[adv exp]

My understanding of duplicate scoring in tourneys was you compared your score to every other pair sitting your way. If you had the highest cumulative score of all hands played by those pairs sitting the same direction, you took first place and were awarded master points accordingly. I noticed early on that there seemed to be a bonus amount given to the team that had the highest score of all teams, those sitting your way and those sitting opposite. This was a little puzzling because unless you played directly against that/those teams with higher scores sitting opposite you, you really could not do anything to affect their score nor yours relative to theirs. It only made sense to me that the top teams sitting each way should earn the same amount of master points. I wrote a letter to ACBL/bbo asking about this included below:

"I played a speedball tourney today, #8342 Pairs ACBL Thu 6pm, and placed 3rd overall out of about 100 pairs. However we were first of all E-W pairs by about 10 IMPs! Why is this not considered a first place and awarded master points accordingly. It’s not like we could have done anything about affecting the outcomes of those N-S pairs that had marginally higher scores than ours, unless we played against them for a set of three games. Could you please explain the rationale for this scoring which seems quite arbitrary to me? Thanks (aka adv exp)"

A week later I got this reply:

"Hello:

This situation occurred because you were in a section, where scores are
compared to *that* section and where you didn't n qualify to score in the
Overalls.

I'm sure you have been in face to face Regional or National tournaments where,
in your Strata you scored higher than others did in their section and those
players were able to finish in the 'points awards' and that you, even though
you scored higher didn't score any masterpoints at all. I know that this has
happened to me more times than I'd like to remember.

But please understand that this was a singular situation and the next time the
player in a different section will wonder why he/she scored higher than you and
didn't receive masterpoints, while you did.

It's really the luck of the draw as to which section you will be placed - and
the others in your section and Strata and how they played vis a vis how you
did.

I hope this explains things as to why this happens, but please know that you
can always contact me at the BBO email address, jacki@bridgebase.com if n you'd
like to discuss this in more detail.

Best regards,

Jacki
BBO Customer Support

I am more puzzled than ever! Call me naive but shouldn't the responder first acknowledge my point based on the understanding of duplicate scoring and then shouldn't any explanation given not be just "you didn't qualify" without explaining in detail why I didn't qualify! If there is some magical way that determines how a person "Qualifies" please, please, please bestow on us what that way of determining/figuring qualification is!!! Is it my gender, my experience, my cumulative master points, my cumulative master points relative to my opponents, random, the level of each section that a side is divided into, what?? If it is by section, how is the level of each section determined? When I (and I presume all others) play a competitive game of any sort, I want to know ALL the rules including and especially the scoring rules so that I will know best how to maximize my score.

If anyone out there knows the answers to these questions especially about "qualifying" I would really appreciate a response. I am tired of waiting.

adv exp

[johnu]

I'm definitely no expert on movements, but in the Mitchell movement results I have looked at, there are 2 sets of winners for N/S and E/W. Is this actually run as some sort of computer movement? Also, it seems very strange that the B/C winner can earn the same 1.5 MP that the A winner gets. Must be some quirk in the way the ACBL awards master points.

By comparison, the ACBL robot tournaments award points within each section, so the winner of each section gets the maximum number of master points.

[barmar]

Our scoring in speedballs is just like scoring in real life ACBL tournaments. There are per-section awards where you're just ranked with pairs in the same section and direction, and overall awards where all the pairs are ranked together. You get whichever award is higher.

If you have a problem with the logic of ranking NS and EW pairs together, take it up with ACBL. We have to follow their rules to be able to award their masterpoints.

[barmar]

 johnu, on 2017-January-11, 15:17, said:

I'm definitely no expert on movements, but in the Mitchell movement results I have looked at, there are 2 sets of winners for N/S and E/W. Is this actually run as some sort of computer movement? Also, it seems very strange that the B/C winner can earn the same 1.5 MP that the A winner gets. Must be some quirk in the way the ACBL awards master points.

It's just based on the number of contestants you're being compared with. If it's at least 15 pairs, the top award is 1.5 MP. So the lower strats will only get a lower top award if there aren't enough players in the event for them to have 15 pairs/players.

When BBO tourneys are large enough to be split into sections, each section is about 15 pairs, with the strats about evenly distributed. So winning the lower strats usually won't give the same award as winning A, because there will only be about 5 in C and 10 in B.

[barmar]

I just looked at one of the speedballs I played in. The players who got 1.5 MP were the top A and B players in the overalls. None of the sections were big enough to award the max MP in section awards. But when you combine sections, there are enough.

