Whenever I play in a tournament, I always check my own pairings. I always make sure that whatever pairing I’m given makes sense – whether it’s a human or program that generates the pairings. I think it’s fair to expect TDs to at least give the pairings a quick review to make sure they look reasonable.
In defense of some of the more experienced TDs, many of them SAY “the program did it and it’s impartial” just to cut off discussion and show there’s no favoritism. This is especially true when dealing with parents (who have no clue how pairings are done) at scholastic events. When some irate parent is demanding an explanation of why his (or her) son has been given black for the 4th time in the 6th round (which is supposed to start RIGHT NOW), it’s easier to say that than to teach the parent how pairings are supposed to be done.
For the TDs that DON’T do some checking and have no experience pairing by hand, what are they going to do if the computer crashes?
“Should TDs HAVE to manually check computer pairings?” I think that they should check them, but I wouldn’t go so far as to say they are REQUIRED to. Nowadays, the pairing software is standard and the new TDs, quite expectedly (is that a word? ), are learning their job by using them. The new TD jumps right in and lets the computer start spitting out pairings. The rules haven’t mandated that new TDs get experience with pairing cards. Perhaps there aren’t enough documented cases where the pairing program gives bad pairings. I’m really not an expert in the computer pairing field. WinTD, for example, will show a “Pairing Logic” for one to look at, but I’ll be darned if I can make heads or tails of it. As far as Pairing Logic is concerned, I liked PairPlus. I still like PP, but it’s no longer updated.
Even if you never plan on pairing a tournament by hand, you still need to know how to do so. I’ve found that once people realize I can pair by hand and can review and explain a computer’s pairing, they are more likely to accept an explanation like “it’s something that can happen” (a common answer when somebody asks how it is possible for them to get two blacks in a row, or how it is possible for them to win and then move to a higher-numbered board in the following round).
If you don’t know how to pair by hand then that would make it difficult to do a re-pairing if there are known no-shows and you do not want to have players sitting through a round without an opponent. Knowing how to ladder re-pair can make it easier to insert a player who’s upcoming bye was recorded incorrectly and who is really there expecting to play that round.
It is good to review at least the top boards. There aren’t that many and that is often an area that has the fewest alternatives available. If you are going to get a question it is more likely to be on those boards.
Another review I’ve been doing lately is taking a look at the bottom board of the scoregroups. There is one strange bug in some versions of some pairing programs where two players in the exact middle with the same rating can end up paired against each other (I can only guess as to why - it may be that the computer isn’t sure how to split them between the halves, so it does them separately). If it happens then I can get around it by forcing a pairing for the top board in the score group (using one of the two same-rating players), deleting the rest of the score-group’s pairings, and then re-pairing those players).
In team/individual tournaments you need to check the top and bottom boards to see whether or not teammates should be set to play each other (generally last round in the top boards and maybe the entire last half of the tournament in the bottom boards).
For local tournaments you may want to avoid intra-family pairings if there are other pairings readily available (I once did a 12-point transposition to avoid a father-son pairing).
Sometimes you are in the middle of getting a round ready and that is when you get an involved question from a knowledgeable player that will take a while to answer. In that case you either hope that you can answer it while multi-tasking or you hope that there is somebody else available to answer it.
There was one tournament where I was not directing and happened to stop by to say hi to the director (friend) while he was waiting for the pairings to finish printing so that he could post them. When a player came in with a question (from the on-line pairings) that director was happy that I was willing to explain it while he went and posted the paper pairings and linked up with the chief TD (the pairing director would have been able to explain it, but didn’t have the two minutes it took me to explain it followed by the 10-15 minutes to explain that whether or not the rule was wrong and should not have been written was immaterial since it was in the rulebook and was the rule to be used). That question happened to revolve around the new full-color-history rule concerning who got white and the difference between black followed by a bye versus a bye followed by black.
I haven’t tried using the WinTD pairing logic. So far I’ve found that pretty much any time something looks strange in WinTD it can be logically determined why it happened, and then I can make the settings change to allow the correct pairings to be automatically generated.
One that threw me for a few minutes one time was a setting to prohibit interchanges. I was doing the floor instead of the pairings and had to manually figure out what the pairings should be, at which point the person doing the pairings changed the settings to allow interchanges and redid the pairings (matching what I had done manually).
Tom Doan will be doing an update on WinTD in the next month or two to accomodate the new reporting format (so we can get color information for FIDE events), this may be a good time to contact him regarding other wish-list ideas for WinTD.
Something like this happened at a big scholastic a few years back. All the copies everywhere of the files of one of the sections were trashed. We took (a [u]PHOTOCOPY[/u] of) the wallcharts and a pair of scissors, and cut each player's line on the wallchart out, resulting in over a hundred little slips of paper. Each slip became, in effect, a "pairing card" for that player, showing his or her playing history in the tournament. The slips were sorted and shuffled into pairings (index cards do in fact make better pairing cards, it turns out), which were then written out by hand.
I forgot what happend at the end with the tiebreaks, though.
I ran out of ink on my printer at a tournament I ran a couple of years ago, I still used the computer to pair but I had to hand write out pairing sheets and standings. (A couple of people did ask me if I was pairing by hand again.)
