March’s “Just the Rules” is posted! https://new.uschess.org/news/just-rules-re-entry-gambit/
If I didn’t know better, I would have thought reentering after a bad start was an April Fool’s joke.
Michael Aigner
At the very least re-entry ought not to be allowed in small sections. I once saw a section of about 10 players where the winner with 4.5 never played the runner up with 4. He had lost to him Friday night, re-entered and thus could not face him again.
That is a tough one.
If they were 3.5-0.5 (re-entry) and 4-0 or 3.5-0.5 (original winner) then you either force the original winner to win (or draw) against a player already beaten (can be perceived as a burden - especially if it was an upset round one win) or you allow the re-entry to leapfrog over the original winner by outscoring that other player in the final round.
When rating a section of an event, all of the games under the same member ID are pooled together for ratings purposes, regardless of what pairing number they’re under. That covers both events that allow re-entries and multiple-schedule events that are merged together for the final rounds, as the latter generally allow players to play in more than one schedule, deciding which one to take into the merged portion of the event.
So it is possible in a 5 round event for a player to have more than 5 ratable games. This affects how bonuses and norms are computed for that player.
It is worth noting that because of how events are reported, we don’t currently know what schedule of a multiple-schedule event someone was in. Hopefully a new reporting format will correct that at some point, though I’m not sure if WinTD currently keeps track of what schedule someone was in.
That’s good to know (actually, I already knew it), but that’s just for ratings purposes. By the time the crosstable is received at MSA, all the pairing questions, prize eligibility, etc, have already been determined at the tournament.
I suppose that would be nice, mainly for FIDE purposes, since the time controls used in the faster schedule might not satisfy FIDE’s requirements for the particular type and level of the tournament.
Bill Smythe
I seem to recall when one playing schedule is not FIDE ratable (time controls!) and it is merged into the FIDE rateable time control schedule there is a hoop to jump through before it is submitted to US Chess; i.e., the TD must separately ID the players in the non FIDE schedule. That-a-way only the FIDE rateable games get left and submitted to FIDE. Did I remember that correctly?
Not the players, the games. With WinTD, if you properly tag a section as non-FIDE rateable, then any games paired in that section will get tagged as non-FIDE rateable. If that section then gets merged into a FIDE rateable section, the further games by the players will be tagged as FIDE rateable. The rating report will show 0 - 0, 0 - = or 0 - 1 for the non-rateable games, which are basically the bye/unratable game codes for no credit, draw and win. (The usual fields are <> <> <>, i.e. something like 21 b 0 for played 21, black, loss.)
Back in 2010, I played the same opponent twice in a smallish Master section (14 players) of a class tournament. I reentered after starting out 0.5/2. From the perspective of my opponent, we played in both round 1 and round 6 (last round), splitting the two games (he won the first). The tournament situation was critical as we played for a share of second place.
If I remember correctly, my opponent questioned the last round pairings and the directors spent considerable time debating and verifying by hand, delaying the round. For once, I kept my trap shut, as the pairings I honestly expected were even worse (a strong student of mine). I don’t remember the exact line of reason, but the pairings stood and I won the rematch.
http://www.uschess.org/msa/XtblMain.php?201002159321.1
Standings after round 5
IM DeGuzman 4.0 (played Liou, Aigner, Sandberg)
NM Liou 3.5 (played DeGuzman, Mackenzie)
NM Aigner r/e 3.0 (played DeGuzman, Mackenzie**)
NM Loanzon 3.0 (last round bye)
IM Shipman 3.0 (played Sandberg)
Sandberg 3.0 (played DeGuzman, Shipman, Mackenzie)
Mackenzie 3.0 (played Liou, Aigner**, Sandberg)
everyone else 2.5 or less
Posted round 6 pairings (without colors)
DeGuzman vs Shipman
Liou vs Sandberg
Aigner** vs Mackenzie**
Alternate round 6 pairings (without colors)
DeGuzman vs Mackenzie**
Liou vs Shipman
Aigner** vs Sandberg
I can’t remember exactly why the directors rejected the alternate pairings, but it was probably a +2 color conflict or 3 in a row. The directors ruled that having us play again was the lesser of evils. If it matters, the section was FIDE rated.
