[thepossum]

Hi all

I still regard myself as the world's most experienced beginner or an intermediate with extremely variable bidding and particularly declarer play. Every time I think I'm starting to make progreess biddinng contracts, games, slams and other contracts missed by others, improving accuracy of contracts, winning the occasioonal instant tournament when along comes a slam contract which should go down 1 but I manage to ruin an entry and go down 6 (NB I only wanted to be in 4 but GIB pushed on to 6) On average I've improved to not missing out by more than 1 trick very often on good-average contracts/play

So, if you were analysing average IMPs and average MPs over time on BBO tournaments do you think it can give a guide to player rating. For example does an average of an IMP per hand give any guidance, or an average of 40%, 45% etc in MPs. Obviously I know what the MPs means in terms of percentile but in terms of ranking myself. However in terms of IMPs don't really know ho to use them to understand, and obviously it depends to an extent who is in the tournaments and what the average BBO player is ranked at. I'm also trying to reduce my variance obviously. When I say ranking I don't mean international stuff, just intermediate average club etc.

PS I've decided to call myself intermediate but am still prone to beginner errors from time to time. Most errors seem to relate to having the right plan but such a poor memory I make an error.



regards P

[pescetom]

If you use the search feature for ranking or self ranking you will pull up this topic plus endless other discussions. The bottom line is that BBO don't want automatic ranking.

BBOskill gave quite credible results until BBO pulled the plug on it a few years ago, nominally because the crawling was slowing down their database. It was probably unfavourable towards those that mainly play with random partners, but on the whole it gave advanced to players who really were advanced, and so on.

[mikeh]

average imps per board are completely dependent on the relative skill level of one's partner and opponents, if playing as a pair (which is the usual format in BBO) or one's partner, and teammates and opps, if playing a head to head team game. If playing a tournament, with real people, you are still very much bounded by the quality of the field and, often, the randomness of the opps (and partner).

A WC pair playing another WC pair will, on average and in the long run, score pretty close to 0 imps per board. The saem WC pair playing a less-skilled pair will, on average and in the long run, score as plus...the amount of the plus being inversely related to the skill level of the opponents.

Playing a lot of bridge online, whether against bots or random opps, and with random partners, is not, imo, a way to improve one's game. The hard truths are that one's ability to recognize mistakes is limited by one's knowledge of the game, and that seeing how one might have done better on any given board is no indication that one 'should' have done better.

As an example of the latter, many clubs in NA offer printouts of the hands at the end of the game, and each printout lists how many tricks EW and NS could make in various contracts. I used to give lectures at the local club, and many times people asked me how they should get to 4S (as an example), and would explain that, firstly, reaching 4S would require very bad bidding and, secondly, that the only way to make 4S would be to badly misplay the hand. Bridge is a game of percentages and one strives to reach the optimal contract based on percentages and to make the best play based on percentages, but the software that tells one that one could, for example, make 4S is based on double-dummy play.

Going back to that first hard truth....not only are you going to be limited in your ability to identify your own mistakes, but so too will be your partners and opponents. Indeed, at your level, most of the 'advice' you get will not merely be wrong but may be actively harmful.

Obviously that doesn't mean you stop playing. It does mean that you should avail yourself of other means of learning. Play to practice...read or watch to learn.

Look for 'good' team games on BBO, with good commentators. This means, usually, major events, but can also mean some of the better team matches, such as those with Cayne and his teams.

Read...read...read. Texts are fine, tho I'd look for books on play more than on bidding. Find old Kelsey books...the bidding is antiquated and of little interest, but the play is the thing. Read in all the bridge forums here, and read on Bridge Winners. If the budget can stretch to it, subscribe to the Bridge World.

Find a partner who has as much interest as do you, and spend time discussing things.

One of the most enjoyable and challenging parts of this game is that the more you know, the more you realize you don't know. I think it was Hamman, the world's top-ranked player at the time, who, when asked about what he hoped to accomplish said words to the effect that he hoped that he'd learn how to play the game.

[diana_eva]

I think tracking your own performance based on the results you are getting is very reasonable, especially when always playing the same type of game. Instant Tournaments are not as random as most free games out there, the robots are consistent, so it makes sense to do what you are doing -- see whether you can keep yourself above 50% for instance, then aim for 55% and so on.

[helene_t]

Daylong tournaments, instant tournaments and robot rebate all give meaningful scores and as Diana suggests, you can try to see if you can improve your long term average in those tourneys. Beware, however, that Daylong has a weaker field than Robot Rebate so you can't compare the results between those two.

Robot Race and Robot Reward are somewhat strange games that reward other skills than the ones you need in real bridge.

Play with human partners and/or opps is more interesting but is not suitable for ranking.

[pescetom]

helene_t, on 2018-October-25, 19:52, said:

Play with human partners and/or opps is more interesting but is not suitable for ranking.

There are obvious limitations (match two world-class pairs against each other for long enough and the result will be 0 IMPs) but even including human partnerships and all types of opposition some truth does emerge. However it was calculated, BBOskill was impressively accurate in my experience - people I judged in real life as intermediate, advanced or somewhere inbetween had scores to match. I imagine some of them played only robots, some only known humans, some only random humans and most some mixture, but the numbers were right all the same.