Pairing System: Pairing Log

The WinTD pairing log can help you to answer the inevitable questions about why the pairings fell the way they did. Check the Log Pairing Logic box in the Pairing Rules section of the Preference Dialog. When you pair a round, information will be placed into a text window entitled PAIRLGxx.TXT (where xx are digits). If you leave this window open, subsequent pairing information will be added to the window. Note that it may not be easy to explain precisely why A was paired with B, since the answer to that is often that C can't play D. It's a bit easier to come up with (at least) some explanation of why A didn't play C.

For each section, WinTD shows first the natural pairings, and then the final pairings.

**********Natural Pairings**********

****Score Group 3.0****

0 EQ OK Berry, Nicholas(1:988,W,OKTEV) vs Flax, Robert(2:1049,W,LNWEV)

Drop 0.5 Points

****Score Group 2.5****

0 OK OK Decker, Samuel(3:948,W,DWYEV) vs Sandler, Justin(4:888,B,LNCEV)

Duplicate

0 OK OK Carfang, Andrew(5:816,W,WLKSK) vs Banzhaf, Alex(6:1042,B,DWYEV)

Drop 0.5 Points Duplicate

****Score Group 2.0****

0 EQ OK Levine, Matt(7:913,BB,LNWEV) vs Rosic, Gregory Geo(9:883,B,LNWEV)

Teammate

0 OK OK Wright, Alrick K(8:885,W,LNWEV) vs Bialer, Jacob(10:852,B,ORREV)

****Score Group 1.5****

0 OK OK Burkhead, Shannon(11:808,b,DWYEV) vs Barzilai, Joshua(12:711,w,KNGEV)

Duplicate

On each line, there are three fields before the names:

1.The rating difference from natural pairings. Since these are natural pairings, these are all 0. A suffix of I means a top/bottom interchange.

2.Color problems. OK is none, EQ means an equalization failure, AL an alternation failure, 33 means three in a row.

3.Rating difference flags. OK means within all limits, LO means over the low limit (but not high limit) and HI means over the high limit.

Each line will always list the higher ranked player on the left (once pairings have been set, what color you get is determined by the Color Allocation Rules). In parentheses is much of the relevant information for pairing: the current ranking (based upon score and rating), rating, color needs (w,b=expects White(Black) for alternation, W,B=needs White (Black) for equalization, WW,BB=needs White (Black) to avoid three in a row, x = no color yet), followed by team and/or club codes if those pairing preferences are being used.

On any pairing which has more serious problems, a line follows indicating what they are. Berry vs Flax is a 0.5 drop, Decker vs Sandler is a duplicate of an earlier game, Levine-Rosic are teammates.

Following the pairings (we've omitted some of them) is a summary table

Total value 74690.285714

0003 Duplicate

0002 Teammate

0003 Wrong Equalize

0001 Three in a Row

0003 Drops for 2.000000

There were some serious problems with the natural pairings, with three duplicates and two teammates out of eleven boards. The final pairings came out as follows:

**********Final Pairings**********

****Score Group 3.0****

101 EQ LO Berry, Nicholas(1:988,W,OKTEV) vs Decker, Samuel(3:948,W,DWYEV)

Drop 0.5 Points

****Score Group 2.5****

60 OK OK Flax, Robert(2:1049,W,LNWEV) vs Sandler, Justin(4:888,B,LNCEV)

129 OK LO Carfang, Andrew(5:816,W,WLKSK) vs Levine, Matt(7:913,BB,LNWEV)

Drop 0.5 Points

****Score Group 2.0****

2 EQ OK Banzhaf, Alex(6:1042,B,DWYEV) vs Rosic, Gregory Geo(9:883,B,LNWEV)

0 OK OK Wright, Alrick K(8:885,W,LNWEV) vs Bialer, Jacob(10:852,B,ORREV)

****Score Group 1.5****

344I OK HI Burkhead, Shannon(11:808,b,DWYEV) vs Berry, Tom(14:737,WW,KNGEV)