[Jacki]

This link might give you the technical data you are looking for:

http://web2.acbl.org...tion/MPBOOK.pdf

This goes into great detail about the ACBL rules for awarding masterpoints in various events. And of course, this is what we follow in issuing masterpoint awards in our games.

It is possible to come out with one overall winner in games where everyone plays everyone else. This would be a Howell movement and is typically used in clubs with a small number of tables. In our large games, running this would mean that a tournament would last until the wee small hours of the morning. And, of course, then it wouldn't be a Speedball.

[mycroft]

It would be possible to run a Mitchell where certain tables were arrow-switched each round (or a "three-quarter" (likely close to one-twentieth) Howell) in order to have a true one-winner movement with comparisons against "all" pairs. BBO already does simultaneous play, thus negating some of the need for the standard Mitchell movement (and the skip, of course); why not this? I realize it would require some coding.

Of course, the cure could be worse than the problem - can you imagine the chaos when people don't realize until trying to read the scores that they played rounds 1, 2, and 3 N-S, but 4 E-W? The calls to the Helldesk? The rants to BBF?

But yes, in games where the ACBL allows overall awards, they are to be given to the pair that scores the best, even if they didn't compare against half the field. Which is definitely an issue if by sheer luck all our line is all the weak pairs, and me and thee - which is why we try to seed the field. Sometimes your 59% is top your direction and 5th best; sometimes first place is 57%, second is 53, and 10th is 52.5.

It's not "fair" (which is why big events score you against different lines in different sessions, and have enough different sessions that it will actually settle out), but it's rarely far out. At least compared to that misplay that slipped the overtrick on 12 or the system fix that rightsided their 3NT against you on 4.

[barmar]

 mycroft, on 2017-January-11, 18:38, said:

Of course, the cure could be worse than the problem - can you imagine the chaos when people don't realize until trying to read the scores that they played rounds 1, 2, and 3 N-S, but 4 E-W? The calls to the Helldesk? The rants to BBF?

That already happens in indies, I don't think we get many screams (actually, for all I know we do -- I don't see emails to support unless they feel the need to escalate them). But our general feeling is that if the simple Mitchell movement is good enough for most f2f club games and tourneys, it's good enough for us.

BTW, it's too bad we don't actually have a physical "helpdesk" (everyone in the BBO organization works from home, there's no office) -- I'd love to enshrine your typo in a poster above their cubicles. "Wellcome to the HellDesk!"

[mycroft]

Ah, it's clear that more than some have misunderstood me. I apologize; my work experience tails with 6 months outbound market research calling followed by 3 years 18-hour on-call IT for the same research company; then 6 years on my current company's support desk (which *starts* at "Level 3 support" and goes up from there, we're kind of special) before migrating to the implementation side of things. I know from HellDesk and am not at all suggesting incompetence from BBO staff.

I use a lot of IT Jargon, usually as a joke, and it has sometimes caused issues when the audience doesn't immediately know it: I remember an argument with some hurt feelings when I referred to something as "b0rked", and the hearer not knowing that "since the default state of IT is broken, unlike the myth of the Inuit, we do in fact have "57 words for snow" - or in this case, "multiple scales of brokenness" so not getting the grading I was putting on it, but also getting the "I'm saying something that clearly I expect 'my people' to understand, deliberately to exclude you from 'my people'" vibe that I didn't intend. I see this was the same; I apologize.

It wasn't a typo, it was a deliberate play on words; I meant it as the Jargon File has recorded it for ever; I assumed that enough people had done customer support to have heard and lived that phrase before - or the rest had done enough equivalent (retail, phone banking, ...) to understand.

As an apology, can I offer the great Trolls' sublime Welcome to the Internet Helpdesk? "Please hold."

[barmar]

 mycroft, on 2017-January-12, 11:05, said:

Ah, it's clear that more than some have misunderstood me. I apologize; my work experience tails with 6 months outbound market research calling followed by 3 years 18-hour on-call IT for the same research company; then 6 years on my current company's support desk (which *starts* at "Level 3 support" and goes up from there, we're kind of special) before migrating to the implementation side of things. I know from HellDesk and am not at all suggesting incompetence from BBO staff.

I didn't think you were insulting anyone, I thought it was a typo, and an amusing one at that.

BTW, in past lives I've been level 3 support for a couple of pretty large tech firms.

[adv exp]

 Jacki, on 2017-January-11, 16:27, said:

This link might give you the technical data you are looking for:

http://web2.acbl.org...tion/MPBOOK.pdf

This goes into great detail about the ACBL rules for awarding masterpoints in various events. And of course, this is what we follow in issuing masterpoint awards in our games.

