PCL Top Performers

|
0

We are 2/3 of the way through the season, and I took a hard look at the top performers in the league, thinking about which individual players were the most valuable so far this season.

The first stat that one could look at would either be points scored or performance rating. One tells you how much total one player has contributed to the team's score, the other gives a measure of how well they have played overall, accounting for strength of opposition, and eliminating frequency of play as a factor.

If you look at most points scored, you are looking squarely at Fabiano Caruana, with a two point lead over Daniel Naroditsky. But while amassing all those points, Fabiano has actually performed a bit below his rating, while some have performed above their ratings over fewer games. Looking at the second stat, performance rating, the league leader would be Anish Giri with 2839. However, with only 7 games played, it's hard to imagine that he has contributed the most of any player in the league this season.

So I wanted to look at a bunch of factors: points scored, performance rating, performance rating differential, how irreplaceable the player is, and more. I made a lot of errors along the way and there is a lot of guesswork, so I'll hone and improve my calculations. But this much can be said for them: they do get the #1 MVP right, undoubtedly. Every other result of turning these numbers around may be silly, but I think if you asked 32 PCL managers which one player they would want to move to their city and play as a local for their team, you'd get about 25 choosing Nika Volkov of the Tbilisi Gentlemen. (The Gentlemen themselves would choose Nakamura or Duda).

A quick word on what each category in my tables means, though I won't take the time to justify every decision I made. Perf Differential is just the player's performance rating minus the official rating the league counts towards the team's weekly 2500 rating cap. v #5 is a measure of how much better this player's Perf Differential is than the #5 player in their rating band. Perf Diff 27 attempts to adjust Perf Differential to recognize that it is far more valuable to have a 2500 who plays at 2750 than a 2200 who plays at 2450. It does so by weighing performance points over 2700 double, over 2650 *1.5, and all the rest normal. Probably over 2500 should be multipled by 1.25 or so, bc it is definitely more valuable than performances under 2500. By how much these should be weighted I don't know, but with my current numbers, several low-rated players with differentials of +200-300 show up near the top, whereas I think it is actually more important to have 2600 and 2700 performances on the team. W/ Frequency finally brings in the points scores/games played aspect, increasing the value of a player based on how many games they have played. Here I chose 4 per game, and the right number might be 3 or 5, or some more complicated formula altogether.

On to best of the best!

1st Category: Highest performers, no rating cap:

 Name Score Perf Official Rating Perf Differential v. #5 Perf Diff 27 w/ Frequency Duda 12.5/15 2837 2700 137 54 274 334 Naka 20/26 2812 2700 112 29 224 328 Iturrizaga 21.5/29 2742 2636 106 23 180 296 Xiong 18.5/26 2745 2651 94 11 163.5 267.5 Caruana 25/34 2752 2700 52 104 240 Wang Yue 8.5/11 2783 2700 83 0 166 210 Saric 16.5/20 2750 2689 61 116.5 196.5

2nd Category: Highest performers, u2650 official rating:

 Name Score Perf Official Rating Perf Differential v. #5 Perf Diff 27 w/ Frequency Shant 18/23 2715 2477 238 153 303 395 Grigoriants 13/18 2738 2543 195 110 283 355 Iturrizaga 21.5/29 2742 2636 106 21 180 296 Naroditksy 23/34 2692 2615 77 115.5 251.5 Sarana 19/27 2693 2634 59 88.5 196.5 Van Wely 7.5/13 2727 2642 85 0 141 193 Oparin 8 of 11 2698 2612 86 1 129 173 Durarbayli 9 of 12 2686 2629 57 85.5 133.5

3rd Catgory: Highest performers, u2550 official rating:

 Name Score Perf Official Rating Perf Differential v. #5 Perf Diff 27 w/Frequency Shant 18/23 2715 2477 238 74 303 395 Grigoriants 13/18 2738 2543 195 31 283 355 Dobrov 19.5/27 2677 2478 199 35 237.5 345.5 Giga 17.5/25 2659 2495 164 0 193.5 293.5 Sadhwani 13/19 2623 2438 185 21 196.5 272.5 Bezold 10.5/15 2650 2487 163 188 248 Gorovets 9 of 13 2630 2512 118 133 185

4th Category: Highest performers, u2450 official rating:

 Name Score Perf Official Rating Perf Differential v. #5 Perf Diff 27 w/ Frequency Nika 20.5/34 2499 2115 384 272 384 520 Grant Xu 10.5/18 2501 2207 294 182 294 366 Tianqi 11.5/19 2558 2312 246 134 246 322 Sadhwani 13/19 2623 2438 185 73 185 261 AADiaz 13/23 2489 2392 97 97 189 Vrolijk 10.5/18 2547 2435 112 0 112 184 Camallonga 14/25 2529 2447 82 82 182 Sheng 12/22 2509 2447 62 62 150

