We are pairing the second round. There are 17 players in the section, but three are taking a half point bye for the second round These are the players to pair (each row is pairing number, rating, first round color, first round opponent):
1.0 score group:
1 2470 W 9
2 2340 B 11
3 2278 W 12
4 2246 B 13
0.5 score group:
5 2199 ---- (half point bye in round 1)
7 2088 B 15
8 2016 W 16
10 1954 ---- (half point bye in round 1)
15 1854 W 7
16 1831 B 8
17 1780 ---- (half point bye in round 1)
0.0 score group:
9 1961 B 1
11 1939 W 2
13 1896 W 4
(Players 6, 12, and 14 are taking half point byes for the round.)
OK, how would you pair the second round? (Suboptimal answers: “choose ‘pair next round…’ from the SwissSys ‘pairings’ menu” and “choose ‘pair a round…’ from the WinTD ‘section’ menu”. )
My intuition said I should keep the 5-10 natural pairing as spliting that pairing does not help with color allocation in any way. Then use a 200-pt interchange between players 8 and 16 to improve color equalization. With the equalization transposition of players 3 and 4, the resultant pairing is:
4 1
2 3
5 10
7 8
16 15
17 11
9 13
But I guess because you posted, I wasnt satisfied. I could see some eyebrows being raised at the 1831 vs 1854 pairing (16 v 15) and players 7 and 8 also frowning. I found that dropping an alternate player (transposing 16 and 17) was a smaller transposition than above. Then I could look ahead from that first natural pairing (5 vs 10), see that I will have a color problem, and transpose 10 and 15 (100 pt differential). The resulting pairing is:
4 1
2 3
5 15
7 10
17 8
16 11
9 13
So… in option 1, I performed a 185 point interchange that produced some weird looking pairings ratings-wise, but no color problems.
In option 2, I performed two transpositions (51 and 100 point) to also arrive at no color problems.
If doing this by hand - I’d go with option 2 if I had enough time to think it through. If I was doing this with round 2 waiting to start - i probably would have gone with my gut.
I would set aside the bye players but look ahead to see if there are potential color problems in the third round once they are reinserted.
4 v 1
2 v 3
5 v 10 Neither one is due a color as each had a half point bye bye.
7 v 15
16 v 8
Player 17 drops to the “0” group. He has no due color as he had a half point bye
9 v 13
17 v 11
No one in this group received a color twice in a row. It was possible if 9 v 17 and 13 v 11
After looking at the pairings, I would project ahead likely results and consider out of curiosity what the situation will look like in round 3 once the second round byes were inserted. If there looked to be major pairing problems, I might tweak a pairing in a lower score group to see if that solves anything, but generally I would be more likely to let them stand and deal with color problems as they arose.
If you are locked into dropping 17 instead of 16 then an 8/10 62-point interchange followed by a 5/7 111-point transposition fixes colors (at the cost of a 7-8 pairing) while a 16->10->15 transposition (123 points) retains the top and bottom halves with no re-matches at the cost of the 8-15 color conflict.
The 51-point 16/17 transposition (to determine the dropped player) and the 100-point 10/15 transposition (to avoid repeating an opponent without introducing color conflicts) seemed less expensive than living with a 7-8 pairing.
I believe you’re scoring the interchange wrong in the first case. It should be 162—you can’t flip flop players whom you don’t pair and then further swap once you’ve done that. (You can, after all, produce any arrangement of the score group with a sequence of adjacent swaps). If 7 plays 8 rather than the natural 15, that’s 162.
I’ll do as some of the rest of you have done, and take a fresh look at this without first looking extensively at the other replies.
It seems undesirable to pair either 7 or 8 vs either 15 or 16, because of bad colors or because they’ve already played each other. The interchange (switching 8 with 15) doesn’t look all that bad – 162 points – but it shouldn’t be necessary.
So, of the four “played” players – those who actually played in round 1 – pair three of them against the “unplayed” players (those who had a half-point bye in round 1), and downfloat the fourth. This gives us:
5 vs 15
7 vs 10
17 vs 8
– and downfloat 16. For colors, pair 16 vs 11, and 9 vs 13.
In addition to getting all colors right and avoiding a half-half interchange, it also avoids two other things that make me a little queasy – pairing two players both of whom had byes in round 1, and downfloating another player who also had a bye. These are just a little too close (for comfort) to pairing two unrateds and downfloating another unrated. (Yes, I realize it’s not quite the same thing, but if you substitute “unplayed” for “unrated” it’s not all that different, either.)
Is this what anybody else did? Is it what everybody else did? I guess I’ll go back now and look at the other posts.
I should point out that the above pairing downfloats a player other than the lowest, and upfloats a player other than the highest. But this bothers me not one whit.
OK, now I’ve looked at the other posts. It seems that great minds think alike.
However:
Before you look ahead to that curve in the road a mile away, you might want to observe the stop sign immediately in front of you. You paired four players – 7,8,15,16 – against the same opponents they had just faced in round 1.
Whoever invented the phrase “look-ahead pairings” probably didn’t have in mind looking ahead all the way to the next round. They probably meant just looking ahead in the current score group, and possibly into the following score group, to see whether the “natural” transposition that cries aloud to be made near the top of the current score group would actually improve anything overall.
On the other hand, I’m not opposed to looking ahead to the next round, especially in small sections (under 20 players), and especially if several byes are involved. In fact, I have espoused this idea myself, in some older threads. As an extreme example, if you don’t look ahead to the next round, you might fall into the infamous 6-player trap in round 4.
On the third hand, I’ve never been a fan of trying too hard to alternate colors in round 3. In a small section, if you use up too many of the relatively scarce “good color” pairings in round 3, you could end up with many bad equalizations in round 4. Save the good-color pairings for the even-numbered rounds. Bad equalizations are worse than bad alternations.
It’s a funny thing, but computer-age TD’s think pairing large tournaments by hand is daunting. I always found it easier to pair large tournaments because there were more opportunities to swap players as needed.
I agree. Human TDs are worse at pairing small tournaments than they are at pairing large tournaments. Then again, so are computers – sometimes, at least.
I treated this as a test, and made the same mistake Bob made. (There are some things computers will never get wrong .) When I fixed the mistake, I ended up with Jeff’s pairings. The bye players helped in optimizing the colors.
That was Jack “narceleb’s” mistake, not mine. Actually I have seen computers pair people who have already played. I’ve also seen a computer pair the same person twice in the same section. Software writers are only human.
Already played - valid if re-entries are allowed and both re-entered since the earlier meeting.
Already played - valid in a double round robin
Already played - valid in a small section with high ratio of rounds to players (mandatory when rounds exceed players)
Double pairing - invalid but would sometimes occur when two cards were made for one person (some software programs give a warning when there are multiple entries for a person).
Old saying - to err is human but to really mess things up you need a computer.
Yes, but this wasn’t one of those cases. An earlier version of SwissSys could have produced a pairing which avoided the rematch but didn’t.
This was at the 2015 New Hampshire Open. I didn’t enter the player twice. Apparently SwissSys got confused about some pairing situation and paired the same player on two boards in the same section. I’ve upgraded to a later version since then. I’ll report the problem if I see it with the latest version.
)
.) When I fixed the mistake, I ended up with Jeff’s pairings. The bye players helped in optimizing the colors.