It is possible to come out with one overall winner in games where everyone plays everyone else. This would be a Howell movement and is typically used in clubs with a small number of tables. In our large games, running this would mean that a tournament would last until the wee small hours of the morning. And, of course, then it wouldn't be a Speedball.

Jackie, thanks for the reference. I did go and read all of the ACBL material and it was completely lacking in the important details of how masterpoints are awarded when there are multiple sections. I have a call/complaint in to them and I expect it will be resolved soon.
In the meantime I did a detailed analysis of the whole tournament results in question and tried to determine your method for assigning master points and I think I may have figured it out.
It appears that what is done when there are multiple sections is that bbo (or bbo’s computer program) distributes teams such that each section has relatively equal distributions of talent/experience/expertise/ability, from very experienced all the way down. I would presume that bbo uses player master points (the big number after a person’s bbo name) as their measure of ability to evenly distribute talent among the sections. Already I have a big problem with this, but I will save that for later. Players are assigned a level based on how big their number is and I believe there are 3 levels/flights. Notice that I keep using terms like “it appears”, “I believe” etc. because despite my specific request for these kinds of details from bbo, they were not given!! So instead I have to intuit the specifics.
BBO’s computer now ranks the various sections somehow, and my best guess is by actual average number of player master points previously accumulated/earned in that section. The section with the highest average is given a priority or “handicap”. The idea being that the level of competition in that section was tougher than in any other section. The winner of that section, no matter how many IMPs they scored (or high high their % was) relative to other section winners, earns first place master points!!!!!.
Again I am guessing here and if I am wrong, please, please, please set the record straight with all of the correct specifics. If I am right, then I have a real problem with how your BBO/ACBLs method works.

The most important is the assumption that a player’s masterpoint number is an accurate measure of their ability. I know, I know heretical right? However I have played enough speedball tourneys to know this truth. If anyone doubts this, go and look at any tourney results, and while you are at it, take this tourney in particular. What you should expect to see is all of the players with larger numbers clustered at the top with the highest tourney scores. But you do not. Ever. Instead the larger numbered player’s scores are randomly distributed throughout the section. Always! They do not perform consistently better than those with less “experience/ability”. So why, why, why are the sections rated in the first place??? This method for assigning priority status to a section is as arbitrary as rubber bridge scoring and, in the interest of full disclosure, you should be telling this to all new speedball players. You should also be telling the specifics details, as I have done above for example, of how you determine master point assignments.
As an aside I have wondered how many of these players got such large numbers after their names when their play is not always superior. I have a couple of possible explanations. One that has been passed around by some BBO players is that they have passed their BBO identity on to relatives, friends or children.

My final complaint involves my specific masterpoint award in this tourney. Again we took 1st overall among E-W teams by 12 IMPs or 33% better than the second place E-W team. This other team actually had higher numbers after their names than I and my partner did, so presumably they are in a higher rank than us. If anyone should be given a “handicap”, it is us! And yet they get the first place MPs! Go figure.
I could even live with this result if I could at least know what we had to do to win 1st place MPs. All players want to know when they finish close to first is how much more they needed to get the top score. When an arbitrary section level assignment is given like this, a player doesn’t know the algorithm and cannot know what they had to do to get 1st place, nor how close they were either. They are left completely in the dark. BBO should quit using the excuse that the ACBL requires them to use their “system”, address the realities discussed above with the ACBL and take responsibility for offering fairer tourney placements. BBO can and should do better than this.

[barmar]

 adv exp, on 2017-January-13, 12:18, said:

In the meantime I did a detailed analysis of the whole tournament results in question and tried to determine your method for assigning master points and I think I may have figured it out.
It appears that what is done when there are multiple sections is that bbo (or bbo’s computer program) distributes teams such that each section has relatively equal distributions of talent/experience/expertise/ability, from very experienced all the way down. I would presume that bbo uses player master points (the big number after a person’s bbo name) as their measure of ability to evenly distribute talent among the sections. Already I have a big problem with this, but I will save that for later. Players are assigned a level based on how big their number is and I believe there are 3 levels/flights. Notice that I keep using terms like “it appears”, “I believe” etc. because despite my specific request for these kinds of details from bbo, they were not given!! So instead I have to intuit the specifics.

For players that are members of ACBL, we use their ACBL masterpoints. For players who are not ACBL members, we use their BBO masterpoints. We then order all the players, and put the top third in stratum A, the middle third in stratum B, and the bottom third in stratum C (in addition, all star players and new BBO members without any masterpoint history are put in A). We then distribute the pairs through all the sections, trying to get roughly equal numbers of A, B, and C pairs in each.