Drop 0.5 Points

0 AL OK Barzilai, Joshua(12:711,w,KNGEV) vs Burt, Alexander R(13:984,WW,LNWEV)

Drop 0.5 Points

****Score Group 1.0****

15 OK OK Siegfriedt, Gordon(15:705,B,LNWEV) vs Marshall, Andrew(17:607,WW,WLKSK)

64I OK OK Starkman, Nick I(16:641,BB,ORREV) vs Walls, Clifford(18:592,W,DWYEV)

46 OK OK Metz, Veronica(19:545,W,DWYEV) vs Fodor, Tim(21:693,BB,WLKSK)

Drop 1.0 Points

****Score Group 0.0****

14 EQ OK Johnson, Derrick E(20:739,B,LNWEV) vs Elyacharschuster, Isaac(22:679,B,WLKSK)

Total value 1127.119781

0002 Interchange

0003 Wrong Equalize

0001 Wrong Alternate

0001 Over High Limit

0002 Over Low Limit

0005 Drops for 3.000000 from Points 9.500000

In order to get pairings which avoid duplicates and teammates, it was necessary to drop more than one player out of the 1.5 score group. Let's look closely at that score group:

****Score Group 1.5****

344I OK HI Burkhead, Shannon(11:808,b,DWYEV) vs Berry, Tom(14:737,WW,KNGEV)

Drop 0.5 Points

0 AL OK Barzilai, Joshua(12:711,w,KNGEV) vs Burt, Alexander R(13:984,WW,LNWEV)

Drop 0.5 Points

Why didn't Burkhead and Barzilai (the two 1.5's) play? Here, the answer is easy: if you check the natural pairings above, that duplicates an earlier pairing. The "cheapest" way to handle this would be to have Burkhead play the highest player in the next score group (Burt) and Barzilai the next (Berry). However, Barzilai and Berry are teammates, so those pairings are switched. The 344 point rating gap on Burkhead-Berry is 984-737 (gap in the lower group) + 808-711 (gap in the higher group). While this is a large gap, the only alternative is to pair teammates.

Let's now try to answer the more interesting question of why the lone 3.0 (Nicholas Berry) doesn't play the highest ranking 2.5 (Robert Flax). You could work through the pairings manually to answer this question, but then why did you buy a pairing program? Instead, put WinTD to work. Here's how you could answer this question. The following is not for the faint of heart, and isn't recommended for anyone who is just getting used to the program.

Make a Game Window for the paired section. Use drag-and-drop to alter the pairings to include the hypothetical matchup. (Don't worry if it complains about a duplicate on the other board involved; we're going to delete that anyway. Of course, if it warns you that the players you're trying to force pair have already played, you already have your answer). Delete all games except the one that you are checking: select by some means and use Edit>Clear to delete them. Then do Section>Pair a Round to pair the round again, but make sure that you change the round number back. WinTD will now pair the round again, but this time will make sure that the game you didn't delete isn't changed.

The pairing log will tell you the story:

**********Final Pairings**********

****Score Group 3.0****

0 EQ OK Berry, Nicholas(1:988,W,OKTEV) vs Flax, Robert(2:1049,W,LNWEV)

Drop 0.5 Points

****Score Group 2.5****

261I OK HI Decker, Samuel(3:948,W,DWYEV) vs Levine, Matt(7:913,BB,LNWEV)

Drop 0.5 Points

0I OK K Sandler, Justin(4:888,B,LNCEV) vs Carfang, Andrew(5:816,W,WLKSK)

****Score Group 2.0****

2 EQ OK Banzhaf, Alex(6:1042,B,DWYEV) vs Rosic, Gregory Geo(9:883,B,LNWEV)

0 OK OK Wright, Alrick K(8:885,W,LNWEV) vs Bialer, Jacob(10:852,B,ORREV)

Remaining pairings are the same as before. WinTD avoids this pairing because rather than splitting the changes rather equally among the three pairings in the top two groups, it loads them onto a single pairing, where Decker drops over two players in his score group and the top player in the next one. If, as we hope, you decide that the original pairings were more reasonable, just use Edit>Undo until you get back to those pairings and move on. If you like the new ones better, they're the ones on the computer right now.