Help
Results of a compilation of BBO Boards from MyHands.
Link removed.
I think the topic has been beaten to death.
Sorry for the inconvenience
Wayne
Page 1 of 1
BBO Boards from MyHands
#1
- Group: Full Members
- Posts: 180
- Joined: 2003-July-10
- Gender:Male
- Location:Henderson, NV
- Interests:Bridge, Poker
Posted 2013-May-22, 23:17
#2
- Group: Advanced Members
- Posts: 3,214
- Joined: 2007-December-16
- Location:Pasadena, CA
Posted 2013-May-23, 04:48
Sorry I don't understand what this table is meant to represent or imply, can you please explain?
#3
- enigmatic
-
- Group: Advanced Members
- Posts: 4,757
- Joined: 2009-July-31
- Gender:Male
Posted 2013-May-23, 06:35
I think it is an attempt to rate players by mining their results from the BBO database.
Life is long and beautiful, if bad things happen, good things will follow.
-gwnn
-gwnn
#4
- Group: Full Members
- Posts: 180
- Joined: 2003-July-10
- Gender:Male
- Location:Henderson, NV
- Interests:Bridge, Poker
Posted 2013-May-23, 08:37
rogerclee, on 2013-May-23, 04:48, said:
Sorry I don't understand what this table is meant to represent or imply, can you please explain?
It is not intended to represent or imply anything. It is a compilation of the results from importing BBO MyHands data into an Excel database and retaining only the boards played by the player and his/her most frequent partners (8 boards or more in the period examined). Boards played with GiBs were discarded.
I think it indicates there is a definite correlation between skill level and IMP/Board averages over the long haul. Without doing all the grunt work I had to do to determine these averages, I also think advancing players can learn much by reviewing (using BBO MyHands) the better boards (+7 IMPs or more) played by really good players. This is like kibbing them after the fact..
Using the data collected and screened I have reviewed many boards with favorable results (7+ IMPs) achieved by some of the best players in the world. This process continues. So far, I have not found the holy grail of bridge; no killer bidding system, no pattern of genius level decision making processes, no apparent ESP. What I have found is that the key to winning bridge at all levels seems to be sound bidding and play; avoiding stupid mistakes and making fewer judgement errors than you opponents. And ....... crouching behind a wall .... LUCK, lots and lots of luck. Boards handed on a silver platter by opponents making stupid mistakes and bad judgement calls. And yes, this does occur at the highest levels of bridge with surprising frequency.
#5
- Group: Admin
- Posts: 21,589
- Joined: 2004-August-21
- Gender:Male
Posted 2013-May-23, 08:40
Wayne_LV, on 2013-May-23, 08:37, said:
What I have found is that the key to winning bridge at all levels seems to be sound bidding and play; avoiding stupid mistakes and making fewer judgement errors than you opponents. And ....... crouching behind a wall .... LUCK, lots and lots of luck.
In other words, if you play better than your opponents, you'll get better scores. Who'da thunk it?
#6
- Group: Advanced Members
- Posts: 3,214
- Joined: 2007-December-16
- Location:Pasadena, CA
Posted 2013-May-23, 14:25
I just don't understand how any information can be gleaned from these numbers if they don't account for the difficulty level of your opponents.
#7
- Group: Full Members
- Posts: 180
- Joined: 2003-July-10
- Gender:Male
- Location:Henderson, NV
- Interests:Bridge, Poker
Posted 2013-May-23, 18:51
rogerclee, on 2013-May-23, 14:25, said:
I just don't understand how any information can be gleaned from these numbers if they don't account for the difficulty level of your opponents.
If the skill level of the opponents could be arbitrarily judged then so could the skill level of the subject player be so judged. The assumption is that the opponents and the subject are of comparable skill levels. In the absence of such an assumption, you are correct: no information can be gleaned. There are certainly lopsided games where one pair is expert and the other pair is novice or worse; but, those games do not go for many boards before the weaker pair runs for cover.
#8
- Group: Advanced Members
- Posts: 3,214
- Joined: 2007-December-16
- Location:Pasadena, CA
Posted 2013-May-23, 19:34
Sorry you are going to have to explain to me what is going on. It seems like you put a solid amount of work into this, so I want to know if there is something useful I can get out of it.
However, they are evidently not playing comparably skilled opponents if some of them are +1 IMP/bd, for example.
However, they are evidently not playing comparably skilled opponents if some of them are +1 IMP/bd, for example.
#9
- Group: Full Members
- Posts: 180
- Joined: 2003-July-10
- Gender:Male
- Location:Henderson, NV
- Interests:Bridge, Poker
Posted 2013-May-24, 09:55
rogerclee, on 2013-May-23, 19:34, said:
Sorry you are going to have to explain to me what is going on. It seems like you put a solid amount of work into this, so I want to know if there is something useful I can get out of it.
However, they are evidently not playing comparably skilled opponents if some of them are +1 IMP/bd, for example.
However, they are evidently not playing comparably skilled opponents if some of them are +1 IMP/bd, for example.
Most likely there is little to nothing to be gained by looking at the numbers on the chart. I thought posting a comparison might be of interest to others, but without the detail I guess it is of little real value.
With the individual spreadsheets I am able to look at the best and worst boards of the players who's data I imported. I can also examine numerous scenarios to see how they do with various contracts as declarers and as defenders.
The real value (I thought) would be to identify really good players on BBO who play significant numbers of boards. Then to review in detail their better boards with hopes of learning what they do to be consistent winners over large numbers of boards.
Few are willing admit that luck is a factor in duplicate bridge, but after reviewing many boards won by large IMPs, I see a significant amount of luck involved. I realize that some of what I might attribute to luck is really a skill I am not able to understand. However, many of the big boards are won by opponents making really dumb mistakes and gifting their opponents with a great score. One set of boards played by one of the top partnerships in the world vs. another top partnership was won by close to 1 IMP per board and the vast majority of the big swings were the result of one pair making mistakes that most intermediate players would have avoided. But, luck is certainly not the main factor.
At any rate, I think this study has been largely a waste of time. There are better ways of improving bridge skills than trying to copy the methods of others, no matter how great they may be. I am becoming convinced that, for the most part, the skills required to become an expert bridge player cannot be learned. The important skills are those you either have or you don't: High IQ, good instincts, the ability to quickly assimilate facts and apply that information to making good judgement calls, a near photographic memory, and other talents that make people "smart".
The most valuable thing to come from this effort is an Excel template that allows me to dissect the boards I have played with regular partner(s) to identify bad trends in our game.
"The sum of all technical knowledge cannot make a master contract player." Ely Culbertson
#10
- Group: Members
- Posts: 10
- Joined: 2013-May-22
Posted 2013-May-24, 12:32
Wayne_LV, on 2013-May-24, 09:55, said:
Most likely there is little to nothing to be gained by looking at the numbers on the chart. I thought posting a comparison might be of interest to others, but without the detail I guess it is of little real value.
With the individual spreadsheets I am able to look at the best and worst boards of the players who's data I imported. I can also examine numerous scenarios to see how they do with various contracts as declarers and as defenders.
The real value (I thought) would be to identify really good players on BBO who play significant numbers of boards. Then to review in detail their better boards with hopes of learning what they do to be consistent winners over large numbers of boards.
Few are willing admit that luck is a factor in duplicate bridge, but after reviewing many boards won by large IMPs, I see a significant amount of luck involved. I realize that some of what I might attribute to luck is really a skill I am not able to understand. However, many of the big boards are won by opponents making really dumb mistakes and gifting their opponents with a great score. One set of boards played by one of the top partnerships in the world vs. another top partnership was won by close to 1 IMP per board and the vast majority of the big swings were the result of one pair making mistakes that most intermediate players would have avoided. But, luck is certainly not the main factor.
At any rate, I think this study has been largely a waste of time. There are better ways of improving bridge skills than trying to copy the methods of others, no matter how great they may be. I am becoming convinced that, for the most part, the skills required to become an expert bridge player cannot be learned. The important skills are those you either have or you don't: High IQ, good instincts, the ability to quickly assimilate facts and apply that information to making good judgement calls, a near photographic memory, and other talents that make people "smart".
The most valuable thing to come from this effort is an Excel template that allows me to dissect the boards I have played with regular partner(s) to identify bad trends in our game.
"The sum of all technical knowledge cannot make a master contract player." Ely Culbertson
With the individual spreadsheets I am able to look at the best and worst boards of the players who's data I imported. I can also examine numerous scenarios to see how they do with various contracts as declarers and as defenders.
The real value (I thought) would be to identify really good players on BBO who play significant numbers of boards. Then to review in detail their better boards with hopes of learning what they do to be consistent winners over large numbers of boards.
Few are willing admit that luck is a factor in duplicate bridge, but after reviewing many boards won by large IMPs, I see a significant amount of luck involved. I realize that some of what I might attribute to luck is really a skill I am not able to understand. However, many of the big boards are won by opponents making really dumb mistakes and gifting their opponents with a great score. One set of boards played by one of the top partnerships in the world vs. another top partnership was won by close to 1 IMP per board and the vast majority of the big swings were the result of one pair making mistakes that most intermediate players would have avoided. But, luck is certainly not the main factor.
At any rate, I think this study has been largely a waste of time. There are better ways of improving bridge skills than trying to copy the methods of others, no matter how great they may be. I am becoming convinced that, for the most part, the skills required to become an expert bridge player cannot be learned. The important skills are those you either have or you don't: High IQ, good instincts, the ability to quickly assimilate facts and apply that information to making good judgement calls, a near photographic memory, and other talents that make people "smart".
The most valuable thing to come from this effort is an Excel template that allows me to dissect the boards I have played with regular partner(s) to identify bad trends in our game.
"The sum of all technical knowledge cannot make a master contract player." Ely Culbertson
I think what you are seeing is that experts just make many fewer dumb mistakes than the rest of us. Note, however, that may seem like a "normal" play to an intermediate or advanced player will often be a clear error to the expert (although maybe the advanced player will see and understand the mistake in the post-mortem). In addition, the expert will also give away less information through tempo and through carding than lesser players. Experts also have better bidding judgment, which may not show up as "errors," but on borderline bids that are 60/40, making the right judgment adds up over time (here, I am talking about situations where the suboptimal bid does not appear unreasonable).
At the margins, they will also have greater technical skill in play and bidding, but something like a compound squeeze or the opportunity to use a deep fancy convention comes up sufficiently rarely, that I doubt these add up to more than 0.2 imps/board in the long run.
This doesn't fully explain outlier pairs like Meckwell, but I think accounts for 90%+ of the difference between an intermediate+ pair that averages 50% and a "normal" expert pair that averages 63%.
#11
- Group: Admin
- Posts: 21,589
- Joined: 2004-August-21
- Gender:Male
Posted 2013-May-24, 15:22
Wayne_LV, on 2013-May-24, 09:55, said:
Few are willing admit that luck is a factor in duplicate bridge, but after reviewing many boards won by large IMPs, I see a significant amount of luck involved. I realize that some of what I might attribute to luck is really a skill I am not able to understand.
Luck is a factor, but in the long run it evens out, so it can't be responsible for players who do well pretty consistently.
It's often said that "you make your own luck". And in bridge, it's often said that the good players achieve much of their success not through brilliant plays (there simply aren't that many opportunities to execute a backwash squeeze), but by making fewer mistakes. The way you make your own luck is by noticing when the opponents have screwed up, and making sure that you take your gift. Poorer players will let their opponents get away with overbidding, good players know when to make them pay for it.
#12
- Group: Full Members
- Posts: 496
- Joined: 2009-December-12
Posted 2013-May-24, 16:39
- Bridge is a zero-sum game, whether head to head, or in a "tournament"
- There are 3 distinct forms of competition, individual, pairs, and teams.
- There are different rules for scoring - MPs, IMPs, Total Points on BBO My Hands. Rubber bridge (on people's hard drives - maybe) and others.
- There are different rules for contests - set by various associations.
- BBO even has games where one player is known to have at least equal to the best hand
- There are aberrations - perhaps 4 people were just practicing bidding - and making claims that made no sense.
- There is luck involved - which "in the long run" should net to zero (if all hands are included).
- The transitivity property is not valid. If A > B and B > C, one can't correctly infer that A > C.
- Players compete in divisions - just like in college or pro sports. Interdivisional play occurs, but it is not especially common.
Most of the above contribute to the fact that BBO, with vastly more data than you used, has not established a player skill rating system, or a partnership skill rating system, or a team rating system. One could argue that even if they could, they might not, for business reasons.
I think taking a small sample (selected or not) is not likely to lead to useful new knowledge. Throwing out the small scores seems inherently wrong. For example, you could select Sabine Auken and Roy Welland matches and collect up a ton of swing boards. Suppose they broke roughly even. Would you rate them as average players? Or their team as an average team? What would you learn from the scores? You can be sure that the matches that they play on BBO are against "Division 1" opponents. Even if they came out even... or on the short end of the stick in your sample...
#13
- Group: Advanced Members
- Posts: 5,808
- Joined: 2003-January-15
- Gender:Male
- Location:USA
Posted 2013-May-25, 20:59
Color me simplistic but I think that when someone plays a lot of hands against a number of different opponents, there might be a correlation between the skill level of a player and the average number of total points earned per hand.
This raises issues immediately, of course, even if you don't instantly yell "BS!"
How many hands do we need to examine? Across how many opponents? How to factor in the opponents?
Assume that a primitive adjustment based on opponent's average total point score has taken place and that 50 boards are enough. This is what a summary of our recently active userbase would look like
Simple tests indicate some correlation between this sort of statistic ( the adjusted average total score per board ) and being starred, ACBL LM status, and number of masterpoints won here.
The people with the ridiculously high +ve scores tend to be, as you'd imagine, cheaters.
Is this a meaningful number? is the percentile a meaningful number? Would people take this number too seriously if we published it in the profile?
Anyway, thought these numbers might be of interest, so I'm sharing.
Uday
This raises issues immediately, of course, even if you don't instantly yell "BS!"
How many hands do we need to examine? Across how many opponents? How to factor in the opponents?
Assume that a primitive adjustment based on opponent's average total point score has taken place and that 50 boards are enough. This is what a summary of our recently active userbase would look like
Simple tests indicate some correlation between this sort of statistic ( the adjusted average total score per board ) and being starred, ACBL LM status, and number of masterpoints won here.
The people with the ridiculously high +ve scores tend to be, as you'd imagine, cheaters.
Is this a meaningful number? is the percentile a meaningful number? Would people take this number too seriously if we published it in the profile?
Anyway, thought these numbers might be of interest, so I'm sharing.
Uday
All players with recent activity and with 50+ boards For example: Someone has played at least 50 boards, and averaged -520 per board. And at the other extreme someone has averaged +340 per board over 50 boards +------------------------------------+-------------------+-------------------+ | adjusted_avg_total_score_per_board | Number_of_players | number_of_players | +------------------------------------+-------------------+-------------------+ | -520 | 1 | | | -420 | 1 | | | -400 | 1 | | | -300 | 1 | | | -270 | 2 | | | -260 | 1 | | | -250 | 1 | | | -230 | 3 | | | -220 | 3 | | | -210 | 1 | | | -200 | 2 | | | -190 | 7 | | | -180 | 11 | | | -170 | 14 | | | -160 | 35 | | | -150 | 37 | | | -140 | 80 | | | -130 | 88 | | | -120 | | | | -110 | | | | -100 | | * | | -90 | | * | | -80 | | ** | | -70 | | ** | | -60 | | **** | | -50 | | **** | | -40 | | ******** | | -30 | | ********* | | -20 | | ************* | | -10 | | ************ | | 0 | | **************** | | 10 | | ************ | | 20 | | ************* | | 30 | | ******** | | 40 | | ******* | | 50 | | **** | | 60 | | *** | | 70 | | ** | | 80 | | * | | 90 | | * | | 100 | | * | | 110 | | | | 120 | 194 | | | 130 | 86 | | | 140 | 94 | | | 150 | 51 | | | 160 | 27 | | | 170 | 25 | | | 180 | 26 | | | 190 | 9 | | | 200 | 5 | | | 210 | 8 | | | 220 | 6 | | | 230 | 3 | | | 240 | 2 | | | 250 | 1 | | | 260 | 1 | | | 270 | 2 | | | 340 | 1 | | +------------------------------------+-------------------+-------------------+ All players with recent activity, 50+ boards, and a star +------------------------------------+-------------------+-------------------+ | adjusted_avg_total_score_per_board | Number_of_players | number_of_players | +------------------------------------+-------------------+-------------------+ | -120 | 1 | | | -90 | 2 | | | -80 | 4 | | | -70 | 10 | * | | -60 | 7 | * | | -50 | 15 | ** | | -40 | 23 | ** | | -30 | 33 | *** | | -20 | 51 | ***** | | -10 | 42 | **** | | 0 | 83 | ******** | | 10 | 69 | ******* | | 20 | 94 | ********* | | 30 | 70 | ******* | | 40 | 58 | ****** | | 50 | 50 | ***** | | 60 | 41 | **** | | 70 | 27 | *** | | 80 | 21 | ** | | 90 | 6 | * | | 100 | 6 | * | | 110 | 3 | | | 120 | 2 | | | 130 | 2 | | | 140 | 3 | | | 150 | 2 | | | 170 | 1 | | +------------------------------------+-------------------+-------------------+ ACBL Lifemasters +------------------------------------+-------------------+----------------------------------------+ | adjusted_avg_total_score_per_board | Number_of_players | number_of_players | +------------------------------------+-------------------+----------------------------------------+ | -120 | 1 | | | -110 | 3 | | | -100 | 5 | * | | -90 | 7 | * | | -80 | 15 | ** | | -70 | 15 | ** | | -60 | 36 | **** | | -50 | 65 | ******* | | -40 | 127 | ************* | | -30 | 155 | **************** | | -20 | | **************************** | | -10 | | *************************** | | 0 | | ************************************ | | 10 | | ******************************** | | 20 | | ************************************** | | 30 | | *************************** | | 40 | | *************************** | | 50 | 161 | **************** | | 60 | 114 | *********** | | 70 | 54 | ***** | | 80 | 45 | ***** | | 90 | 24 | ** | | 100 | 13 | * | | 110 | 9 | * | | 120 | 6 | * | | 130 | 4 | | | 140 | 5 | * | | 180 | 1 | | +------------------------------------+-------------------+----------------------------------------+ Users with '7' and up in their profile +------------------------------------+-------------------+----------------------------------------------------+ | adjusted_avg_total_score_per_board | Number_of_players | number_of_players | +------------------------------------+-------------------+----------------------------------------------------+ | -160 | 2 | | | -120 | 1 | | | -110 | 2 | | | -100 | 7 | * | | -90 | 9 | * | | -80 | 21 | ** | | -70 | 22 | ** | | -60 | | ***** | | -50 | | ******** | | -40 | | **************** | | -30 | | ********************* | | -20 | | *************************************** | | -10 | | ************************************ | | 0 | | ************************************************** | | 10 | | ***************************************** | | 20 | | ************************************************ | | 30 | | ********************************* | | 40 | | *************************** | | 50 | | ***************** | | 60 | | ************* | | 70 | | ****** | | 80 | | **** | | 90 | | *** | | 100 | | *** | | 110 | 18 | ** | | 120 | 8 | * | | 130 | 6 | * | | 140 | 5 | * | | 170 | 1 | | +------------------------------------+-------------------+----------------------------------------------------+ Users with '9' and up +------------------------------------+-------------------+------------------------+ | adjusted_avg_total_score_per_board | Number_of_players | number_of_players | +------------------------------------+-------------------+------------------------+ | -100 | 3 | | | -90 | 4 | | | -80 | 4 | | | -70 | 10 | * | | -60 | 10 | * | | -50 | | *** | | -40 | | ****** | | -30 | | ******** | | -20 | | **************** | | -10 | | ************** | | 0 | | ********************** | | 10 | | ******************* | | 20 | | ********************* | | 30 | | *************** | | 40 | | ************* | | 50 | | ********* | | 60 | | ******* | | 70 | | *** | | 80 | | ** | | 90 | 9 | * | | 100 | 9 | * | | 110 | 7 | * | | 120 | 3 | | | 130 | 3 | | | 140 | 3 | | +------------------------------------+-------------------+------------------------+
Page 1 of 1