[AAr]

I love playing the IMP robot duplicate tourneys, with one minor issue.

When I play the hand well, I make an overtrick while the rest of the Field just makes, scoring +650 while most of the Field scores +620, for a 1 IMP gain.

When I misplay the hand, I go down one while the rest of the Field makes exactly, scoring -100 to the Field's +620, losing 12 IMPs.

But, it doesn't seem like the 1 IMP gains make up for the 12 IMP losses. So, I'm getting punished more for the bad plays than awarded for the good ones. Am I just misplaying the hands at the wrongs times or what?

Really, here are my points.

A. Are some hands likely worth bigger swings than others, at IMPs, even against GIB?

B: Will a 12 IMP loss be more likely to drop me in the standings than a 1 IMP gain is to boost me, even in Robot IMP tourneys?

C: Shouldn't every board be of equal weight, even in IMPs?

D: What should I do?


I also play in the MP tourneys, which is slightly more boring (still fun enough though), but I do tend to fare slightly better in them.

Thanks for your input. I'm just frustrated that when I play a hand well that the Field doesn't, I don't get awarded much (at least not in the IMP ones), but I misplay the hands, it's disasterous. It just seems like being -100 to the Field's +620 is a much bigger adverse effect than being +650 to the Field's +620 is being a good one. I just don't think it's fair that my bad plays are punished more than my good plays are awarded because of the timing of the bad plays vs. the timing of the good plays.


Thanks!

[Stephen Tu]

 AAr, on 2013-March-08, 11:30, said:

A. Are some hands likely worth bigger swings than others, at IMPs, even against GIB?
C: Shouldn't every board be of equal weight, even in IMPs?

Boards being of unequal weight is the main defect of IMP scoring. If you want equality, stick with MPs. MPs are definitely more skill less luck since you get more opportunities to show superiority by taking extra tricks. IMPs is kind of a middle ground between MP scoring and total point scoring, it is used more for team play and meant to mimic old rubber bridge to some extent in that some boards are worth more than others; games swings worth more than partscore swings and overtricks, slams more than games, etc., and making/breaking contracts is vastly more important than overtricks. But it's scaled down so it's not quite as unequal as total points. A couple partial swings can cancel out a game swing at IMPs, while at total points you need 3-4 partial swings.

But the thing is, it's really best for team play, where the team aspect of both scores being under your side's control damps down the luck factor of boards being unevenly weighted. IMP pairs, in the real world, really isn't so popular because it's very much a crap shoot, there's massive luck in being stationed against the bad pair on the board where their misdefense can let a game or slam through for massive points. But for some reason it has gotten popular online, maybe some people like the randomness so they can win occasionally? IMP pairs online with the short 8-12 board tournaments is like the most luck-oriented duplicate bridge played in the world, it's very random, although in the long run the better players will have higher averages and placements. Not as consistently as they would playing MP though.

Quote

B: Will a 12 IMP loss be more likely to drop me in the standings than a 1 IMP gain is to boost me, even in Robot IMP tourneys?

Obviously 12 >>> 1.

Quote

D: What should I do?

- stick to MP tourneys
- concentrate more on boards where the contract can be made/set than on overtrick boards, if you are unable to maintain 100% concentration on all boards. Stop having 12 imp losses! 12 imp loss, sans the robots doing something weird to screw you, are usually a sign you are making a big error of some sort, else other players would be making the same error and tying you, lowering your loss down to like 6 imps. Look what the players who made the contract did, analyze whether their line is clearly superior to yours.

[ArtK78]

Boards being of unequal weight is the main characteristic of IMP scoring of pairs tourneys, not the main defect. Just like matchpoint pair events are a different game than IMP team events, IMP pair events are a different game than matchpoint pair events.

One must recognize the differences in the different forms of scoring and adjust accordingly. The play of the hand at IMP pairs is, in almost all cases, identical to the play of the hand at IMP teams, which is considerably different than the play of the hand at matchpoint pairs. At matchpoint pairs, the difference between scoring 10 or 11 tricks in a 4♠ contract may be considerable, and may be worth the risk of going down. At IMP pairs, that is not the case. While the overtrick may be worth an IMP (and it may not be), going down one in a cold 4♠ will certainly be a disaster. Furthermore, even if you are more than a 12-1 favorite to make the overtrick, that doesn't mean that you should risk the contract for the overtrick IMP. As I alluded to earlier, the overtrick may not be worth an IMP. Furthermore, gaining an IMP doesn't mean that you will move up in the rankings, while losing 12 will almost certainly cause you to drop several places.

