, and Inarkiev is now at 55.2% to advance to round two in the tie breaks. The ripple effect of this small shift is that Ding’s projected second round opponent is slightly better, making Ding’s odds of winning that match and reaching the THIRD round lower. Ding is now 83.1% to reach the third round, down from 83.8% after yesterday’s games. Also bad for Ding was Matlakov’s win today. Matlakov was upset in round 1, and yesterday we thought there was only a 10% chance he’d make it through to round 3 (where Ding would be his presumptive opponent). However as we saw above, he won today to force tie breaks, and was the second biggest odds gainer of the day, and now if Ding reaches round 3 he’ll have to face Matlakov 27.5% of the time. This means that not only does Ding face a potentially tougher second round, but also a potentially tougher third round. He is now a 60.4% favorite to reach round 4, down from 62.6% yesterday.

So it’s not through any fault of his own that Ding now has a 10.0% chance of reaching the finals, where it was at 11.2% yesterday. His path just looks that much harder now, as we project who his later opponents might be. It’s no coincidence that Topalov, who is in the same bracket as Ding, gained 1.2 percentage points in his chances. His path is easier because if he reaches the quarterfinals there’s only a 34.2% chance he’ll have to face Ding (his toughest potential quarterfinal opponent), down from 36.3%. He’s also slightly less likely to have to face either Giri or So in the semifinals, should he make it that far, which also improves his chances to reach the final.

What can we expect to see tomorrow? Here are all 24 matches where tie breaks proved necessary, and how likely the favorite is to win each:

Favorite Seed Underdog Favorite Odds
 Alexander Grischuk (RUS) 7  Yusup Atabayev IM (TKM) 97.6%
 Evgeny Tomashevsky (RUS) 12  Ziaur Rahman (BAN) 93.9%
 Dmitry Jakovenko (RUS) 10  Ilia Iljiushenok IM (RUS) 93.8%
 Boris Gelfand (ISR) 13  Cristobal Villagra Henriquez  IM (CHI) 91.8%
 Teimour Radjabov (AZE) 17  Samuel Sevian (USA) 87.1%
 Leinier Dominguez Perez (CUB) 18  Federico Perez Ponsa (ARG) 85.8%
 David Navara (CZE) 25  Tamir Nabaty (ISR) 81.2%
 Nikita Vitiugov (RUS) 28  Samvel Ter-Sahakyan (ARM) 79.2%
 Laurent Fressinet (FRA) 31  Ante Brkic (CRO) 75.7%
 Wang Hao (CHN) 33  Milos Perunovic (SRB) 74.3%
 Alexander Moiseenko (UKR) 32  Lu Shanglei (CHN) 73.3%
 Rustam Kasimdzhanov (UZB) 35  Anton Kovalyov (CAN) 72.6%
 Ian Nepomniachtchi (RUS) 34  Zhao Jun (CHN) 71.8%
 Gabriel Sargissian (ARM) 44  Mateusz Bartel (POL) 65.2%
 Maxim Matlakov (RUS) 40  Gadir Guseinov (AZE) 64.9%
 Igor Lysyj (RUS) 43  Constantin Lupulescu (ROU) 62.8%
 Hou Yifan (CHN) 46  Rafael Leitao (BRA) 61.2%
 Alexander Onischuk (USA) 54  Andrei Volokitin (UKR) 56.4%
Ernesto Inarkiev (RUS) 57  Yuniesky Quezada Pérez (CUB) 55.2%
 Baskaran Adhiban (IND) 71  Vladimir Fedoseev (RUS) 54.6%
 Lázaro Bruzón Batista (CUB) 59  Vidit Santosh Gujrathi (IND) 52.5%
 Csaba Balogh (HUN) 60  Eltaj Safarli (AZE) 50.6%
 Ivan Bukavshin (RUS) 64  Sergei Zhigalko (BLR) 50.3%
 Alexander Motylev (RUS) 61  Boris Grachev (RUS) 50.0%

Do note that in one case (Adhiban versus Fedoseev) the lower seed is the higher rated player, and is thus favored to win the tie break round. So far 8 top seeds have been eliminated. 24 more are in play tomorrow; how many will fall? On average, the answer is probably around 7 more, but we shall see what happens. Be sure to tune in and find out!


