Notes From the Stats Guy

                by Russ Mollot

I’d like to start with a quick “thank you” to Greg, not only  for putting this article up on the USCHESSLEAGUE site, but for permitting CXR (Chess Express Ratings) to continue tracking the League’s games (we now have both seasons’ results in our database). (ed - and the league would like to thank CXR for rating the games as well!)

The expansion teams have added a lot of excitement and the high caliber of the participants across the nation has made this a positively amazing season.   One can easily spend hours on the web-site, which is full of interesting games, of course, plus all the “bio’s” of the players, and so on. By the way, you can see the complete season’s wall-chart at www.cxrchess.com/WC155.html where you can click on any player’s name to see their statistical profile.

This article will attempt 3 things:

1)       Explain how we dig through our statistics to do a rough analysis of a player

2)       Give a “snapshot” analysis of each player on the San Francisco and New York rosters.

3)       Make a guess on the outcome.

Here at CXR, we’ve been quietly absorbing all the data, and we’ve cranked out some very interesting “stats” on the more active players.   Now, for sure, we’d like to take a look at the likely line-ups for the upcoming Finals match between San Francisco Mechanics and the New York Knights.  For the ensuing discussion, I recommend opening another browser to our web-site address, www.cxrchess.com, where you can follow along with how we can do some very cool analysis.   I’ll “walk you through” this process for two of the players, then you’ll be able to do it for any players you like. 

One thing that jumps out at you is the “Hot List” on the upper right portion of the home page.  The Hot List highlights those players who have gained the most rating points in the current month.   Okay, I know that you’ve spotted the “Top Active Players” report right below the Hot List, and you can definitely find all your favorites there, but please do that later and stick with me for now.  Each of these are really miniature versions of the full reports, and if you click "view more” you get to the full report.  Looking at the Hot List, we can spot several of the U.S. Chess League participants there, by scanning down the column labeled “Affiliate”.  As of this writing, Matthew Herman of the New York Knights is tied for 10^th place with Chris Safranca, a youngster in Hawaii.  By clicking on Matthew Herman’s name, you are taken to his Player Profile page, where you will see a bar chart called the “CXR Ratings Spectrum”.  Nearly all the statistics on this page are based on what we call “official results”, which excludes the activity of the latest 2 weeks (because such recent activity is subject to corrections).  Looking at the bar chart, what jumps out at you is the very long bar labeled “Unofficial Rating 2147”.  This number happens to include all activity (even that of the latest 2 weeks).  This shows that Matthew’s rating has leapt forward in the past 2 weeks (and that is why he’s on the Hot List).  Note the third bar: “Best 2094”.  This was Matthew’s official Best (highest) rating achieved … that is, until now.  We see that he has surpassed his personal best.  Note also that the bars labeled Uphill, Downhill, Black, and White are all about even.  This shows that Matthew has a “balanced” game, equally comfortable playing stronger or weaker opponents, equally comfortable playing Black or White.  This is the hallmark of the very best players. 

On the upper right of this web page you should see a link labeled “View Game History”.  Clicking this takes you to a very interesting report.  The Pie chart at the top shows an equal proportion of wins and losses – a seemingly average performance.  But scan down at the details of Mathhew’s games.  The first four games were played in February of 2005, toward the end of the New York Masters tournament.  Matthew lost 3 games against much stronger opponents, but won 1 game against a weaker opponent.  That was then – this is now.  In the current U.S. Chess League Season, Matthew played six games – all against stronger players.  He only lost the first game, won 3 and drew two.  Since September, Matt has gained roughly 90 rating points.  Two Pie charts at the bottom of the report confirm that Matt’s results were the same with White as with the Black pieces.   Note that this analysis is, unfortunately, based only on ten games.