Michael Aigner
Fascinating story. I betcha if you tell me the colors you played in each of your seven games, I’ll be able to figure out the colors for all the players in all the rounds. And then we can all have a furious and interesting debate over which set of pairings was better.
Just to confirm: This was a 3-day tournament with a 2-day alternate schedule, merging after round 2, right? It would also be helpful if you could tell me (or if I could figure out) which players played in which schedule.
Let’s have at it!
Bill Smythe
I found some emails from February 2010 in an old account. The format was a typical 3-day tournament (Sat, Sun, Mon) with a 2-day schedule (Sun, Mon). The merge came on Sunday late afternoon after round 3.
My colors were B, bye, reenter, W, B, W, B, W and B. Yes, not only did I play Mackenzie twice, I played black in both games! Rule 28S3 clearly states the colors in the first game should not impact the second game. I don’t know anyone else’s color history.
My old email says two players appealed the pairings, including my opponent. Supposedly I told my three students in the section that I expected the pairings to change upon appeal and that I would likely play one of them. For that reason I did not intervene in the proceedings at the TD table.
The primary reason for the Chief TD’s ruling was that Rule 28S1 states “they should not be paired against each other for a second time.” In other words, the pairing is undesirable, but not illegal. The other factor was giving the tournament leader his preferred pairing in the money round of a Grand Prix event.
The appeal was denied and the computer generated pairings stood.
Have at it.
Michael Aigner
OK, I think I’ve figured out the missing colors, and which schedule each player played in. The second column below represents the schedule (3-day or 2-day). The last-listed player joined the tournament only after the merge, so he didn’t really play in either schedule specifically.
I looked up all the ratings in the February 2010 supplement. Those ratings (listed below) are the ratings that presumably were used (or should have been used) for pairing purposes at the tournament.
I have listed the players in rating order rather than score order. That makes it easier to see why the pairings were made the way they were. In each result column I have listed the opponent number and the cumulative score after that round. Again, that makes it easier to understand (and justify) the pairings.
[size=150][code]
colors rd 1 rd 2 rd 3 rd 4 rd 5 rd 6
------ ------ ------ ------ ------ ------ ------
01 (2) RICARDO DE GUZMAN 2443 bwbwbw 11 1.0 2 2.0 13 2.5 7 3.5 6 4.0 5 4.5
02 (2) MICHAEL AIGNER 2234 wbwbwb 13 1.0 1 1.0 11 1.0 12 2.0 9 3.0 8 4.0
03 (3) MICHAEL AIGNER 2234 b----- 8 0.0 H 0.5 U 0.5 U 0.5 U 0.5 U 0.5
04 (3) ARJOE LOANZON 2230 wbwb-- 9 1.0 7 1.5 6 2.0 8 2.5 H 3.0 H 3.5
05 (3) WALTER J SHIPMAN 2217 -bwbwb H 0.5 12 1.5 7 1.5 9 2.0 10 3.0 1 3.5
06 (3) YIAN LIOU 2211 bwbbww 10 1.0 8 1.5 4 2.0 11 3.0 1 3.5 7 4.0
07 (3) EVAN SANDBERG 2162 wwbbwb 14 1.0 4 1.5 5 2.5 1 2.5 8 3.0 6 3.5
08 (3) DANA N MACKENZIE 2139 wbwwbw 3 1.0 6 1.5 9 2.0 4 2.5 7 3.0 2 3.0
09 (3) SAMUEL SEVIAN 2133 bwbwb- 4 0.0 10 1.0 8 1.5 5 2.0 2 2.0 H 2.5
10 (3) LUKE HAR-VELLOTTI 2076 wb-wbb 6 0.0 9 0.0 B 1.0 14 2.0 5 2.0 12 3.0
11 (2) KYLE SHIN 2052 wbbwwb 1 0.0 13 1.0 2 2.0 6 2.0 12 2.5 14 3.0
12 (3) JACK QIJIE ZHU 1999 -wbwbw B 1.0 5 1.0 14 1.5 2 1.5 11 2.0 10 2.0
13 (2) JEFF MCCANN 1968 bwwbb- 2 0.0 11 0.0 1 0.5 15 1.0 14 1.0 B 2.0
14 (3) DANIEL LIU 1940 b-wbww 7 0.0 H 0.5 12 1.0 10 1.0 13 2.0 11 2.5
15 (-) THEODORE BIYIASAS 1813 —w-- H 0.5 H 1.0 U 1.0 13 1.5 U 1.5 U 1.5
[/code][/size]
Bill Smythe
EDITED to correct the 3-day and 2-day, which I had backwards.
I came up with another possible set of alternate pairings, including colors:
DeGuzman - Shipman
Liou - Sandberg
Harmon-Vellotti - Aigner
Zhu - MacKenzie
Liu - Shin
These pairings seem to avoid all the problems. You don’t have to play the same opponent twice, there are no +2 color conflicts or 3 in a row, and the tournament leader has his preferred pairing, assuming that means he is paired against the highest-scoring, highest-rated opponent he has not already faced.
Of course, I may have made a huge mistake somewhere.
Bill Smythe
First, I think you have the 2 and 3 days schedules reversed—see Aigner’s description of his results in the 3 day and 2 day sections. (Most players were in the 3 day, which is why the pairings in rounds 2 and 3 of the 2 day are so weird—De Guzman and Aigner play in round 2).
Second, your alternative pairings are wrong as you have 2.0’s mixed in with the 3.0’s. This looks like the only way to pair the 3’s and up without duplicating either a main run or re-entered pairing.
[code]Test Reentry 3-Day - Round 6 Pairing Sheet, Page 1
Board White Black
-
___ De Guzman, Ricardo (4.0,2443) ___ Mackenzie, Dana N (3.0,2139)
-
___ Liou, Yian (3.5,2211) ___ Sandberg, Evan (3.0,2162)
-
___ Shipman, Walter (3.0,2217) ___ Aigner, Michael-RE (3.0,2234)
-
___ Liu, Daniel (2.0,1940) ___ Shin, Kyle (2.5,2052)
-
___ Zhu, Jack Qijie (2.0,1999) ___ Harmon-Vellotti, L (2.0,2076)
McCann, Jeff (1.0,1968) Please Wait
[/code]
One understands the wiggle room in the rule book phrasing, since you might have to double downfloat tournament leaders if it was an absolute prohibition (and, of course, rules lawyers might argue that in a really tight situation, you should duplicate a main tournament pairing before a re-entry if there is an explicit rule that you must not repeat a pairing involving a re-entry).
If you’re going to drop two players from the 3.0 score group, then why not pair DeGuzman vs Shipman, Liou vs Aigner and drop the lowest two Sandberg and Mackenzie?
As already mentioned, DeGuzman and Aigner played in the rapid games of the 2-day schedule.
Michael Aigner
Which means that you played under your current entry, so the special re-entry rules don’t come into play. The only two players at 3.0 and above that DeGuzman can play under the standard Swiss rules are Shipman (which requires that you then would play MacKenzie) or MacKenzie. Given that there is a pairing of the 3.0’s and above which avoids a re-pair with the re-entry, that should have been chosen (in my opinion). Switching Shipman (the highest ranked 3 available to DeGuzman) for MacKenzie is around an 80 point swap (I assume Bill is taking the MSA ratings, which wouldn’t be the ratings under which it’s played, but probably will be close). One would do that to fix a flipping alternation error, so doing it to avoid a (re-entry) re-pair seems like a no-brainer. Note that the wording in the rule book is a bit too rigid:
First, what does “to equalize or alternate colors” actually mean? If the player being downfloated is due Black and the highest ranked player is due Black, but the second highest ranked player is due White, this sounds like you should upfloat #2. But what if the next score group down has an excess of Blacks? That just shifts the bad color down to the next score group.
Also, you obviously should change the upfloats for lots of other reasons, which the examples below discuss, but the rule (as written) doesn’t.
You’re right – thank you. I’ve edited that post to correct the errors.
More on the other points later (maybe).
Bill Smythe
Nope. As I said at the top of that post, I actually looked up the February 2010 ratings of all the players, and used those in my crosstable.
Bill Smythe
But if those are the ratings under which the tournament was played, the wrong person gets the bye in round one of the 3-day schedule and there was at least one other pairing in a later round that wasn’t right. So while those are close, they aren’t the ratings used.
It would be nice if a new upload format included the rating that was used by the event for pairing/prize eligibility purposes, which is often not the same as someone’s official (or unofficial) rating at the time of the event.
There may still be other reasons why attempts to recreate pairings fail, though.