Sections in Robot Tournaments

[boni58]

I ve read Artk78 and I can only agree!

[uday]

OK, I'm convinced.

 

I'm going to change the way it works as follows

 

1. Ensure that no section is less than 15 tables ( this is the point at which section award max out at 0.90 )

 

This won't address all the issues I've been hearing about but it will address most of them.

This isn't ideal from the perspective of maximizing total masterpoints but I've come to believe that this is what we want as a group. And it isn't so bad.

 

 

2. Stratify the sections. This is unrelated but was always in the works. Along the way we'll balance the sections by masterpoint holdings.

 

 

The upshot will be that it will be no worse than before most of the time, and will sometimes be better.

 

Sound good? We'll be done with the development in a day or so. I don't like fiddling with the system on weekends, so I'll probably push the change out early next week.

 

3. We'll continue to seek the ability to issue overalls. When we achieve that, I still think we should go back to smaller sections + overalls, but history suggests that this won't happen for a while, so let's not worry about that.

 

 

Thanks to everyone who twisted my arm with equal amounts of passion and logic to effect this change. It is nice to be reminded that my wife is wrong, and that I occasionally need to make a U-turn once in a while.

 

U

uday@

[ArtK78]

Is there any reason to keep the limit on the number of players in a Robot tournament at 48?

[uday]

No reason at all. We dont usually hit more than that. I'll bump it up along the way.

[bkenner]

2. Stratify the sections. Along the way we'll balance the sections by masterpoint holdings.

 

Based upon BBO MP or ACBL MP?

[glaurie]

I hadn't made up my mind about the new scoring, however after reading all these comments, I now see both side of the issue. However, I do feel that there are now too many types of ACBL tournaments. 12 board, 18 board, human declare, no human declare, match point, IMP. I just signed in and saw each of these six options offered in the next half hour, and there were between zero and five people registered for each. I don't want to play without a field of opponents!

Glaurie

[johnu]

Since the only reason for sections seems to be to award more masterpoints, the fairest method would be to just sort the players by their final scores, and then for 2 sections, place every other player into a different section. For 3 sections, place each group of 3 consecutive players into different sections (ie the top 3 finishers would go into 3 different sections, the next 3 finishers would go into 3 different sections). Adjustments could be made for flights A, B and C to go into different sections for additional balancing.

 

If you are going to determine section placement before play starts, total ACBL points seems like a bad method to seed players. Maybe total ACBL points could be used for provisional placing for newbies, but I think some kind of seeding based on average IMP game, average MP per game, or overall average placing would be best

[GreenMan]

The changes uday describes seem to go in the right direction.

 

The problem with the old-new system was that although the inequities would in theory balance out in the long run, we don't play tournaments in the long run. We play them one at a time. If we see inequities in the standings every time, they don't seem to balance out, they seem to go on forever. Hopefully the new-new system will lessen the downside considerably.

[barmar]

Since the only reason for sections seems to be to award more masterpoints, the fairest method would be to just sort the players by their final scores, and then for 2 sections, place every other player into a different section. For 3 sections, place each group of 3 consecutive players into different sections (ie the top 3 finishers would go into 3 different sections, the next 3 finishers would go into 3 different sections). Adjustments could be made for flights A, B and C to go into different sections for additional balancing.

Hmm, this makes sense to me. In the pair games we need to assign sections ahead of time, because it's used for the movement. But there's no movement of players in robot games, so sections are pretty arbitrary.

 

But I wonder if the ACBL would see it that way.

 

If you are going to determine section placement before play starts, total ACBL points seems like a bad method to seed players. Maybe total ACBL points could be used for provisional placing for newbies, but I think some kind of seeding based on average IMP game, average MP per game, or overall average placing would be best

We don't currently keep statistics like that, although we obviously could. But the problem with this is that it only reflects your play on BBO, not the rest of your ACBL playing experience. If you have 10,000 ACBL masterpoints, but only recently joined BBO, should we really ignore the masterpoints?

[TylerE]

Since the only reason for sections seems to be to award more masterpoints, the fairest method would be to just sort the players by their final scores, and then for 2 sections, place every other player into a different section.

 

That's not sections, that's printing masterpoints.

[johnu]

That's not sections, that's printing masterpoints.

 

I see you understand the problem :lol: Without overall awards, the skewed results in section awards because almost all the top scores end up being in the same section brings a randomness and fairness perception problem to the awards.

[johnu]

Hmm, this makes sense to me. In the pair games we need to assign sections ahead of time, because it's used for the movement. But there's no movement of players in robot games, so sections are pretty arbitrary.

 

But I wonder if the ACBL would see it that way.

 

 

We don't currently keep statistics like that, although we obviously could. But the problem with this is that it only reflects your play on BBO, not the rest of your ACBL playing experience. If you have 10,000 ACBL masterpoints, but only recently joined BBO, should we really ignore the masterpoints?

 

In my defense, I did suggest using ACBL masterpoints as a temporary measure for BBO newbies. On reflection, you can't eliminate the randomness of section awards if you predetermine sections before play begins, no matter how sophisticated your seeding method may be, but having maximum overall awards and smaller section awards would go a long way on the perceived fairness issue.

[itryhardr]

Again, same thing most times I play.. I was 9th out of 42 or 43 players, got no masterpoints, but three people in another section got points, with percentages far less than mine. If this cotinues, I will no longer support BBO. It's no fun.

 

As stated above.. why not just pull out of a hat those who get masterpoints.

[dobiefan1]

I am finding that my percentages have sometimes been higher than those in the other section, yet I am getting a lesser award. I do not really want to hear that it will balance out in the long run. Also, I thought through the discussion that there would only be sections if there were no less than 16 per section. I played in one last night with 12 in one section and 14 in another. I am starting to prefer the smaller tournaments where I know there will only be one section. I may not get points, but at least I know that it has been fair. I agree that there should be overall awards.

[ArtK78]

I have this to say to those complaining about the randomness of the awarding of minor section awards. Grow up!

 

This kind of thing happens all of the time in tournaments. In fact, if it didn't happen, it would be statistically improbable (approching impossible as the number of events approaches infinity).

 

That is not to say that it is not right to point out the randomness of section awards. That has been accomplished. Now live with the result.

[dobiefan1]

Let's not get nasty and personal, Art.

[Leo LaSota]

One other flaw with the current system is that you place players with a star all above players without a star. This distorts the separation of players at the top. This means that a Grand Life Master with 20,000 masterpoints that has not asked for a star will be seeded below anyone else that enters a game that has a star. I also find it strange that BBO still cannot assign a player both a star and a ranking. I have the "A" next to my profile, but do not have the star also. This means that when I enter a game with others that have a star, I am placed below them when determining the sections.

 

 

 

Hmm, this makes sense to me. In the pair games we need to assign sections ahead of time, because it's used for the movement. But there's no movement of players in robot games, so sections are pretty arbitrary.

 

But I wonder if the ACBL would see it that way.

 

 

We don't currently keep statistics like that, although we obviously could. But the problem with this is that it only reflects your play on BBO, not the rest of your ACBL playing experience. If you have 10,000 ACBL masterpoints, but only recently joined BBO, should we really ignore the masterpoints?

[barmar]

I think the issue with star players is that many of them don't play in lots of BBO masterpoint tournaments. And if they're not American, they also may not have lots of ACBL masterpoints. But if someone has won a bunch of European championships, he deserves to be seeded highly.

 

In our user database, we do track being a star separately from the masterpoint ranking level. I think it was just felt that showing both would be redundant, and waste valuable screen space (maybe not so bad in the pop-up profile, but where would you show both in the player lists?). And kind of disingenuous -- there would be many players with a star and a masterpoint level like [2] or [3] because they don't play in our pay tourneys often.

[itryhardr]

Art, above, mentioned how all local tournaments have sections, which is true. But in tournaments, they make sure that the number of A, B, and C players in each section are equal. I don't see them doing it here... I've been in sections with all highly ranked players, while the other section had few ranked players, thus my gripe (which rude Art obviously didn't realize when he told all of those not happy with sections to "grow up"). It appeared to me once, when I kept track of the order in which players signed on, that the sections are set AS the players sign on (basically the first 15 to join are in one section, the next 15 in another section, etc.). Maybe it was just a coincidence. But I'm still not happy with sections as set up now by BBO.

[ArtK78]

The division of robot games into sections provides more opportunities (not necessarily equal opportunities) for players to earn masterpoints. And, as barmar has mentioned, the sections are divided roughly equally based on the strength of the players in the field. If I understood barmar correctly, the division is based on the players' rankings in BBO - Star, A, K, Q, J, 10, etc. That may or may not correspond to the players' ranking in the strata A/B/C. I totally endorse any attempt to property distribute the players among the sections by ability on a roughly equal basis. Whether the method being used is the best available is not really the issue. It is a reasonable method.

 

As for my comment "Grow up," I stand by it. You may consider it rude, but you are getting a benefit from the division of players into sections. Quibbling with the details or how you were placed into a section of better players is kind of silly. I admit that I often go back after the results were posted to see how my score would have done in the other section(s), but that is just out of curiosity. I don't look to see if I was in the stronger or weaker section. The only cure for that would be to eliminate sections and play as one section.

 

Besides, I have my own opinions about who the strong players are and who the weak players are, having played with a number of them in pair events and, when I finish a robot game early and have an opportunity to kibbitz the remaining players, by observing their play. And my opinions do not necessarily correspond to the BBO designations assigned to them.

 

By the way, I can see how you might interpret my "grow up" remark as nasty, but personal? I didn't mention any names.

 

Brings back memories of the days when there were separate smoking and non-smoking events at regional tournaments. There were more than a few players who tried to determine whether the smoking events were stronger than the non-smoking events, or vice-versa, and would enter the event that they perceived was the weaker of the two.

[Leo LaSota]

I think the issue with star players is that many of them don't play in lots of BBO masterpoint tournaments. And if they're not American, they also may not have lots of ACBL masterpoints. But if someone has won a bunch of European championships, he deserves to be seeded highly.

 

In our user database, we do track being a star separately from the masterpoint ranking level. I think it was just felt that showing both would be redundant, and waste valuable screen space (maybe not so bad in the pop-up profile, but where would you show both in the player lists?). And kind of disingenuous -- there would be many players with a star and a masterpoint level like [2] or [3] because they don't play in our pay tourneys often.

 

 

Wouldn't it be simple to sort players in the ACBL robot tournaments by their averages in previous ACBL robot games? At least after a player has played a suitable minimum, this would be a good way to assign the players to different sections. The fact that Justin Lall uses an id in the ACBL robot tournaments that does not have a star designation does not mean that he should be placed below all of the stars. I think that he said he averaged close to 68% in 360 hands over 2 days recently.

[runewell]

I played a robot tourney against my friend adelbe. The results were surprising... :blink: I'm not against the new stratification since there simply aren't overall awards, but it would seem fairer if jccasper and SandraGeb could have each shared 0.90 and Arnie and I could have each shared 0.63. Oh well :D

 

Name Score ( % ) Rank

A B C

Prize Points

jccasper 65.27 1 0.90

SandraGeb 64.65 2 0.63

adelbe 61.60 3 0.45

 

Section 2

Name Score ( % ) Rank

A B C

Prize Points

runewell 59.46 1 0.90

chablis314 53.81 2 1 1 0.54

patosa 53.80 2 1 0.54

Edited July 21, 2013 by runewell

[johnu]

I don't understand the points awarded for flight winners. IIRC, In one tournament summary I saw, the winner of flight B came in third in section, about 5th overall, and won .90 masterpoints, same as the flight A winner. 2nd place in B got the same award as the 2nd place flight A player.

[barmar]

As long as there are at least 15 players in a stratum (including the lower strats) in a section, placing in that strat pays out the maximum award. So if there are 15 in B+C, and 25 in A, you get 0.90 for coming in first in A or B.

 

See the links to ACBL's masterpoint awarding rules I posted in http://www.bridgebase.com/forums/topic/61266-where-is-the-detail-published/

 

If we get permission to issue overall awards, that should improve things, since you get the max of your overall and section awards, and overalls pay more.

[dwar0123]

There is something icky about dividing into sections after the fact, A better solution might be to not score across sections. There are certainly reasons to score across sections when there are overall awards, as you are actually competing against everyone. But they don't apply here.

 

Make each section their own little game, that way comparing percentages across different sections is far less valid.

 

15 minimum per section, stratification, remove the 48 limit and no scoring across sections = everyone wins!