Of course that’s better than the year I showed up at the tournament site with my computer and printer, but without the diskette that had the registration data on it (I had done that on another computer.)
A computer crashed last year when they were doing the blitz events at the National Elementary.
To make matters worse, it apparently messed things up enough that they couldn’t get an upload file for that section, either. (It was missing all the round detail.) I think Phil wound up doing the entire section over on a different computer a couple of weeks later.
So, the moral is probably that TDs should know how to pair their events by hand if that bccomes necessary, and should also know enough to check the top boards. (And always bring another ink cartridge.)
We did eventually reconstruct the tournament in the computer and were able to run normal tie breaks. We did have to do one or two rounds that way until I got caught up.
I get hives just thinking about that tournament!!!
Great improvising! Another area that should be practiced by hand are tie-breaks. If you always rely on the computer, what happens when it crashes? You have to do the tie-breaks by hand, and knowing how to do them IN ADVANCE, keeps the tournament end from being delayed too long while looking in the rulebook to learn how they’re done.
I think it’s import for TDs to check pairings, particularly in later rounds and in the score groups that are in contention for prize money. In a large section the TD may not be able to check from top to bottom, but at least make sure scores and colors are reasonable.
When I’m playing I do look at the wallcharts and try to come up with a ballpark pairing for the next round. Usually I’m pretty close, but sometimes the computer does something totally bizarre as the following illustrates.
I had a very strange pairing in the last round of the NYS Championship. In the under 2000 section I was the lowest rated player in the 2.5 score group, and I had 3 blacks and 2 whites. I was expecting to play the highest rated 2 pointer who was due black. Instead I ended out with the lowest rated 2 pointer. Not only that, my opponent had taken 2 1/2 point byes and had played black 2 times and white once. I ended out with white, equalizing my colors but leaving my opponent with a 3rd black out of 4 played rounds.
I asked the TD about it since I was very curious why I played the bottom of the lower score group. At the time I was not even aware of my opponent’s color situation. The TD looked at the pairings, and was totally baffled by what Swiss-Sys did. He said it appeared the program went out of its way to make the worst pairing possible both rating and colorwise. There were plenty of other scenarios that would have equalized colors and put me against an appropriate 2 pointer.
I got an easier pairing then I should have. Unfortunately I did not capitalize on the easier pairing and only got a draw. Had I won I would have tied for Top under 1800 and won some money.
In one of the earlier versions of WinTD I had a crash when it was trying to do board tie-breaks for a fixed-board team event. Doing them by hand was tedious, but I could start doing them while the team trophies were being awarded and that gave me enough of an edge to barely keep ahead of the board trophy awards. I later determined that two players had been transposed in a lineup in round one and that was the likely culprit. Since then I’ve avoided that issue and the problem has not cropped up again.
The title of this thread sounds like inspiration for a TD Tip! In fact it is a tradition that we TDs at the National Open observe for the last two rounds and the top score groups.
Oops! Sorry, Tim, I had just changed the title when I read your post. It used to say “have to” and now says “be expected to”. This seemed like a more appropriate way to word the title, more in line with the suggestion from another poster earlier in the thread.
I will at least check the top boards and go into WinTd pairing logic to see what is going on for an odd pairing. You do have to learn how to read the thing. Usually the oddest parings are in even rounds later in the tournament sorting out colors, but they are rarely something that one would not have done by hand.
It is a real challenge in a G30 scholastic with rounds as soon as possible and a bunch of 3rd graders waiting to play to really study the pairings in depth. I honestly spend less time on those than in our state scholastic championships when I have scheduled rounds and more time to review.
You do need to know what the different settings for pairings and administration will do in WinTd because they all sure do matter.
Manually checking computer pairings is a double-edged sword. There is at least a 50-50 chance that the TD’s revision will be worse than the computer’s original. This is especially true of less-experienced TDs, although I have seen senior TDs (and higher) fall into various traps.
Usually there is a good reason for whatever the computer did. The programs do a better job of looking ahead than humans do. A case in point was discussed in a recent thread here, when the program paired the lowest group A player against the third-highest groop B player instead of the second-highest (the highest would have resulted in bad colors). Everybody wondered why, until we finally saw the crosstable, which revealed that pairing against the second-highest would have forced another transposition elsewhere, because two tentative opponents would have already played each other.
Only a TD with a LOT of hand-pairing experience (and this doesn’t even include all NTDs) should consider overriding computer pairings. This is becoming more and more true as the pairing programs become increasingly sophisticated.
Bill
I agree. I didn’t say I overrode them. I just said I checked them so that I was clear and understood what the computer was doing, not just pushing the button adn saying - oh well thecomputer did it.
), are learning their job by using them. The new TD jumps right in and lets the computer start spitting out pairings. The rules haven’t mandated that new TDs get experience with pairing cards. Perhaps there aren’t enough documented cases where the pairing program gives bad pairings. I’m really not an expert in the computer pairing field. WinTD, for example, will show a “Pairing Logic” for one to look at, but I’ll be darned if I can make heads or tails of it. As far as Pairing Logic is concerned, I liked PairPlus. I still like PP, but it’s no longer updated.