Quote

BBO’s computer now ranks the various sections somehow, and my best guess is by actual average number of player master points previously accumulated/earned in that section. The section with the highest average is given a priority or “handicap”. The idea being that the level of competition in that section was tougher than in any other section. The winner of that section, no matter how many IMPs they scored (or high high their % was) relative to other section winners, earns first place master points!!!!!.
Again I am guessing here and if I am wrong, please, please, please set the record straight with all of the correct specifics. If I am right, then I have a real problem with how your BBO/ACBLs method works.

No, there's no ranking of the sections or handicapping. Pairs are simply ranked within their section, and get the appropriate award based on their place there. In addition, all pairs are ranked together for the overall awards. You get either your section award or your overall award, whichever is higher.

Quote

The most important is the assumption that a player’s masterpoint number is an accurate measure of their ability. I know, I know heretical right? However I have played enough speedball tourneys to know this truth.

It's well known that masterpoints are only roughly correlated with ability, but it's the only measure we have. These are ACBL games, we're following ACBL procedures. Don't you have the same complaint when you play in f2f clubs and tournaments? Or maybe you only play online, so you didn't realize that this is just the way the game is played everywhere, not specific to BBO.

Quote

As an aside I have wondered how many of these players got such large numbers after their names when their play is not always superior. I have a couple of possible explanations. One that has been passed around by some BBO players is that they have passed their BBO identity on to relatives, friends or children.

I'm a below-average flight A player, but I have a 26 next to my name. How did I get that? By playing lots of robot games. Since I'm a BBO employee, I get to play for free, so I've been earning 30-40 MP per month for the past 5 years, almost all from robot duplicates and instant duplicates.

Quote

My final complaint involves my specific masterpoint award in this tourney. Again we took 1st overall among E-W teams by 12 IMPs or 33% better than the second place E-W team. This other team actually had higher numbers after their names than I and my partner did, so presumably they are in a higher rank than us. If anyone should be given a “handicap”, it is us!

There's no handicapping, although stratification is kind of similar, but less granular. If you're in stratum B, you can get the top award just by beating all the other B players, you don't have to beat the A players as well.

And the overall awards don't distinguish EW from NS pairs, that's only done in the section awards. Just like real-life tourneys.

Could you tell me the specific tournament you're talking about, and I'll explain how the awards were calculated.

[adv exp]

h

[adv exp]

 barmar, on 2017-January-13, 13:23, said:

For players that are members of ACBL, we use their ACBL masterpoints. For players who are not ACBL members, we use their BBO masterpoints. We then order all the players, and put the top third in stratum A, the middle third in stratum B, and the bottom third in stratum C (in addition, all star players and new BBO members without any masterpoint history are put in A). We then distribute the pairs through all the sections, trying to get roughly equal numbers of A, B, and C pairs in each.

No, there's no ranking of the sections or handicapping. Pairs are simply ranked within their section, and get the appropriate award based on their place there. In addition, all pairs are ranked together for the overall awards. You get either your section award or your overall award, whichever is higher.

It's well known that masterpoints are only roughly correlated with ability, but it's the only measure we have. These are ACBL games, we're following ACBL procedures. Don't you have the same complaint when you play in f2f clubs and tournaments? Or maybe you only play online, so you didn't realize that this is just the way the game is played everywhere, not specific to BBO.

I'm a below-average flight A player, but I have a 26 next to my name. How did I get that? By playing lots of robot games. Since I'm a BBO employee, I get to play for free, so I've been earning 30-40 MP per month for the past 5 years, almost all from robot duplicates and instant duplicates.

There's no handicapping, although stratification is kind of similar, but less granular. If you're in stratum B, you can get the top award just by beating all the other B players, you don't have to beat the A players as well.

"I find this sentence exasperating because it continues the practice of vaguely restating the result without explaining in detail how and what the final “scores” of my team and the other relevant teams who finished with a higher "score" than us, yet had much lower IMPs than us. I think your use of the term "granular" epitomizes this; impressive but useless for conveying relevant info. Could you please explain what my final "score" was and exactly how my final score was calculated and compared to the E-W team that received more MPs than us. They finished 7th overall and I believe 2nd E-W. What was their final score? Could you also show how the 7th place (overall) team's score was calculated and why they earned 1st place MPs? What was my particular stratum? What were the stratums of the 3rd and 7th place overall teams. If my stratum was not the lowest third, why isn't it? Can we know what our stratum is before any particular tournament starts? If not, why not? The tournament in question was speedball on Jan 5th at 6:00 PM #8342 ACBL".

"What happens if you are in the highest stratum and you beat all of the other teams, and…… so did a stratum B team beat all other B teams, and…….. so did a stratum C team beat all other C teams. Now who walks away with 1st place MPs and why?

Thanks, cranky old adv exp"


And the overall awards don't distinguish EW from NS pairs, that's only done in the section awards. Just like real-life tourneys.

Could you tell me the specific tournament you're talking about, and I'll explain how the awards were calculated.

[barmar]

The pair you're talking about was in stratum B.

In each stratum, 1st place pays 1.50, 2nd place pays 1.13, 3rd place pays .84. So you got .84 for coming in 3rd in A, they got 1.13 for coming in 2nd in B. The fact that you got more IMPs than them is irrelevant, because you're in a higher stratum, so they don't have to compete against you. Although if they did do better than you, they would get the A award -- stratification allows you to get a benefit from beating players in the higher strats, but doesn't penalize you if you lose to them.

You can't know your stratum before the tournament starts because it's not known. We calculate the strats when the tourney starts, by ordering all the players by their masterpoints and breaking them up into equal groups.

[barmar]

I'm looking into the details of how the strats were determined for that tourney, I'll get back to you on that.

[barmar]

The reason you were put into stratum A is because you were subs. All subs go in A, because they're not part of the original ordering of players that was used to calculate the strats.

I made one small error above. The masterpoints that are used for deciding strats are the higher of BBO and ACBL points, not just ACBL points.

[johnu]

 barmar, on 2017-January-13, 13:23, said:

I'm a below-average flight A player, but I have a 26 next to my name. How did I get that? By playing lots of robot games.

When you get to 100, I'll buy you a drink the next time we are at the same tournament

[adv exp]

 barmar, on 2017-January-15, 12:22, said:

The pair you're talking about was in stratum B.

In each stratum, 1st place pays 1.50, 2nd place pays 1.13, 3rd place pays .84. So you got .84 for coming in 3rd in A, they got 1.13 for coming in 2nd in B. The fact that you got more IMPs than them is irrelevant, because you're in a higher stratum, so they don't have to compete against you. Although if they did do better than you, they would get the A award -- stratification allows you to get a benefit from beating players in the higher strats, but doesn't penalize you if you lose to them.

You can't know your stratum before the tournament starts because it's not known. We calculate the strats when the tourney starts, by ordering all the players by their masterpoints and breaking them up into equal groups.

Thanks for all of your info. I now understand how all of the MPs are awarded. I looked over all of BBOs tourney info and also ACBLs and have never seen anything close to the description you have above. You and other BBO admins seem to be under the impression that all us tourney players know this stuff but I would guess otherwise. I play with several others in speedball tourneys and not a one of them knew these "rules". Also you said that "We calculate the strats when the tourney starts" so in fact you do know it pretty much at the beginning of the tourney. How hard would it be to inform players just as soon as you get the stratums, which would still be soon enough for players to know it before playing at least their second hand? One last thing. I would amend your masterpoiont award explanation slightly this way;

"In each stratum, 1st place pays 1.50, 2nd place pays 1.13, 3rd place pays .84. 1st place in a stratum is defined as that team that scored the most IMPS or had the highest percentage no matter which direction they sat in N-S or E-W. 2nd place in a stratum is defined as that team that scored the second most IMPs or had the 2nd highest percentage no matter which direction they sat in, N-S or E-W. 3rd place etc. You will not necessarily win any masterpoints for finishing 1st in a division if there are at enough teams finishing higher than you in the same (or lower) strata in the other division. Even though this is duplicate scoring and you might win your division by a considerable margin, if there are enough teams scoring higher than you in the other division you possibly could not earn any MPs." If you sub in a tourney, your stratum will automatically be the highest, i.e, A stratum, no matter how inexperienced you are or low your ability".

Full disclosure modeled in my description above is a good policy and should be readily available on your site somewhere.

adv exp

adv exp

[barmar]

 adv exp, on 2017-January-15, 19:49, said:

How hard would it be to inform players just as soon as you get the stratums, which would still be soon enough for players to know it before playing at least their second hand?

What good would it do? You're not going to withdraw if you discover you're in a higher strat than you hoped, are you? Will you play differently if you discover you're in strat A instead of B?

And with a handle like "adv exp", I'm surprised you have an issue with being in A.