Let us look at New York’s 1^st Board player, GM Pascal Charbonneau.  From our home page, click “Look Up A Player” (look in the left column, underneath “Ratings & Stats”) and type in the first few letters of the last name – i.e., charb should be enough.  Click the Search button and Pascal Charbonneau comes up.  Click on the link to his Profile.  We are going to learn some important things here.  The graph labeled “CXR Rating Over Time” shows that since slumping below 2400 last season, GM Charbonneau has toughened up noticeably, peaking earlier this season at 2564.  However, since then, he has fallen back about 50 points.  The mini-pie shows he is about even in wins and losses, with a big slice of “grandmaster draws”.   The Ratings Spectrum (bar chart) shows that his Black game is better than his White (the spread used to be greater, but Pascal’s recent results have been somewhat more balanced), but it appears that his “Downhill” number is low.  This indicates that GM Charbonneau has given up more points to weaker players than should have been expected.  Note the blue box labeled “PERFORMANCE PERCENTAGE OVER TIME”, where we see that he scored only 50% in the last 5 games, as compared with scoring 69% in the previous 8 games; this is a weakening trend.  Now, at the top right of the page, click “View Game History” and let’s see what we can learn there.  Our Overall Results pie chart is repeated here, followed by the game results that we have on record.  Further down we see the Additional Analysis section.   Although we saw in the pie chart that Pascal’s wins and losses were roughly equal, here we see that he actually has had better results with the Black pieces,   Having Black, he has scored 60%, gaining an average of  5 rating points per game, whereas with White, he scored 42% and lost 4 points per game on average.  In the beige box, we see again that Pascal has done better defending (i.e., having Black), in both King Pawn and Queen Pawn openings.  Finally, please scroll all the way to the bottom and we have two new pie charts:  Results With White and Results With Black.  We can compare them to each other and also against the Overall Results pie.  Take a look at the Red (Games Lost) slice on each pie chart.  GM Charbonneau has lost more than half his games with White, but only 20% with Black.  In other words, playing Black, Pascal Charbonneau is very unlikely to lose!

Now that you’ve seen what the procedure is for navigating the CXR web-site, I will skip the navigation remarks and stick to the statistics.  What I will do is give some numbers and quick analysis on those players most likely to be in the lineup for the Finals. 

Let’s look at the available players.

Team Roster:

 IM Josh Friedel – USCF 2513

Josh is the most likely 1^st Board for the Final match.  He will play the White pieces, and he does not tend to lose games with White.  In 7 League games we have in the database where Josh played White, he won 5 and drew 2.  I believe that  IM John Donaldson, the team’s manager, chose White for boards 1 and 3 in order to play Josh Friedel on Board 1. 

 IM Vince McCambridge - 2502

With only 8 games played, McCambridge has won half and lost half his games.  I don’t have enough data to make a judgement.   

 IM Vinay Bhat - 2463

IM Bhat hasn’t lost any of the 8 League games that we have on record.  He has 3 wins and a draw with White; 3 draws and a win with Black.  So Obviously he’s stronger with White.  However, he can practically  guarantee a half-point for Black – which might earn him the 2^nd Board slot. 

 IM David Pruess - 2459
We have quite a bit of data on IM Pruess, dating back to the New York Masters.  He has about 47% losses on Black; but with White, David has only lost 12% of his games, with 60% wins, and 28% draws.  He would be ideally placed on 3^rd Board, but not 2^nd Board (where he often has played) which would give him the Black pieces. David has also played stronger in his last 25 games than his long-term “norm”.

 IM-elect Dmitry Zilberstein - 2435

This season, Donaldson has assigned Dmitry Zilberstein the Black pieces 5 out of 6 times “at bat,” resulting in 2 wins, 2 draws, and 1 loss (he won his game with White). 

 
 NM Mark Pinto - 2217

Only 6 games in our database for Mark, but he has won all 3 of his games with Black, including a win over New York’s FM Boris Privman! 
 

Sam Shankland – 2106

Sam has won 5 of his 6 League games this season, and has a 4-game win streak going.  But his single loss was on Black.  Still a possibility, and Sam’s low rating would help keep a top-heavy lineup within the team rating ceiling.

 Daniel Naroditsky – 2055

With only three games this season (and the team’s lowest-rated player), Daniel is unlikely to play in the Final match.

