Back when I first joined the USCF, in 1972, you had to wait months to find out your new rating after you played in a tournament. Nowadays, if the tournament directors do their job quickly enough, you can find out overnight.
In fact, Michael Aigner found out even before I did! When I logged onto Facebook this morning, the very first comment I saw was from him: “2199!”
This was a disappointment to me, for sure, but I knew that something like that might happen. It became even a little bit more disappointing when I saw that my new rating was estimated in the crosstable on the Bay Area Chess website… at 2203. This estimate was calculated by Tom Langland’s Swiss Sys software. It seemed like a pretty big discrepancy to me: a 17-point rating increase (USCF) versus 21 points (Swiss Sys).
However, the subsequent discussion on my Facebook page clarified things for me. Swiss Sys does not use the latest USCF formulas, but outdated ones from the 1990s. When I calculated my rating using the calculator on the USCF website, it came back as 2199.
I thought there still might be a shred of hope for 2200 because Michael Aigner said that when you play a provisionally rated player, or a player whose rating jumps by a large amount, they recalculate your rating via a feedback loop. Indeed, my round three opponent (from Argentina) did have a provisional rating, which increased from 2248 to 2284 at this tournament. Could the feedback put me over 2200? Alas, according to Tom, the answer is no… in fact, without the feedback my rating comes out to 2198 and the feedback loop adds 1 point to it.
As it turns out, where I was really unlucky was my choice of an Argentinian opponent. The other Argentinian player, whom Michael played, came into the tournament UNRATED and had a post-tournament rating of 2357. The feedback made Michael’s rating come out 5 points higher than expected. “My” Argentinian had an even better performance rating and better score at this tournament, but because he already had a USCF rating, his post-tournament rating was not as high.
Another thing I learned today was that you can actually see your rating to hundredths of a point on the USCF website. My post-tournament rating was not 2199 but 2198.74. So if we could find some way to get me another 0.76 points (basically three-quarters of a rating point) that would get me to 2199.50, which would presumably be rounded up to 2200.
Well, all this agonizing and stressing over three-quarters of a rating point is pretty comical. It reminds me a lot of a close election. My exit polls (the Swiss Sys software) show me with four more votes! I demand a recount!
Normally I really wouldn’t care whether my rating varies by 1 point or 4 points from what I expected. But that last step between 2199 and 2200 is a pretty big one psychologically. It’s been almost twenty years since the last time my rating was over 2200, and I am sooo ready to get over that hurdle. However, in the long run it’s absolutely meaningless. What’s important is the process, not the results, as I wrote in my last blog post. The process gave me a 2374 performance rating at this tournament. And that’s my real goal — not 2200, but 2300. If I keep playing the way I did this weekend, I will get there eventually.
So I am now officially over it. I will enjoy my Sandbagger’s Special rating of 2199 while it lasts.
Addendum (added later in the day): I just couldn’t resist my curiosity. I started wondering, what if some of my best tournaments from the past were re-rated using today’s rating system? Presumably over the long run my rating would have been the same, but over the short run there might have been some changes.
So I decided to look at the four-tournament run that led to my highest rating ever (2257, after the 1994 Cardinal Open in Columbus, OH). This was right in the middle of the time when we had a higher K-factor, allowing for bigger rating changes and quicker progress.
First, my rating changes for those tournaments as computed by the USCF at the time:
1993 Ohio Chess Classic. Performance 2359, rating change 2183 -> 2202
1993 Roosevelt Open. Performance 2413, rating change 2202 -> 2220
1993 Trick or Treat Open. Performance 2280, rating change 2220 -> 2226
1994 Cardinal Open. Performance 2445, rating change 2226 -> 2257.
Now if we re-compute those ratings using the 2014 rating system, starting from the same base, we get:
1993 Ohio Chess Classic: 2183 -> 2198
1993 Roosevelt Open: 2198 -> 2212
1993 Trick or Treat Open: 2212 -> 2218
1994 Cardinal Open: 2218 -> 2247.
So my lifetime peak would have been ten points lower. It’s interesting to note that the first tournament in this series of four was almost a perfect mirror of the tournament I played last weekend, where I had a 2374 performance rating and went 2182 -> 2199. I think it’s fair to say that under the old system my rating would have gone to 2204, which pretty much agrees with the Swiss Sys calculation of 2203.
But what’s really important is that I followed up that 1993 Ohio Chess Classic with three more good tournaments in a row. And that made it academic whether I went over 2200 at that tournament or not. Let’s hope that last weekend’s tournament will also be a springboard to three more good tournaments (at least!)
