Sunday, December 02, 2007

Blitz almost lose, almost win, and finally lose

Cancel the plans for the rolling rally. Don't bother waiting for another Chris Williams Riverdance. And forget about a celebratory Championship video (which probably isn't a bad thing, since several Championship spectators expressed their displeasure with the Eastern Division Championship version). They're holding the parties at UTD, while Bostonians are left to contemplate the reality that a single blitz game shattered their US Chess League Championship dreams.

The storyline is fairly well known by now. The Blitz appeared well on their way to defeat after GM Christiansen could achieve no more than a draw by repetition against Boskovic on Board 1 and things looked dire for the good guys on all three remaining boards. Sammour-Hasbun was fighting an uphill battle after Kuljasevic sacrificed a piece in a French Tarrasch; Shmelov's promising position had given away to a strong attack by Stopa against his King; and William's opponent, Zorigt, had secured an advanced passed pawn and nice piece outposts for the exchange.

However, it wasn't that long before the tables had turned completely. First, Shmelov managed to find a nifty resource to secure a drawn position. Next, Jorge finally coordinated his pieces, broke through and traded off to a won two pieces vs. a rook ending. Suddenly, the Blitz were just a draw away from the title and fortuitously, Williams' game had suddenly become eminently drawable. Yet it was not to be, as Chris blundered away his chances and allowed the Destiny to the tie match.

So, off they went to blitz tiebreaks. While there are certainly many paragraphs that could be written about the tiebreak games, in my view the USCL tiebreak system makes only the top board matchup relevant, by design....
I suppose an aside to explain this view is warranted, since I'm sure some will disagree with my supposition. While it is theoretically possible that a team's 1st board might be defeated by the 2nd or 3rd board of the other team, I suspect the actually probability of this is fairly low. Therefore, more often than not, it doesn't really matter what happens in the preliminary blitz games because eventually the 1st boards will face each other for the title. So far, this is how things have played out the two times tiebreaks have been used. It's only a guess on my part, but I'd bet that over the next several decades, this will be the case more than 9 out of ten times (we should only be lucky enough that the USCL shows enough longevity to generate such a large sample).
....In the game that mattered, GM Christiansen eschewed a draw and played on seeking winning chances. Instead, it was Larry who made the critical blunder and a minute later Boston's season was over. Some might claim that Christiansen's loss was attributable to the fact that he had to play three blitz games as opposed to his opponent's one. I find this highly unlikely. Larry may very well have been tired, but only because it was 2:00 am, not because he had to play an extra 30 minutes of chess.
You can find analysis of key positions from the match at Braden's blog and Arun offers his overall impressions. I thought I might try to get inside the heads the combatants themselves. While everything that follows is conjecture and speculation, there are two key moments which I'd like to explore.

First, why did Williams throw away his rook with 50.Rxf7? It doesn't take a master to see that there was no winning follow up to this sacrifice and Chris is more than strong enough to have realized this. Instead, I think he thought he had found a quick way to draw the game, only to realize after the sacrifice that his key move wasn't possible.

NM Williams - WFM Zorigt
After 49.Qd3

When he played 49.Qd3, not only was he offering to trade queens, but he also had setup the variation 50.Rxf7 Rxf7 51.f6+ Kg8 52.Qd8+ Kh7 (if Rf8 then 53. f7+) 53.Qd3+ with perpetual check. However, only after he sacrificed the rook did he realize that Black's 49th move, Qxa5, covered the d8 square. There was no check and no perpetual. Time pressure certainly played a role, but I think impatience was a factor also. Chris was looking for a quick way to bring the game to a close since the alternative was likely returning the exchange for the pawn on d2 and playing on for awhile to draw a slightly inferior but eminently drawable endgame.

Second, I've been pondering why Larry decided to risk it all in a totally equal position in the last blitz game when he could have simply taken a draw and tried for more in the next one. There are certainly plausible explanations like it was late and he just wanted to get things over with, or he thought he could outplay Boskovic in the ending, or he didn't think his chances of winning with Black in the next game were particularly high. While one or more of these may very well have been a factor, I think an off-board occurrence may well have influenced the final outcome the match.

GM Christiansen-IM Boskovic
after 31...Rd7

Larry certainly could have
considered taking a draw here

You may recall that there was an extra long delay before the last blitz game commenced. My understanding is that this was due to the fact that the UTD Student Center, where Dallas was playing, was being shut down for the night and the team needed to move to another location. While there had been no internet connection problems throughout the long match, there was one after Dallas moved to the new location, right in the middle of the final blitz game. Why did this matter? Because early on in the game, Larry had been using much more time than his opponent. At the time of the disconnection, he had only about two minutes left to his opponent's three and a half or so. I contend that if that deficit had remained the same, Larry would likely have acquiesced to a draw later on and played another game. Instead, because of the disconnection, he was given an extra 2 minutes. As a result, at the key moment (when he had to decide whether to take a draw or risk going for a win) he had a time advantage and this may very well have influenced his decision to play on. Maybe UTD Security should be given an assist for their role in securing Dallas' victory?

All speculation as I said before, but perhaps the players in question will shed some light on what was actually going through their minds.
I suppose my comments above on the tiebreak system might be viewed as a critique. While I actually don't mind the current system and found watching the tiebreaks to be very exciting, I do think there is at least one viable alternative which, among other things, would probably take less time. What about just switching colors and having the same players play each other again (1 vs 1, 2 vs 2, etc.) at a time control of G/10? If the match draws again, switch colors again and repeat at G/5. Keep going until one team wins. I think this makes sense and would likely take far fewer rounds.
So that's it for the 2007 season, except for post season awards and the USCL hot stove season. Dallas and Boston both had great seasons and played an exciting Championship match. The league itself also continued to make major strides forward and the USCL blogosphere had some significant growth as well. Where's it all going to be years from now? Beats me, but it's a good bet that things will be bigger and better in 2008.

BCC Weblog provides independent coverage of the United States Chess League. It is not affiliated with the USCL or the Boston Blitz.

No comments: