An interesting situation just arose in round 5 of a tournament.
There are seven players to pair. Three players have a score of 3.0. Two players have a score of 2.5. One player (“Alpha”) has a score of 2.0. One player (“Bravo”) has a score of 1.5.
Bravo has already been given a full point bye. Alpha has a forfeit win because an opponent failed to appear.
Rule 28L3 reads: “28L3. Players ineligible for full-point byes. A player must not be given a full-point bye more than once, nor should one be awarded to a player who has won an unplayed game due to the opponent’s failure to appear.”
In the above situation, does the word “should” in rule 28L3 mean that the bye should be awarded to one of the players in the 2.5 score group? Or must Alpha be given the bye?
Of the solutions you propose, I think it’s better to give the bye to one of the 2.5’s.
However, there are lots of ways to avoid full-point byes. If some of these had been used in earlier rounds, maybe there wouldn’t be a problem this round.
For example, if there are only seven players, why doesn’t the TD jump in and play a game? For pairing purposes, he could give himself a “pairing score” for the unplayed games. Otherwise, the player paired against the TD might be getting an unfairly strong opponent.
Or, maybe one of the players who received an earlier bye or forfeit win might be asked if he wants to play two games simultaneously, to make up for the missed game.
How many players dropped out? The ones you list have 17.5 points total, but usually should have no more than 16 after round 4 if there were originally 7 or 8 players (three rounds with one of the eight players taking a half point bye and leaving another player to get a full point bye could do it though).
In a small Swiss you sometimes see people a half-point off the pace being in the bottom half of the active standings and thus getting a bye.
Having 87 players (16 in the top section that is under discussion) cuts down on the feasibility of the TD playing a game to eliiminate byes.
The lower rated of the 2.5-1.5 players had also had a bye, so the TD had the option of giving the bye to the 2-2 player that had already received a forfeit win, or giving it to the second highest rated player in the tournament (2382 with two prior half-point byes) who was in the top third and only a half point out of first place. The first withdrawals came in the final round, and encompassed 8 of the 16 players. A ninth had already scheduled a final round half-point bye, and had also received a full point bye earlier.
In the interest of having the leaders meet each other I would have been really tempted to also give the bye to the 2-2 player. Otherwise the 2382 gets a full point bye (to 3.5 points) and his actual 2333 opponent (which he drew) instead plays a 2070. Since the top two players drew, the 2333 winning would take clear first with a win, and otherwise the 2382’s bye would put him into a four-way tie for first.
Among the top seven players (two of whom played every round) there were 8 half-point byes, 3 full-point byes and 2 forfeit wins.
It’s too bad the U2000 section didn’t have an odd number for the 1782 1.5-2.5 to possibly play in to even out (in the event he would have been interested in it in the final round).
Correction******
If the 2382 (2.5-1.5) received the bye then his 2333 (3-1) opponent would instead play the 2019 (2.5-1.5) and the 2070 (2-2) would play the 1752 (1.5-2.5)
In either case the 2520 (3-1) plays the 2354 (3-1) [they drew]
If the 2333 then beat the 2019 instead of drawing the 2382 then he would have taken clear first while the 2382’s full point bye would have moved him into a three-way tie for second (with two of those three players each playing only two of the five rounds and getting two of their three and a half points from byes/forfeits).
I shouldn’t have rushed a response before my lunch hour ended. Mea culpa.
End Correction*******
The 9-player U1400 section was interesting. It had 9 players who took 9 half-point byes, were given 2 full-points, had 4 unplayed games and had 2 forfeit wins.
What a mess!!! It seems like almost everyone had unplayed games for various reasons. I’d be inclined to give the bye to the 2-2, or jump in the tournament for pairing purposes. The TD playing may not work well if he is much lower or higher rated then the players involved.
Assume that there were no available options for a house player. If you look at the crosstable, you’ll notice the TD is the chump in the under 1700 section in the very last row, with zero points. The TD is not amused.
Anyway, yes, it was a mess with more players missing than present for the last round. And don’t even begin with the under 1400 section. The TD is not amused about that, either.
At the chess club I used to attend, we sometimes had sections like this but made them strict in that lower players could not “play up”. In the case of an odd number in the top section, we would pair the first place player in the lower section with the last place player in the higher section. Players in the lower section were happy with this as the player in first would play someone higher rated than everyone in their section and therefore it wasn’t an easy pairing. It also helped keep the top section paired “normally.” The result of the game stood for the relevant standings in their particular section.
I could see a similar idea cascading down to various sections if the odd man problem remained. Of course this will always mean the bottom section has to have the potential full-point bye, but it also means only one full-point bye instead of potentially more if all the sections had an odd number.
Of course it becomes problematic if you allow players to “play up” and so potentially they could play someone “easier.”
Except, of course, for the section leader, who would probably howl.
But the total number of byes would be fewer. Additionally, all byes would be in the bottom section, where they are easier to handle. Games in that section tend to finish more quickly, making cross-round pairings (and other similar tricks) more feasible.
But that’s what you want, isn’t it? It would answer my first objection above.
Yikes! Your original post made me think you were conducting a small, single-section 5-round swiss with 7 players, perhaps on weeknights. The full crosstable presents a different perspective.
Still, when a TD really wants to avoid byes, he often can. He might view potential house players a with a little more hunger, pressuring them just a bit harder to join the fray. Or, he might actively ask players whether they are willing to play their next round early (cross-round pairing), or if they’d like a real challenge, such as playing two games at the same time, or if they’d really object to playing an opponent from a different section. Or, the TD could even ask himself, do I care more about the tournament, or about my personal results, and maybe do the magnanimous thing.
I don’t think Ken was particularly concerned about his score. It looks like he only played in the U1700 section to avoid giving any byes to that section. There were 16 players in the open (with a full point bye every round), 24 in the U2000 (only two full point byes for the tournament) and he made it 38 in the U1700 (no full point byes for the tournament) rather than playing in the U1400 (two full point byes for the tournament). He did not play the only round that there were otherwise an even number of players in the U1700.
I guess one option would have been for him to play in the open in the final round since he wasn’t playing in the U1700, but that makes the assumption that the five last-round withdrawals in U1700 were known in time for him to realize that was a viable option (I’m ignoring the final round mismatch that would have caused - and his available opponent was playing up to face strong competition and may have been upset at the idea of being paired down about 400 points). I also wouldn’t be surprised to learn that the open section had already started by the time the second or third repairing of the U1700 section was completed. As it turns out, the U1400 section in the final round had an even number of players paired without him.
Even with a lot of people leaving, the only bye in the final round was in the open section, so it doesn’t sound like there were any house players available (other than the opponents of the two forfeiting players - and those wouldn’t have been known at the time the pairings were done).
Jeff is much too kind with respect to my intentions. I did not play the last round because I was getting ready to drive from MA to the U.S. Open the following morning, leaving at oh-dark-thirty. In any case, the pairings SwissSys produced would have given a player rated 2391 in the open section a bye. Is it really worth that player’s time to play a 1350 or so opponent as a house player? Not to mention that the game would have been somewhat less than a thrill a minute for me, although at least it would be over quickly.
In any case, there were no house players present. It was the last round of the tournament, so it was not possible to make a cross round pairing.
I would like to ask to refocus on my original question. Without discussion being diverted by suggestions of house players, I’d like to ask whether the rules for selecting a player to receive a full point bye allow the TD to go to a higher score group.
Again, with seven players to pair, the lowest score group (1.5) comprised one player who had already been given a full point bye. The next score group (2.0) comprised one player who had not been given a full point bye, but who had had a forfeit win because of a no-show opponent. There are two players in the 2.5 score group.
Does rule 28L3 mean that I should give the full point bye to someone in the 2.5 score group, even though there is a player in the 2.0 score group who has not had a full point bye? (Again, it’s the meaning of the word “should” in rule 28L3.)
The rule says “A player must not be given a full-point bye more than once, nor should one be awarded to a player who has won an unplayed game due to the opponent’s failure to appear” (emphasis mine).
When “must” and “should” appear at opposite ends of the same rule, one suspects that the rule writer either:
A. was unsure of himself, or
B. wanted to make the “should” prohibition a little less absolute than the “must” prohibition, or
C. did not want to paint the TD into a corner too strongly, or
D. realized there may be unforeseen circumstances where the “should” provision ought not to apply, or
E. all of the above.
In this case, I’d say the rule strongly encourages (but does not absolutely require) the TD to give the bye to a player with 2.5.
When the number of players in a section is both odd and small, stuff like this happens. Eventually the byes go up to players in the top half. It’s a groaner for TDs and rulemakers alike.
Exactly. We once had a bizarre situation in a small open section where the two leaders played in the fourth round. The winner, who was rather ill and had a long trip home, elected to sit out round 5. The loser got the bye. Thus they finished tied for first.
That’s why it makes no sense to have too many sections if attendance won’t support it. In the case we’re discussing, it was the withdrawals that messed it up, but it can get awkward even without them.
And don’t even get me started on what happens when you allow re-entries in that setting. We once had a small class section we’re the top 2 finishers did not officially play each other. They met in round 1 and the loser re-entered. His new identity couldn’t be paired with the other fellow.
For a moment, you made me doubt my faith in re-entries, a device I’ve always believed in (even though I’ve never re-entered as a player).
Then, on second thought, I realized that this was really a small-section problem rather than a re-entry problem.
But, on third thought, it dawned on me that exactly the same thing could have happened in a larger section. If a first-round loser re-enters, his loss doesn’t count and he doesn’t have to face that opponent again.
Next, on fourth thought, I figured that, in a larger section, even if the player doesn’t have to face the same opponent again, he’ll have to face some other strong opponents if he is doing well. Thus, it’s a small-section problem after all. (In a small section, there may not be any other strong opponents available.)
Still, on fifth thought, if it is a class section, all the ratings are close together. It would be feasible for the re-entered player, even in a large section, to end up in a perfect-score tie with his first-round opponent, even if the number of rounds would normally be sufficient to prevent perfect-score ties.
Mulfish, why do you have to be such an expert at making me doubt myself?
In this section, if the higher rated player had only had 1.5 points and there was a 1-pointer with a forfeit win and no byes then I would not have had any qualms about giving the bye to the 1.5-pointer and would question any TD who did not do that.
The difficulty here is that the desire to pair the section leaders with each other runs against the desire to not give a bye to a player that had a forfeit win.
With that conflict existing, I can’t really fault a TD who gives the bye to the higher-rated player or a TD who gives the bye to the player with a forfeit win.
There is an old thread that Bill Smythe started about point-count pairings that is marginally relevent. In that there was an attempt to figure out how “bad” each variation from a perfect pairing would be (where perfect would be considered what the pairings would be if there were no color considerations and no considerations about players previously meeting). In some circumstances it is “better” to violate some pairing rules so that it is easier to obey others. As an extreme example, in a 16-player blitz tournament around round 9, you may need to pair the tournament leaders with each other even if they have already played (probably because they both already played the next half-dozen players in the standings).
Personally, I don’t approve of re-entries and never allow them in my tournaments. But if they are going to be allowed, you’ve pointed out a good reason why 28S1 should be dropped. (28S2 says that if both players have re-entered, they can be paired again, but 28S1 says that if only one has, they cannot – which, as we see above, can lead to a paradox.)
Dropping 28S1 would have the effect of allowing a player who has not re-entered to be paired against the same opponent again. In my opinion, this “cure” would be worse than the disease.
