Issues with an unusual tournament. 3 sets of quads.

Running Chess Tournaments
1

I’ve got access to a new venue for the fall, and I’m thinking about holding an event there. I have a room all day long, from morning till midnight. (And a skittles room as well, all at low cost.)

I like fast games. G/30, maybe G/45. Nothing longer. So I’m inclined to throw the sort of events I like. Although my normal tournaments are five round Swiss affairs, after some experience, the quad format has some significant advantages over a Swiss. However, a quad G/30 isn’t exactly a long format event. I have all day. I should use more than 3 hours of it.

My solution? Three sets of quads! Three tournaments. One day. Enter 1, 2, or 3 quads. The first quad runs 10am-1pm. The next runs 2pm-5pm. The last runs 7pm-10pm. 9 rated games of Chess in one day! Prizes (if any) are awarded for each quad.

Ok, so a lot of people would find that hideous, but they either won’t show up, or maybe they just come for one quad.

Then I got to thinking of something. Norms. I know the title system hasn’t exactly swept the USCF world by storm, and a lot of people don’t care about them, but, personally, I like them. It’s not a huge deal, but it adds something to an event for those people who pay attention. There aren’t enough rounds in a quad to award a norm.

My solution? When entering data into the rating system, enter the whole event - all three sets of quads - in one section. Now it’s a nine game event. I’ll get a huge list of warnings from the system because there will be lots of pairs of players that play more than once, and some players will have 3 games, others 6, and others 9. Nevertheless, it should come out ok.

I don’t think I’m breaking the spirit of any rules here. Everyone knows what’s going on in advance. Ratings will be computed in a similar fashion, with the exception that a perfect score in one quad might not earn the “bonus points”. It really shouldn’t anyway. A three game win streak isn’t all that impressive, and if you are really that good, you’ll win 6 or more games anyway, and your rating will increase just as much without the bonus points. As for norms, they aren’t awarded for three game events because three wins could just be a statistical fluke, as opposed to a real indicator. Performance in two or three quads is a better indicator, anyway.

So, I feel on good ground with respect to any ethical considerations. I’m not giving away rating points or titles to people who don’t deserve them, nor depriving anything to people who do deserve them. However, am I on shaky ground with respect to technicalities? I’ll be awarding prizes based on a partial event. MSA will show one event, but there were actually 3 distinct competitions, not to mention that each quad is normally a section unto itself since the players normally play only each other. I don’t know of any specific rule that would be violated, but I figured I might ask the community.

Then, there’s one more issue. Would you attend? Does this sort of event sound appealing, or is this another case where Dave Lame has once more gone just a bit too far outside the box for the Chess world?

2

If players Brown, White, Green and Blue play in the same 3 quads and you lump them together into 1 section for ratings purposes, they won’t be eligible for either bonus points or norms, because they will have played another opponent more than twice.

3

Why not also run three separate tournaments at the same time? One could be G/30 quads, another G/60 quads, another G/90 quads. Each of these would be run three times, morning, afternoon, and evening. That’s nine separate tournaments. If there are, say, five quad sections in each, that’s 45 separate quad sections.

You could also allow a player to play in all three time controls at once. He could end up playing in all nine tournaments.

Bill Smythe

4

Thanks, Mike. I wasn’t aware of the rule, but it makes sense.

I don’t see you you could run a G/90 quad three times. I am, however, thinking of running a G/90 section simulaneously with each of the G/30 sections. The G/90 players play one game while the G/30 players play 3. The biggest problem with that is I have only one “quiet room”, so the G/90 players will be disturbed by the setup of the G/30 players coming in to play their second and third games. For that reason, I’m leaning against it.

5

I was considering doing something similar here. I’ve never run a quad before. My only idea was that I was going to go G/15 or G/20.

My opinion is that either method is fine as long as you’re up front about the method. If you split the time, though, would you have to break the sections or would you treat it as a multiple-time-control section? (And never done a multi-tc section so I’m a little rusty… Would that mean that the G/30 players would get rated at G/90?)

As to participating, I was reminded of a tournament in Springfield in 2011. Three of us carpooled from up here in Bloomington down to Springfield. It was supposed to be a straight quick tourney.

It turned out that it was the weekend of the Route 66 Road & Car show, and it was only the organizer and the three of us who showed up at the downtown library. Sooo… Split into a morning and afternoon session, we played five rounds against each of the other three players, sort of a quintuple RR / quad of 15 rounds of G/10. Had a blast, got an excellent gyro and saw cars during lunch, and have wanted to play in quads ever since. Haven’t seen one advertised within driving range when I was able to play.

So yup, were I in your area I’d happily come along and make my decisions on a quad-by-quad basis - gut reflex is I’d play the first and third ones. (Though 8 or 12 rounds of G/30 in a day might make me loopy…)

6

I would not recommend holding a single or double round robin event (including a quad) where the time control is not the same in every round.

The rules at this time are:

A Blitz event must have all rounds at the same time control. (I believe one of the reasons the Rules Committee included this was the presumption that many Blitz events would be round robin events.)

A Quick-only event must have all rounds at quick-only time controls, but the time control can vary from round to tound.

A dual rated event must have all rounds at dual-ratable time controls, but the time control can vary from round to round.

A regular-only event may also have time controls that vary from round to round and may have some rounds at dual-ratable time controls and other rounds at regular-only time controls.

As a reminder, for Blitz time controls MM+SS (total time per player in minutes plus increment or delay in seconds) must be between 5 and 10, with MM at least 3.

For Quick-only time controls, MM+SS must be between 11 and 29. MM must be at least 5.

For dual-ratable time controls, MM+SS must be between 30 and 65. MM must be at least 5.

If MM+SS is greater than 65 (or if there is not a sudden death final time control), it is a regular-only time control. MM must be at least 5.

7

OK, here’s yet another (wacko) idea. Run quad sections at G/30, and as soon as four players (from any combination of quads) are finished with their 3rd game and want to play in another quad, form one right there and start playing. You could start another quad any time four players are available who want to play in it. Keep doing so all day and into the evening. Some players might get in 6, 9, 12, or even 15 games.

Bill Smythe

8

I’ve seen some Insanities run along similar lines, Bill, but that was a LONG time ago.

As I recall there was a strange prize structure where you got points based on how many quads you played in plus how well you did in them. (Something like 2 points for a clear first, 1 point for a tie for first.) But a strange prize structure in an Insanity was fairly commonplace. There was a TD in Houston who was very inventive.

These days you could run it at G/3;d2 and probably get in a couple dozen quads in an afternoon with 16 players.

9

I have advocated allowing norms from 3 round tournaments due to the many quad events there are, to increase interest in titles, and to prevent people from being undertitled. To make up for 3 round events not being as statistically significant, the requirements to earn a norm from 3 round tournaments should be sligtly higher.

10

Do you know what ‘statistically significant’ means, and why a 3 round event is not likely to qualify as statistically significant even with a much higher threshold?

11

You could argue that 4 rounds isn’t statistically significant enough to award a norm either but why are norms available from 4 round events? It’s due to the many 4 round events there are, to increase interest in titles, and to prevent people from being undertitled.

12

A minimum of 4 rounds (and the threshold of an actual score more than 1 point above expected score) were chosen after several weeks of analysis by the Ratings Committee. Norms are NOT SUPPOSED to be easy to get!

But politics often trumps logic or analysis in the USCF.

13

Methinks that Bill finds this tournament too far outside the box.

It really shouldn’t be. It’s just a quad. There just happens to be three opportunities to play in one day. If such an event were advertised in the area, not run by me, I would probably attend the 2:00 start time and the 7:00 start time. I’m not a morning person, and my brain would be fried if I tried to play from 10:00 am to 10:pm, even with breaks. Other people would welcome the chance to play a quad and be home in time to watch Michigan play football at 3:00.

Of course I haven’t even gotten to the weird part. The weird part is that one of those tables would be occupied by four players playing Shogi instead of Chess.

14

Yet another mathematical limitation against earning norms in quads is this:

In a quad, one’s opponents are predetermined, and often are in the same general ratings range.

In a Swiss, one’s opponents are not predetermined, they are a function of how well someone is doing. That means someone who is doing well (ie, winning) will face stronger opposition than someone who is not doing well. Since norms are determined by the strength of one’s opponents compared to the standard of how a 1200/1400/1600/1800/2000/2200/2400 player is expected to do against those opponents, someone who is doing well will have a stronger set of opponents and thus a greater opportunity to earn a norm.

If we set a somewhat higher bar for a quad, say, requiring a score of greater than 1.5 above the expected score, would most quads even afford the opportunity to earn a norm?

Suppose the players in a quad are all rated 1600. The expected score for all players will be the same, 1.5. Thus if the threshold is ‘must score greater than 1.5 above expected score’, it would be mathematically impossible for a 1600 player to earn a Category 2 norm from that event.

If there is some minor variance around 1600 in player ratings, then it seems likely that only the two lowest rated players in the quad would even have the opportunity to achieve a Category 2 norm at the 1.5 threshold level.

Now, if someone faces four 1600 players, then the expected score is 2.0, so a player needs to do better than 3.0 (eg, 3.5 or 4.0) to earn a Category 2 norm. The addition of the fourth game provides more opportunity to have a statistically significant result and thus earn a norm.

15

Translation - “No, I don’t know what statistically significant means.”

16

I’m reminded of a set of tournaments that the Oklahoma Chess Association did back in 2004. I don’t remember the time control(s), and they were swisses rather than quads, but take a look at these crosstables:

http://www.uschess.org/msa/XtblMain.php?200412114640
http://www.uschess.org/msa/XtblMain.php?200412114650
http://www.uschess.org/msa/XtblMain.php?200412114660

Alex Relyea

17

If you have the room until midnight, you could run two 4ss, Game 30 tournaments with a quad tournament in between the two. For example, the first tournament would have rounds at 10-11:15-12:30-1:45. The quad could run from 3-4:15-5:30 pm. The last event could have rounds at 7 pm-8:15-9:30-10:45. This would give the players choices of events, time to rest, and the harried TD the time to do pairings and have time for everybody to clean up at the end of the events. That would be 11 rounds of chess in one day for some of the players, with two chances to make norms. You could run 3 tournaments on one day as 4ss with tighter round times, but that would be pushing the sanity of the TD as much as the players.

With the above format, you could have prizes based on performance in all three tournaments, prizes based on each tournament performance, or some other way that you find appealing to the players. Also have door prizes for each event to encourage the crazies to play in all three. You could make a fortune with a concession selling food and Red Bull.

18

Moderators, this is a personal attack. Please delete this post.

19

Tempting.

I’ve decided, though, that Swiss doesn’t work so well for me. My tournaments aren’t very big, and they have a wide rating range. The result is that the first round is practically a throw-away round, with 500 points or more as a typical rating spread. I’m experimenting with pairing methods to get rid of that, but it’s hard to find the right balance and keep the essential characteristics of the Swiss.

20

Looking at your last three tournaments, it seems that you could have a swiss with two sections, say an open section and an U900 section. That would get rid of the “gross mismatch” problem. Another thing to consider is that if you have an eight round swiss, say, having one round be a “throw-away” isn’t such a big problem, not as bad as having to play three players three times each.

Alex Relyea