5th Category: Highest performers, u2350 official rating:

 Name Score Perf Official Rating Perf Differential v. #5 Perf Diff 27 w/ Frequency Nika 20.5/34 2499 2115 384 138 384 520 Manu David 11/18 2348 2044 304 58 304 376 Grant Xu 10.5/18 2501 2207 294 48 294 366 Tianqi 11.5/19 2558 2312 246 0 246 322 Manukyan 7.5/11 2468 2196 272 26 272 316 ZhuJiner 13.5/23 2474 2264 210 210 302 Song 19.5/30 2467 2305 162 162 282 AnnaS 10.5/22 2410 2231 179 179 267 Aaron Jacobson 7.5/15 2388 2255 133 133 193 Bordi 5.5/11 2314 2288 26 26 70

6th Category: Highest performers, u2250 official rating:

 Name Score Perf Official Rating Perf Differential v. #5 Perf Diff 27 w/ Frequency Nika 20.5/34 2499 2115 384 112 384 520 Manu David 11/18 2348 2044 304 32 304 376 Grant Xu 10.5/18 2501 2207 294 22 294 366 Ravuri 11/30 2295 2050 245 245 365 Bloomer 8.5/26 2222 2000 222 222 326 Grabinski 3/12 2274 2000 274 2 274 322 Manukyan 7.5/11 2468 2196 272 0 272 316 AnnaS 10.5/22 2410 2231 179 179 267 Inna 5/20 2292 2183 109 109 189

Overall Most Valuable Players in the League

 Name Score Perf Official Rating Perf Differential v. #5 Perf Diff 27 w/ Frequency Nika 20.5/34 2499 2115 384 272 384 520 Shant 18/23 2715 2477 238 74 303 395 Manu David 11/18 2348 2044 304 58 304 376 Grant Xu 10.5/18 2501 2207 294 182 294 366 Ravuri 11/30 2295 2050 245 245 365 Grigoriants 13/18 2738 2543 195 110 283 355 Dobrov 19.5/27 2677 2478 199 35 237.5 345.5 Duda 12.5/15 2837 2700 137 54 274 334 Naka 20/26 2812 2700 112 29 224 328 Bloomer 8.5/26 2222 2000 222 222 326 Tianqi 11.5/19 2558 2312 246 0 246 322 Grabinski 3/12 2274 2000 274 2 274 322

I do think that if you think about other factors, Nika is a deserving MVP if the season ended now. He is irreplaceable (only one other player u2250, Grant Xu has a comparable performance, but he's played twice as many games and costs his team 100 less points in rating cap), and specifically on his team, they don't even have a single legal lineup without him playing! On top of that, his team is #1 in the league without having a single 2700 player or player performing over 2700. Jobava and Giga have played well, but Nika has contributed a huge amount to his team's success, and almost any other team that had him would instantly become the best team in the league.

I think Shant Sargsyan also deserves a spot around #2, with his consistent 2700+ performances from board 3 putting his Armenia Eagles at the #2 spot in the league right behind the Gentlemen.

On the other hand, I think these numbers have overvalued the 2300 level performances of Manu David and Vinesh Ravuri. 2300 level performances from 2000 rated players are not quite enough to push a team to the top without some 2600+ players on boards 1-3 delivering major points. For that reason, I think my current system has undervalued Nakamura and Duda who have the two strongest performances in the league. I do think Grigoriants and Dobrov have scored enough points that they belong in the top 10 for sure. If, as a manager, I had to pick someone to join my team, and Nika Volkov were already taken, I would hesitate between Nakamura and Shant Sargsyan. Nakamura has been willing to play at 4 in the morning from tournaments, which may cost him a few points of performance rating compared to Duda who has prioritized his classical tournaments, but means that the Sluggers can always count on him to power them. To me that would be a tiebreaking factor in favor of Naka over Duda.

Grabinski popping up-- no offense to him, he has been valuable-- shows an error in the math as well, because 3 points just are not worth that much in the league; it hardly matters if the performance was 2220 or 2300 at that point. The teams who are at the top of the standings have been powered by Sergei Azarov's 19.5 points, Xiong and Conrad Holt's 18.5 points, etc. (with the exception of the Pandas who have 6 different super-GMs showing up for ~10 games each, and winning regardless of who plays which week). So when I revisit this work, I will look at why Iturrizaga and Xiong don't show up on the current top 10 list.

What do you think? If you put yourself in a manager's shoes, which one player would you most want? What aspects did I not look at? How should I weight the value of higher performances v lower ones?

Blogs