[Leo LaSota]

 Stephen Tu, on 2013-March-08, 12:25, said:

Boards being of unequal weight is the main defect of IMP scoring. If you want equality, stick with MPs. MPs are definitely more skill less luck since you get more opportunities to show superiority by taking extra tricks. IMPs is kind of a middle ground between MP scoring and total point scoring, it is used more for team play and meant to mimic old rubber bridge to some extent in that some boards are worth more than others; games swings worth more than partscore swings and overtricks, slams more than games, etc., and making/breaking contracts is vastly more important than overtricks. But it's scaled down so it's not quite as unequal as total points. A couple partial swings can cancel out a game swing at IMPs, while at total points you need 3-4 partial swings.

But the thing is, it's really best for team play, where the team aspect of both scores being under your side's control damps down the luck factor of boards being unevenly weighted. IMP pairs, in the real world, really isn't so popular because it's very much a crap shoot, there's massive luck in being stationed against the bad pair on the board where their misdefense can let a game or slam through for massive points. But for some reason it has gotten popular online, maybe some people like the randomness so they can win occasionally? IMP pairs online with the short 8-12 board tournaments is like the most luck-oriented duplicate bridge played in the world, it's very random, although in the long run the better players will have higher averages and placements. Not as consistently as they would playing MP though.



Obviously 12 >>> 1.


- stick to MP tourneys
- concentrate more on boards where the contract can be made/set than on overtrick boards, if you are unable to maintain 100% concentration on all boards. Stop having 12 imp losses! 12 imp loss, sans the robots doing something weird to screw you, are usually a sign you are making a big error of some sort, else other players would be making the same error and tying you, lowering your loss down to like 6 imps. Look what the players who made the contract did, analyze whether their line is clearly superior to yours.

You are completely incorrect with your statement that the IMP pair online tournaments are high luck oriented and very random. It is not random that I win over 1/3 of the IMP pair online tournaments that I participate in. You fail to recognize that in the robot IMP pair tournaments, everyone has the same partner and the same opponents on every single board. Therefore, to win requires a great deal of skill, just like robot MP tournaments. Yes, there are certain aspects of the game that are more critical in IMP pair scoring as opposed to MP pairs. However, both forms of the game greatly measure skill since everyone has the same partner and same opponents on every hand.

The reason that IMP pairs in live bridge is very random is because everyone has different opponents on every board and tons and tons of imps can be gained based on who your opponent happens to be on a swingy board.

[Stephen Tu]

 Leo LaSota, on 2013-March-08, 18:47, said:

You are completely incorrect with your statement that the IMP pair online tournaments are high luck oriented and very random. It is not random that I win over 1/3 of the IMP pair online tournaments that I participate in. You fail to recognize that in the robot IMP pair tournaments, everyone has the same partner and the same opponents on every single board. Therefore, to win requires a great deal of skill, just like robot MP tournaments. Yes, there are certain aspects of the game that are more critical in IMP pair scoring as opposed to MP pairs. However, both forms of the game greatly measure skill since everyone has the same partner and same opponents on every hand.

The reason that IMP pairs in live bridge is very random is because everyone has different opponents on every board and tons and tons of imps can be gained based on who your opponent happens to be on a swingy board.

Yeah, you're right, I was thinking of the human IMP tourneys when I wrote that and forgot that original post was specifically about the robot tourneys; I should have been more specific. The identical opps/pd reduce the luck a lot. Still, I think IMP bot tourneys are going to be more random than MP bot tourneys. You bid some nice vul slam that the field doesn't, but goes down on 4-1 break or whatever. Boom, you lose a ton of imps, even though you are gaining 1.5 imp per board on the others. Your shot for that tourney is blown, although over the long run you'll do better bidding the slam than not, helping you on average. Or some weirdness in how the robot's random seeding based on what random spot card you played leads it to go ahead and find the winning defense to 3nt, while it lets it through at other tables, even though your random card isn't technically worse or better than the card other people chose.