6 thoughts on “World Cup Recap: Day 2

  1. Lu Shanglei might actually have a better chance to advance than Moiseenko at this point if we take into account rapid and blitz ratings (which I’m aware, of course, that you decided not to include in your model for the sake of simplicity).


    • Based on blitz ratings maybe (Lu is 2780, Moiseenko is 2607). Rapid is less clear though, Lu is actually only 2565 rapid (but based on only 24 games, 9 in August 2012 and 15 more in July 2014, hardly enough to know his actual current rapid strength) while Moiseenko is 2677 rapid (and actually has a decent sample, though no games since January).

      Since Lu has such a small sample of rapid games, we could perhaps theorize that his blitz rating indicates he’s also stronger at rapid than the number indicates, but I’m not completely convinced that standard ratings aren’t a better indicator of rapid skill anyway.

      The first and second rounds of tie breaks are rapid, not blitz. It’s probably safe to say that Lu is a favorite if it reaches that point, but I suspect that Moiseenko is a pretty solid favorite in at least the first (longer) pair of rapid games. Are the odds exactly what our model shows? Probably not… but I think Moiseenko is probably at least a little bit more likely than Lu to advance.


      • Well, it’s a moot point now, since Lu Shanglei won, but according to your model, there was a slightly better than 1 in 4 chance for that to happen anyway, so he may have just gotten lucky. Either way, I’m just glad about the exciting chess.


      • Yeah, one challenge in dealing with probabilities is it’s hard to know if you were actually right about things. Lu advanced in rapid games, not in blitz. I still think that at G/25 time controls Moiseenko is probably stronger than Lu, but that the gap is probably smaller than in a game at classical time controls. Moiseenko was probably a smaller favorite than the model indicated, but I suspect still a favorite.

        On the other hand, there’s no way to be sure. You claimed Lu might be the favorite, I maintained that Moiseenko was still the favorite. Lu won. Does that mean you were right and I was wrong? Maybe… but I imagine we both agreed Gelfand was the favorite in his tie breaks, right? From your posts it’s clear you recognize this uncertainty as well, which I appreciate. This inability to be sure about anything, even after the fact, is why I enjoy trying to analyze probabilities so much 🙂


  2. In a way, calculations of tie break chances should be based on rapid ratings, instead of classical ratings. However, even if this change is made, I still don’t succumb fully to statistical permutations because it do not take into consideration the playing mode and motivation of a player.

    For instance, Grischuk is obviously again not in peak form but is 97.6% favourite tomorrow. The same argument goes to Toma, Jako, Gelfand, Radjabov and Dominguez → all of them have historical ‘on’ and ‘off’ peak modes. I am saying their statistical % are highly skewed (even if they manage to win tomorrow).

    The other argument (motivation) applies to the Top 4 seeds (Topalov, Naka, Anish and Fabiano). Whilst half of their mind is targeting the BIG, BIG $$$ pots, but the other half tells them to play easy because they have already qualified the Candidates (Topalov and Anish shall eventually qualify by ratings). Hence they may breeze their groupings but not necessarily win the Cup.

    IMO additional strong contenders to barge into the Finals are the ‘hungry’ and ‘wounded’ lions, e.g., Karjakin, Wesley, MVL, Ding and maybe Wei Yi → again not necessarily winning the Cup.

    Don’t get me wrong. Statistical numbers, at the end of the day, are good starting points when evaluating a player’s chance.


    • I agree with ChessNumbers here. “Form” is difficult to quantify statistically, and sometimes it’s not there at all. Keep in mind that humans have a tendency to see patterns in even random noise, and many examples of good or bad “form” in chess (for example, Aronian’s performance at Sinquefield) may be because, unsatisfying as it may sound, a particular player got lucky.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s