Thursday, May 3, 2012

Repairings at Junior High Nationals

So I have a rules question for everyone.
You may remember a post from two years ago, called Repairings in Elementary Nationals, in which James Black's first round opponent did not show up, and insanity ensued.
This time no one seemed to have had malicious intent, but a strangely similar thing happened to the same person last weekend:

James Black was the second seed in the K-9 open. Five pre-registered players did not show up for round one, and it is USCF policy to repair players without opponents. The 5 players had ratings of 2300 (James), 1800, 1600, 1100, and 900. The TDs paired the 1800 with the 1100, the 1600 with the 900, and gave James a bye. Their reasoning was that this kept the pairings as close as possible to what they would have been originally (the cut for round one was about 1730). I protested, as it seems very strange to me that the top seed should be left unpaired, especially as that has a serious potential effect on his tiebreaks. Modified median is the first tiebreak, and while MM drops the lowest score, it seemed like a big disadvantage to concede that in round one. It also seemed like it would potentially affect the 2nd seed much more than a 900, and so the bye should have gone (as it pretty much always should) to the lowest rated player.



Keith Ammann said...

With late arrivals, it depends on what order the players arrived in. They're paired as they show up. James Black was originally entered in the K–8 section and switched at the last minute, correct? So he might have been construed as a "late entrant" to K–9. If the other four players were entered in that section to begin with, or if one or more of them were switched before James was, they might have been given precedence on those grounds.

Elizabeth Vicary said...

James was not a late entry. Players are not paired as they show up.

Chris said...

Hi Elizabeth,

Since I was the TD that did the repairing I guess I can shed some light on the situation and give my perspective. However, the first thing is to correct the ratings of the players I had to do repairings for in round 1 as that makes a substantial difference to your version of events.

Here are the 5 players I had available due to 4 players switching sections after the pairings were done and one no-show:

Black (2312) - due to play a 1710
Wu (1944) - due to play a 1430
Kluska (1844) - due to play a 1380
Bolshakov (1480)
Hernandez (1156)

The cut off point for the top half/bottom half was 1714 and we were told in our pre-event meeting that no top half vs. top half, or bottom half vs. bottom half pairings should be done when repairing. I also had no issued with players being from the same team or state, which are other things to take into account when repairing on the floor.

With that in mind, I paired Wu with Bolshakov since Wu was due to play a 1430 and Bolshakov was within 50 points of that and Kluska against Hernandez, which was about 200 points lower than he should have played.

Since Black was due to play someone around 1710, giving him an opponent 230 points lower (Bolshakov) seemed to not be the most ideal pairing when compared to the other pairings I had available.

Obviously I could not pair him with Wu or Kluska, the latter of which I believe is the player you wanted him pairing against since all 3 players were in the top half of the section. This also meant that one of the top half players would have to end up with the bye.

Your argument on site was that since Black was the #2 seed then he should have a higher priority than other players. My philosophy was that all the players are treated equally and I find the repairing that best matches the original pairings if possible.

In the end it turns out that Black finished with the first place trophy on tie-break, the first one of which is modified-median. With that in mind I'm not sure your main argument that having an unplayed game ruins your tie-breaks holds much weight.



ppmint said...

I am not a TD and has no formal training as a TD but as a veteran chess parent and an elementary school chess coach, I find this repairing in question somewhat strange.

The norm in any given tournament is that if there're odd number of players in the same section (in this case, the repairing pool), the lowest rated player should almost always get a full-point bye. Is this a rule in the USCF rule book or TD training materials? Probably not, but the reasoning is quite simple - the lowest rated guy (bottom half) would "appreciate" the free full-point since he/she is less likely to get it normally by playing a top half player in Rd 1.

Of course, ideally, James Black should play someone around 1710, but there are only 5 players that need to be repaired, the pool is limited and the situation is not "ideal". The notion of pairing everyone according to who they're "due to" play is unrealistic. The best way is to treat treat them as a separate group - in this case, Black (2312) and Wu (1944) are in the top half and Kluska (1844) and Bolshakov (1480) become the bottom half.

Giving the 2nd seed a bye in the first round is certainly a disgrace to his chess ability but in my opinion, the money question is: how much weight does a full point bye carry in calculating the tiebreak? Was James's tiebreak calculate based on 6 games or 7 games? Did the 1st round bye hurt him or help him? Judging from the result, the bye certainly didn't hurt him.

Michael Aigner said...

The problem here is not the pairings assigned. There is no suitable pairing for Black. In fact, the desired pairing of Black vs Kluska has the opposite effect of helping Black's tiebreaks. (The average 1844 would score about 0.5 higher than a 1710.)

The real issue is how modified-median is calculated. All unplayed games, including full point byes, count as 0 tiebreak points. Normally the bye goes to someone at or near the bottom of the standings, someone who gets a free point that they likely would not have earned. Alas, Black would be heavily favored to beat any first round opponent.

Given the policy of repairing in round 1, the only fair solution is to give Black the bye and 3.5 tiebreak points (0.5 per round). Singling him out as the only one without an opponent and downgrading his tiebreaks is double jeopardy.

Elizabeth Vicary said...

I had responded to Chris on facebook and forgot to do so here. It's an interesting exchange, but I would summarize my points as
1) Matthew Kluska is a 318 player so he can't play James.
2) I don't see why anyone has a right to play a similarly rated person as they were initially paired against, or I don't see why that the good of that remaining unchanged trumps basic pairing rules. Rule of law here, guys. It's only the way it is because some coaches/parents who are ENORMOUS BABIES complain when their child has to play someone unreasonably skilled at chess. Because of the team conflict, I would have paired 1 with 4, 2 with 3, and given 5 a bye.

Michael -- That is exactly what I asked the TDs for -- 3.5 tiebreak points. Unfortunately, while Phil Smith and Jon Shacter thought this was the rule at first, the USCF rulebook is very clear that you get no tiebreak points for unplayed games.

Elizabeth Vicary said...

You know, it's like *this is a chess tournament*. You got unlucky that your kid's opponent didn't show, that's unfortunate, bad luck, but this is a chess tournament, not a therapy session. Your child is here to play chess, not to avoid any hint of a challenge until Sunday morning. Given that the USCF policy is to do repairings, your child will repaired. It's unfortunate the TDs can't produce a replica of the lost opponent's skill level, but they can't, and that's bad luck for you but it's not an injustice.

Steve Immitt said...

I recall recommending something like the Modified Modified Median tiebreaks for this situation the first time this issue came up under the heading "Repairings At Elementary Nationals" probably two years ago.

Elizabeth Vicary said...

I remember. It was a surreal experience to be discussing the same situation and the same kid with a similar-- but not quite identical-- set of people who seemed to be saying a very different thing than they said two years ago.

Unknown said...

I'm hard core - forfeit the no shows. Didn't BF have that happen for his no show? I don't think they would have repaired the US Championships if Kamsky did not show up . . .