If you are winning 1/3 of the IMP tourneys, I expect you win a greater percentage of the MP tourneys.

[Leo LaSota]

 Stephen Tu, on 2013-March-08, 19:16, said:

Yeah, you're right, I was thinking of the human IMP tourneys when I wrote that and forgot that original post was specifically about the robot tourneys; I should have been more specific. The identical opps/pd reduce the luck a lot. Still, I think IMP bot tourneys are going to be more random than MP bot tourneys. You bid some nice vul slam that the field doesn't, but goes down on 4-1 break or whatever. Boom, you lose a ton of imps, even though you are gaining 1.5 imp per board on the others. Your shot for that tourney is blown, although over the long run you'll do better bidding the slam than not, helping you on average. Or some weirdness in how the robot's random seeding based on what random spot card you played leads it to go ahead and find the winning defense to 3nt, while it lets it through at other tables, even though your random card isn't technically worse or better than the card other people chose.

If you are winning 1/3 of the IMP tourneys, I expect you win a greater percentage of the MP tourneys.

I win a greater % of the robot IMP games that I participate in than the robot MP tournaments. When everyone has the same opponent, each board is weighted the same in MPs. In IMPS, some boards are alot more pivotal than others. Let us say that on average, about 9 of the boards in a MP event vary alot in the outcomes of the participants. Let us say that in an IMP pair tournament, an average of 4 boards are pivotal near double digit IMP swings. The total number of swingy boards may be more in the MP tournament, but the importance of each swingy board in IMPs is magnified. If a player has an outstanding result on a board in MPs and everyone else has the same lower score on that board, you start about 1/2 of a board ahead of your competition. In IMPS, if there are 15 imps that you gain for an outstanding result, it is much more difficult for the other players to pick up those 15 imps on the other boards if you play well. The top players will frequently get these swingy boards correct. Therefore, the top players you see near the top of the leaderboard in IMPS even more than you do at MPs in the robot tournaments where everyone has the same opponents, same partner every hand.

[impie]

I enjoy the robot tourneys because they are quick and can be played at any time. However i've had some very weird and frustrating results playing both IMP & MP. Twice on one day i got a bad result in a hand that i really didn't think could have been bid or played better and when i checked the results at the other tables i found i was the ONLY one where the robot had led a completely different card. On the other occasion i was also the only one where the robot had bid something really odd even though i'd bid the same as the other tables. Can they be that inconsistent??? Also can anyone explain to me what the A, B and C results mean? Obviously the C list is the hand just played but the others? Thanks.

[Bbradley62]

 impie, on 2014-July-30, 04:36, said:

I enjoy the robot tourneys because they are quick and can be played at any time. However i've had some very weird and frustrating results playing both IMP & MP. Twice on one day i got a bad result in a hand that i really didn't think could have been bid or played better and when i checked the results at the other tables i found i was the ONLY one where the robot had led a completely different card. On the other occasion i was also the only one where the robot had bid something really odd even though i'd bid the same as the other tables. Can they be that inconsistent??? Also can anyone explain to me what the A, B and C results mean? Obviously the C list is the hand just played but the others? Thanks.

In robot tournaments (except Instant tournaments), all robots will behave exactly the same in the same circumstances. If your robots did something different than the others, then either your auction was different or the previous tricks were somehow different. Please provide an example if you think otherwise.

A/B/C refer to strata (level divisions) of players. Masterpoints are awarded on this basis. Players who qualify for C (the lowest) compete against only each other for masterpoints... Players classifies as B or C compete for the B-level points, and everyone competes for the A points.

[gordontd]

 AAr, on 2013-March-08, 11:30, said:

A. Are some hands likely worth bigger swings than others, at IMPs, even against GIB?

B: Will a 12 IMP loss be more likely to drop me in the standings than a 1 IMP gain is to boost me, even in Robot IMP tourneys?

C: Shouldn't every board be of equal weight, even in IMPs?

D: What should I do?

A: Yes of course

B: Yes of course

C: No

D: Play some other game

PS Just realised this thread is a necro.

This post has been edited by gordontd: 2014-July-30, 06:50