## The Bayesian Trap

I didn't say it explicitly in the video, but in my view the Bayesian trap is interpreting events that happen repeatedly as events that happen inevitably. They may be inevitable OR they may simply be the outcome of a series of steps, which likely depend on our behaviour. Yet our expectation of a certain outcome often leads us to behave just as we always have which only ensures that outcome. To escape the Bayesian trap, we must be willing to experiment.

### People who liked this video also liked

### Comments

8 comments posted so far. Login to add a comment.

58

#### 2. thundersnow commented 7 years ago

#1 I love his videos (love him too ) because difficult concepts are explained so well, on a level I can understand without it being too watered down, no loud distracting music either, overall just great lectures, as you mentioned. However I admit that there are still many parts of his lectures that I don't understand, but that's okay, good lectures are like good stories, meant to be listened to again and again, with time I do understand things better. In fact it's been quite a while that I submitted this, so time to watch it again.

20

#### 3. iggyneus commented 7 years ago

It would be interesting if this theorem was applied to the chance of a winning lottery ticket.

I work in a field of absolute data where a person may be told that he has a chance of on to a hundred

of winning / loosing / catching something etc. So this person goes home and is delighted with the Bayesian

formula. However, back at the lab is generally understood that one percent means that exactly one to a hundred.

Every one of a hundred persons will be a winner / or looser depending on the subject matter.

I work in a field of absolute data where a person may be told that he has a chance of on to a hundred

of winning / loosing / catching something etc. So this person goes home and is delighted with the Bayesian

formula. However, back at the lab is generally understood that one percent means that exactly one to a hundred.

Every one of a hundred persons will be a winner / or looser depending on the subject matter.

44

#### 4. kirkelicious commented 7 years ago

#3 Applying Bayesian heuristics to lottery chances does not make too much sense. We can determine the probabilities a priory analytically. We could however use Bayesian methods to determine if a lottery apparatus is biased, but the number of samples we'd need to come to meaningful conclusions would make that pretty impractical.

52

#### 5. Natan_el_Tigre commented 7 years ago

Seems Bayes invented or at least inspired Minesweeper too!

42

#### 6. Austin commented 7 years ago

#4. kirkelicious ‘#3 Applying Bayesian heuristics to lottery chances does not make too much sense.’

Indeed. Nice comment. To understand a chance of winning in a lottery one really just needs to have an understanding of probability, sample size, distribution, randomness, odd, etc etc. Often the probability of winning or odds are presented right on the back of lotto tickets. One has no control over a lotto draw. But that doesn’t stop people ;-) Numerous research papers have shown that in big national lottos that no strategy, i.e. picking a random number is significantly better (ok the odds are still quite low) than trying to apply a ‘sure thing method’ (a stats approach).

Bayesian approaches to probability and calculating probabilistic outcomes – especially with respect to a cost benefit type analysis would be useful to someone trying to decide if he or she should play the lotto versus investing the money elsewhere. I think a card game like poker is a good example where Bayes theorem works nicely. Yes you have odds but you also have players making decisions based on what they believe to be true or possible and reacting to decision made by others doing the same calculations. Logic, odds, is he/she bluffing, what has been played so far.. all of those factors influencing what card you play (or not), how much to bet, when to fold or go all in – now that that of probabilistic thinking and mental calculus can be quite ‘Bayesian’ depending on the person.

Indeed. Nice comment. To understand a chance of winning in a lottery one really just needs to have an understanding of probability, sample size, distribution, randomness, odd, etc etc. Often the probability of winning or odds are presented right on the back of lotto tickets. One has no control over a lotto draw. But that doesn’t stop people ;-) Numerous research papers have shown that in big national lottos that no strategy, i.e. picking a random number is significantly better (ok the odds are still quite low) than trying to apply a ‘sure thing method’ (a stats approach).

Bayesian approaches to probability and calculating probabilistic outcomes – especially with respect to a cost benefit type analysis would be useful to someone trying to decide if he or she should play the lotto versus investing the money elsewhere. I think a card game like poker is a good example where Bayes theorem works nicely. Yes you have odds but you also have players making decisions based on what they believe to be true or possible and reacting to decision made by others doing the same calculations. Logic, odds, is he/she bluffing, what has been played so far.. all of those factors influencing what card you play (or not), how much to bet, when to fold or go all in – now that that of probabilistic thinking and mental calculus can be quite ‘Bayesian’ depending on the person.

27

#### 8. Wotty1 commented 7 years ago

Sadly his Math is not correct.

He assumes that your probability of you having the disease is the same for the probability of the incidence in the population.

OK.......

BUT

You have gone to the doctor with an ailment, the rest of the population have not.

AND

The doctor thinks you may have the disease, they do not think that of the rest of the population

SO

The probability therefore is many times that of the incidence in the population.

So in total:-

Not 9%, more like greater than 90%

He assumes that your probability of you having the disease is the same for the probability of the incidence in the population.

OK.......

BUT

You have gone to the doctor with an ailment, the rest of the population have not.

AND

The doctor thinks you may have the disease, they do not think that of the rest of the population

SO

The probability therefore is many times that of the incidence in the population.

So in total:-

Not 9%, more like greater than 90%

## +6 1. Austin commented 7 years ago