Data issues in the new programming

Running Chess Tournaments
1

Here are a few data issue that have come up with the new programming:

  1. According to the Ratings Committee, the new formulas adjust for the number of rounds, so there is no need to break large events up for rating purposes. There might be other reasons to do so, because a 32 player round robin would take a fairly long time to run, even in 10 minute chess. (A double round robin would be even longer.)

I’m also not sure what upper limits WinTD and SwisSys will handle, I’ve asked Tom and Thad to comment on that.

  1. I ran a test to see if the current programming will ignore additional data field in the dBase files currently generated. They are ignored. That means I can add some additional fields to those files now without breaking the current programming.
    (That may not be critical since I’m still on schedule to switch over some time in September.)

The most important field to add is probably a player name field in the TDEXPORT.DBF file, but player state and rating would also help identify additional potential incorrect ID’s. I also want to ask for the time control and event dates for each section.

  1. The new programming will handle both round robin and double round robin crosstables.

  2. The validation checking in the crosstable edit program does not look for cross-round pairing. How may people actually use them? (I’ve suggested to Tom and Thad that we add fields to indicate the color by round and to denote cross-round pairing.)

  3. After some discussion with Tim Just and others, we’ve decided to accept the following TD information for experience requirements:

A. The Chief TD of the tournament plus an Assistant Chief TD.
B. The Chief TD of a section plus an Assistant Chief TD.
C. Any other assistant TD’s who worked at the tournament.

2

Mike,

Is this regarding Local TD upgrade? Language changes to the rules, etc?

3

Think it is for everyone when it is about upgrade, as it deals with assistant tournament directors, as most club tournament directors are not on the top of the list – as it shows only the chief tournament director and the assistant, not the minor directors on the floor.

From page 244 to 246, there is zero information on the tournament report to place that information of the subcategorys for floor dutiers or backroom duties. If that information was on the tournament report form then it could be placed on the MSA.

4

I don’t know whether TDCC will revise their experience requirement rules or not, that’s not my committee.

The data that the old file structure gave us was wholly inadquate as it could only identify 2 TD’s per section. It wasn’t clear who the overall chief TD was and there was no list of additional assistants.

The list we settled on was much shorter than an earlier list, which had categories like ‘floor td’, backroom td’, co chief TD’s, co assistant chief TD’s, etc.

This won’t enable them to completely map someone’s experience, but I think that would be overkill anyway.

5

I assume you are referring to the fact that such a tournament may be played over several weeks’ time, and that it may be desirable to submit part of the tournament after a few weeks, another part after a few more weeks, etc, so that the ratings will be more current.

I don’t see how this is a data issue. The organizer would simply have to submit partial results every so often, which USCF could handle as a complete tournament each time.

In the case of a round robin or double round robin, of course, the partial-results solution would be trickier for the organizer. He might have to convert the crosstable to Swiss format. Or, he could submit it in round robin format with a bunch of “holes” in it for games not yet played. But in that case, when he submits the second report, he would have to convert the former holes to non-holes, AND vice versa.

Still, I don’t see how this is a data issue, nor do I see a solution. I suppose it would be nice to be able (at the USCF end) to “glue” the two (or more) reports together eventually for display purposes, but this sounds like a huge hassle, and would probably be a little-used feature.

I think the limit is about 12 or 14 rounds, but that’s not your problem, is it? If the authors of these programs don’t remove these restrictions, organizers will just have to find another way to submit long tournaments. (I assume there will soon be an on-line, semi-manual method of submitting tournaments through the USCF website.)

How about expiration date also? Even though this changes every year, it could still be helpful. Also, how about both regular and quick ratings, regardless of which tournament type is being reported?

Hallelujah !!

Now, what happens when somebody wants to submit a triple round robin? :slight_smile:

I think almost everybody uses cross-round pairings, at least for some of the lower-rated players in the larger tournaments.

I have written mickey-mouse crosstable-checking programs, and can handle cross-round pairings with no special markings attached.

What you can do is this: First go through the crosstable looking for matches. When you find that player X won from player Y in round N, look for its mate, player Y losing to player X in round N. If the match is found, mark BOTH table entries in some way as already having been matched. (Later in the search, when you come to an entry that has already been matched, just ignore it and go on to the next one.)

After going through the entire crosstable in this way, go through it a second time, skipping any entries that are already marked as matched. When you find an unmatched entry showing player X winning from player Y in round N, look for a mate, player Y losing to player X in ANY round (look only among those entries not already marked as matched). If the match is found, mark both table entries as matched, just as before. (If you like, you can mark it with a different color pen, so to speak, so that you can later list the cross-round pairings as warnings on a separate error report.)

This would be good. Even though color assignments don’t affect ratings, they provide historical information of considerable interest.

Good work, Mike! (just as I expected, of course).

Bill Smythe

6

Bill Smythe wrote:

I guess I don’t see how the expiration date would be of much help, especially since’ it would be coming from USCF records anyway and our records are more current than yours. Does it help narrow down which Michael Smith is playing? I don’t think that helps nearly as much as things like state or rating. And I don’t see much value in reporting both ratings, one should be more than sufficient, especially since I’m likely only going to use it as the basis for a range check since it too could be out of date by the time it gets to us.

Swiss-Sys’s player data box holds bye data up to round 14, so I would guess it can handle at least 14 round events.

The length of an event in terms of the number of days may be more important than its length in rounds, though for a MAJOR event (think AVRO) the players aren’t going to be in a lot of side events, so there aren’t going to be problems with them playing in multiple events that overlap each other. I would imagine that Jay Bonin, who plays about 750 rated games a year, has a LOT of overlapping events, though.

The ‘edit’ program could be used to enter an event via the web, and is likely going to be similar to the program that the office will use for entering events submitted on paper. Whether it will be as fast as the current data entry programis unclear, probably not because it has some limitations imposed upon it because it is web-browser based. But I can enter a quad in under a minute, including ID’s.

The upload and edit programs are in alpha test this weekend. One of the testers wants to know when an event can be submitted and rated using it, though I thought I made it clear that’s still 2-3 weeks away.

The good news is that no major problems have surfaced yet (in less than a day of testing), and that may mean I can move on to another task on my list. (I’m trying to advance several aspects of the overalll project at the same time, kind of like a simulaneous kingside and queenside pawn storm.)

7

Hey Bill:

That is a great idea for the club tournament director. Send in the round robin in a swiss format for half the tournament and the other half in a swiss format for the other half. That would be two swiss tournaments and double the entries.

Wounder if Terry Winchester is going to use the words “FAKE TOURNAMENT”.

8

I would think it could, if there is a Michael Smith with the same expiration date (or exactly a year different), and another Michael Smith with a quite different expiration date. I agree, though, that state and rating would likely be more helpful in most cases.

Bill Smythe

9

That was my phrase, not Terry’s. Dividing a long round-robin into two halves, and putting each half on a Swiss wall chart and submitting them a few weeks apart, is certainly not fake. Up to now, in fact, that was the only way to submit a long tournament, since the old USCF software was limited to 12 rounds (I think). It was also the only way to submit a double round-robin, since the old software could handle only Swisses and single round-robins.

Bill Smythe

10

So, suppose you give me an ID for one of the 125 Michael Smith’s we have in the database. The expiration date doesn’t match ANY of them. (There are 97 different expiration dates among those 125 Michael Smith’s)

Or worse, it matches SEVERAL of them, and is exactly one or more years off from more of them.

Does this help me determine if you’ve given me the right one?

Birthdate would be a far better discriminator, we have birthdates for 109 of the Michael Smith’s, with 91 different birthdays. Only two of them were born on the same day and year and those records have the same address, so this is probably a case of duplicate IDs (one from 1991, one from 1994.)

11

Under the old system could only go up to 20 rounds. Even on a double round robin you could have section 1 and section 2: the first section for the first round robin and the second section for the second round robin. Come on Bill your a senior tournament director you can have 300 players without a assistant. Now if it is a club tournament director with 30 players in the round robin you would be right.

12

OK, I’ll take your word for it. Whether the limit was 12 or 20, the point is that tournaments longer than the limit must be divided into two tournaments.

I like your idea of splitting a double round robin into two single round robins for reporting purposes. That works when the number of games per player is between the limit and twice the limit.

Bill Smythe

13

I was thinking more of a case where the name is less common than Michael Smith. Suppose there are just two matching names, from the same state, with similar ratings (neither of them identical to the rating on the report). Here, I would think, the expiration date might be useful to separate them.

I agree, except that birthdate is unlikely to be on the tournament report, unless the record was obtained directly from USCF files, in which case the ID number would be there anyway.

Bill Smythe

14

When we get to the new data format, it will have a separate record for member data which could include things like expiration date, birthdate, FIDE rating information, etc.

In terms of what to add to the current dBase TDEXPORT.DBF file, I think it is better to keep that to the minimum that is truly helpful. Name/State/Rating should catch most of the potential ID errors that aren’t already being caught by the other checks built into the program, just the name should probably catch 95% of them.

BTW, a few TD’s have already figured out that they can use the ‘lookup’ feature of the membership module to look up ID’s and ratings ahead of their events and help resolve duplicate ID or similar name issues. They don’t have to actually submit the batch for processing, they can just have it checked and then delete it.

15

The only problem with having birthdates on the MSA – could lead to indentity theft. Could lead to someone stealing the identity of a player that just left the federation, or knowing the person that has passed on: just to use their rating for a tournament. Then again someone could know the persons social security number, use the MSA to gain information on the birthdate – then use the information too conduct criminal acts.

16

MSA does not and will not give birthdates. However, nothing should stop a TD from asking for that information, especially since it is needed to establish eligibilty for several membership types.

17

If the question becomes for the directors with age groups, for scholastic events and some states closed tournaments (like the ‘Michigan Senior Championship’) for the states senior tournament(s). Since the MSA has given information about the membership that have a life membership. The MSA could have membership types (senior, regular, youth, scholastic) on the MSA.

Some problems, as a scholastic player could have a youth membership: could make the question of youth and scholastic types a little pointless. Or have a title that only informs about a age group on the MSA. Name of a title for age groups (like 10 and under, 14 - 11, 16 - 15, 18 - 17, 21 - 18, 29 - 22, 39 - 30, 49 - 40, ect) would help the director understand if there are two players with the same name. As ‘Mike Smith from Michigan’ could be in the 39-30 age group and the other ‘Mike Smith from Michigan’ could be in the 14 - 11 age group. What the titles would be, that is up to the computer programer.

With this system, would help in not given the birthdates out to the tournament directors. Would not need to give personal information out too active tournament directors. As someone could just join the federation, then become a ‘club tournament director’ just to gain information about birthdates and other information. If the system of given personal information out to the directors, then a ‘club tournament director’ would need to become a established over-the-board player. As a person that is looking for indentity theft, would not spend the time to get a established rating just for birthdates. This is the reason that having tournament directors get information about other members birthdates, could lead to great abuse in the wrong hands.

True, having age groups on the MSA: would still have a few members with the same name, from the same state, in the same age group. With age groupings, would take a problem and make it a lot smaller problem. It is not a fail safe system, then again nothing is a fail safe system.

18

Not at all. I’ve done this many, many times myself. But not for the purposes of attempting to get around the rules.

19

The limit for SwissSys is 14 rounds. I suppose this might come up once in a while for a round robin, but I haven’t seen one that size in years. (And who needs a computer to run a round-robin?) Haven’t seen a Swiss that long since they abolished the Big Bonus.

20

Nobody, but you might want one for record-keeping, bulletin publication of crosstable, and/or rating report submission.

Bill Smythe