Wednesday, November 08, 2006

Break out the slide rules

Here is some input for those of you contemplating what lineup the Boston Blitz should employ in the US Chess League playoffs. Below are regular season statistics for each member of the team sorted by the differential between their Performance and USCL Rating (Data format: Player Name, Score, USCL Rating, Performance Rating, Differential):

NM Riordan, 4-1, 2272, 2565, +293
FM Kelleher, 4-1, 2402, 2595, +193
NM Krasik, 4.5-2.5, 2162, 2330, +168

NM Martirosov, 3-3, 2259, 2307, +48
GM Christiansen, 4-2, 2633, 2669, +36
GM Perelshteyn, 3-1, 2614, 2628, +14

IM Foygel, 2-2, 2533, 2443, -90
FM Winer, 1-2, 2422, 2189, -233

Conventional wisdom says that the Blitz will be using their two GM lineup (Christiansen, Perelshteyn, Martirosov, Krasik - avg. rating 2400) throughout the playoffs, and that may very well turn out to make the most sense. However, notice that this lineup does not include either of the best performing players on the team. In addition, while NM Martirosov does have a positive performance differential for the year, he has only two draws in his last four USCL contests. Further, if opponents know exactly what Boston's lineup will be in advance they can carefully select their own lineup in a way which improves their chances, e.g., using a balanced lineup against a top heavy one or guaranteeing a specific matchup on a particular board.

With that in mind, let's consider different lineups which include the top performers and still maintain a high overall average rating. There appear to be two feasible alternatives, though each has its drawbacks:
  1. Either GM, Foygel, Riordan, Krasik - 2389
  2. Either GM, Kelleher, Riordan, Martirosov - 2381
Whether either of these makes any sense will depend on specific matchups. For that, we need to wait for the outcome of tonight's contest between New York and Carolina.

No comments: