[pfatseas]

I have been playing Robot Bridge,four deals for scores for a LONG time.
I am getting frustrated because of the losing finesses. Now a finesse has a 50% chance of winning. But after playing this game so many times, I startimg to realize that my finesses were losing in over 75% of the cases. I mean over at least a year that I noticed.
So I put it to the programmers that playing these robot games is not fair. Try it yourself if you don't believe it. I have started to keep track of how many times a finesse will lose. Upwards of 75% of the time over a year!!
Why play online here if statistically they are not fair?

[smerriman]

We have put it to the test. Many people claim this, and each time a proper analysis is done - rather than what you 'feel' is true, which is always heavily biased - everything works out at perfectly 50%.

The fact people keep claiming it is unsurprising, since studies have shown humans are very poor at recalling this type of information correctly, remembering the failing cases far more than the successful cases.

It is, of course, possible that you are taking poor finesses. For example, from the opening lead, if the robot leads a suit and you have AQ in dummy, the finesse will almost surely lose - because the robot almost never leads away from a king. This doesn't mean finesses fail more than 50% of the time.

[johnu]

Finesses only lose 75% of the time? Actually that's quite good for users that aren't members of the Super Latinum Platinum Select Members Club. It used to be that finesses lost 90% but there were so many complaints that I guess BBO dialed it back to 75%.

For the record, member of the Super Latinum Platinum Select Members Club average 85% successful finesses, and never get 4-1 or 4-0/5-0 trump splits, and opening leader never has a singleton to lead. Membership has its privileges!

[thepossum]

One thing I noticed recently is that the Advanced Robot loses or tries fewer losing finesses than the Basic one. Just a small sample

One of the biggest myths in Bridge is that finesses are 50%

It's not like Schrödinger's Cat

Or maybe it is. You don't know until you do it

Or wait until the end of the hand to find out

[AL78]

Finesses win when you have underbid and missed a game. Finesses lose when you have bid aggressively to a tight game or overbid to a bad game.

If you believe your finesses are always losing, look out for elimination and endplays or squeezes instead. Making on a simple finesse is boring, anyone can do that. The expert play is to go down on an attempted strip squeese instead.

[pescetom]

AL78, on 2023-March-04, 02:36, said:

Finesses win when you have underbid and missed a game. Finesses lose when you have bid aggressively to a tight game or overbid to a bad game.

If you believe your finesses are always losing, look out for elimination and endplays or squeezes instead. Making on a simple finesse is boring, anyone can do that. The expert play is to go down on an attempted strip squeese instead.

And then there are hands where you have the cards and need the tricks, but don't even have a finesse available in the first place.
At least they should trigger your imagination.

[msheald]

Hello! I'm curious about folks who claim that finesses typically work less than 50% of the time.

The 50% successful finesse claim is in all the teaching books that I've seen. Were our forbears so poor at bridge statistics that they made the 50% claim even when it was wrong?

In my robot play, I average about 10% successful finesses. Because of that, it has taught me to look for alternate ways to make tricks in human games. As a result, I am a better bridge player in human games - a great thing!

I suspect that those who claim that BBO robot games have 50% finesse success rates look at all possible finesses in a hand - including those that do not need to be won - and come up with the 50% number.

BBO robot play has a reputation among those I play with as "stacking the deck" against finesses and creating deals with card distributions that occur much more often than by chance. That is why they do not pay money for robot games. Whether these perceptions are right or wrong is beside the point - the perceptions are out there. BBO has not gone after that market, either by publicizing that the perceptions are false (by showing the actual statistics of successful finesses and deal distribution frequencies) or by changing the deals (either globally or in sub-games) if the perceptions are accurate.

A business decision for BBO really, and I support them in that decision. Best regards,

Mike

[smerriman]

msheald, on 2023-November-04, 06:36, said:

BBO has not gone after that market, either by publicizing that the perceptions are false (by showing the actual statistics of successful finesses and deal distribution frequencies)

Uh, this has been done *countless* times. People come here with such a claim, get proven beyond doubt with statistical analysis and even their own archived hands that they are wrong, and then either disappear without acknowledging the response, or say OK, but how about this instead.. and move on to a new theory. Once in a rare while someone actually acknowledges they were wrong

As you can tell from the world in general, conspiracy theorists believe what they believe and tend to ignore all evidence against them no matter how strong the proof. That and confirmation bias (along with a general misunderstanding of what randomness means) are the reasons this perception will never go away.

If you actually look at hundreds of your own hands and count the finesses properly, you won't get 10%, which is a number you just made up (feel free to deny that and present your analysis ). Of course, it may be less than 50% if you're only counting finesses you take, if you're making poor decisions influenced by other facts, like the example I stated in the earlier post above (finessing with AQ in dummy in the suit that the robot lead, so you know they don't have the king). The other half which balances that, when you hold AQ in hand in that suit, you're probably not even mentally counting that as a finesse, just a free trick - but the finesse was going to work.

[johnu]

msheald, on 2023-November-04, 06:36, said:

In my robot play, I average about 10% successful finesses. Because of that, it has taught me to look for alternate ways to make tricks in human games. As a result, I am a better bridge player in human games - a great thing!

See my post #3. Looks like you are not a member of the Super Latinum Platinum Select Members Club. I can only say that if you don't want to pay to be a member, the best percentage strategy is to always play for the drop.

[pescetom]

--

[mycroft]

All I know is that two-way finesses for the relevant Q fail at least 65% of the time.

I'd think that BBO had loaded BridgeMaster code to ensure it when it's not the correct play, but it happens with physical cards too.

[LitaBC]

I would say I lose 90% of finesses. I have gotten to the point where I assume I will lose so I don’t bother. I also don’t bid as high as I would in a real game. It isn’t good for improving in real life games.

[barmar]

LitaBC, on 2024-September-21, 11:40, said:

I would say I lose 90% of finesses. I have gotten to the point where I assume I will lose so I don’t bother. I also don’t bid as high as I would in a real game. It isn’t good for improving in real life games.

Maybe you're taking too many finesses to begin with.

There was a series of articles the past few months in the ACBL Bulletin titled something like "Just (Don't) Do It -- Finesse, That Is". They showed how to take more effective lines of play than finessing, such as ruffing a long suit good, or dummy reversals.