NEW YORK KNIGHTS

Team Roster:

 GM Pascal Charbonneau - 2500
As we saw in the analysis at the beginning of this article, Pascal’s strength is on the Black side of the board.  At Board 1, he will have this advantage.  This should prove interesting assuming that Josh Friedel plays Board 1 for the Mechanics, as the latter has not lost (this season) with White.  Although the USCF ratings of these two players is nearly the same, Charbonneau sports a CXR rating of 2540 with the Black pieces, whereas Friedel’s CXR White rating is higher still, at 2585.  

 IM Irina Krush - 2445

IM Krush is Jennifer’s likely pick for Board 2.  Irina is playing in top form, with her rating at its highs.  She is at her best on White, losing only 21% of her games with this color.  That means she will play her better color as Board 2. 

 FM Robert Hess - 2369

Young Robert has gained some 300 rating points in the last 3 years and is still improving.  He won his game in the Wildcard round and drew in the Semi-Finals.  FM Hess is playing better with Black.

 IM Jay Bonin - 2354

Jay has probably played more rated games than any other person, living or dead.   Naturally, that means we have a lot of data on Jay.  He sometimes plays at grandmaster strength, and his CXR rating has carried above 2550 recently, though now back to around 2450.  Jay plays decidedly better with White, so playing Jay as Black on Board 3 may be problematic.  With 41% losses on Black, versus only 39% wins (and 20% draws), Jay would have to pull the proverbial rabbit out of his hat (and he doesn’t wear a hat).

 NM Mackenzie Molner – 2367

NM Molner had a bad season with the USCL this year, as, unfortunately he was given the Black pieces every time.  Mackenzie is actually quite strong with White.

 FM Boris Privman - 2338

Color doesn’t matter to FM Privman, he is quite able on either side of the board.   His rating chart looks like a cyclical stock, as he seems to have many “hot” and “cold” streaks.   It looks like his latest “hot” streak has tapered off, and he may be just about to slide.  Caution!

 WGM Jennifer Shahade - 2318

Jennifer is off her best form, as much of her time has been taken up by her duties with Chess Life and promoting chess.  She has used herself as a “pinch hitter” twice this season, but does not intend to play in the Final match.

 Matthew Herman - 2172
As noted earlier in this article, Matthew is equally adept on White or Black, and since losing his first match this season, he has been undefeated since.  He won his game in the Wildcard round and also won his game in the Semi-Finals.  He is hot. 

Well, here we are just before the Final match.   Let’s examine a tentative lineup:

Unfortunately, there is a problem with these lineups – they are both illegal.  Remember that the average rating for each team cannot exceed 2400 (with a 10 point bonus for female players, New York could take advantage of a 2410 point ceiling).  So let’s try another lineup, balancing higher-rated players with lower-rated players in order to remain under the rating “cap”.

Now we are legal.

With the above hypothetical match-ups, I think Josh (ideally placed with the White pieces) has a better chance for the win on Board 1, even though Pascal is very strong with Black. Since GM Charbonneau has played King-pawn openings (both as White and as Black) almost exclusively this season, I suspect that Josh is going to surprise his opponent.  On Board 2, I am predicting a Draw.  Even though Vinay Bhat is undefeated this season, having the Black pieces he has only won 1 game and drawn 3 times.  Irina’s record shows she is unlikely to lose with White, but because Vinay’s CXR Black rating is 36 points higher than Irina’s White rating, I think he can stop her from scoring a full point.  On Board 3, David Pruess is almost unbeatable with White, giving Robert Hess a very difficult task.  Going strictly by probabilities, we should write David in for a win.  On Board 4, both Matthew Herman and Sam Shankland have been “hot” this season.  Each has only lost one game.  But while Matt had 3 wins and 2 draws, Sam had 5 wins.  That would make Sam “hotter”.  So what do we have?  Possibly the following….      

So, if these are the lineups, I am predicting San Francisco to win the championship  3 ½  to  ½ (and I am a New York Knight fan!).   At least we won’t have to worry about those crazy Tie-Break rules! Well, that’s it from the “Stats Guy”.  Like you, I am eagerly awaiting this battle to begin!