Replies to post #92234 on Avid Bioservices Inc (CDMO)

[iwfal]

At the current P value the odds of this Phase II study being duplicated favorably in a larger Phase III study of similar design are 98%.
I don't think that's true. The probability of 98% is based on duplicating the same trial and having it reach the same HR of 0.524. Since their new phase 3 trial will be designed to return statistically significant results at a much higher HR, it makes the probability of success much greater than 98%.

You are committing the the Base Rate Fallacy. If you:

a) have 1000 drugs entering their first randomized trial

b) only 1% are useful and the rest are no better than sugar pill

c) a drug shows stat sig (p=0.04) in this - its first randomized trial.

Then what is the chance that the drug is no better than placebo?

The answer is NOT that that there is only a 2% chance (1/2 of the 4%) that it is just placebo. The chance that it is placebo is much much higher than that.

[iwfal]

At the current P value the odds of this Phase II study being duplicated favorably in a larger Phase III study of similar design are 98%.
I don't think that's true. The probability of 98% is based on duplicating the same trial and having it reach the same HR of 0.524. Since their new phase 3 trial will be designed to return statistically significant results at a much higher HR, it makes the probability of success much greater than 98%.

The calculation you just made for the chance of success for PPHM is incorrect because you are committing what is called the Base Rate Fallacy. To illustrate, if you:

a) have 1000 drugs entering their first randomized trial

b) only 1% are useful and the rest are no better than sugar pill

c) a drug shows stat sig (p=0.04) in this - its first randomized trial.

Then what is the chance that the drug is no better than placebo?

The answer is NOT that that there is only a 2% chance (1/2 of the 4%) that it is just placebo. The chance that it is placebo is much much higher than that.

Obviously we can argue about what the odds are for PPHM, but the one thing we can be sure of it that the odds are not just 'less than 2% that the next trial will fail'.

PS Note this is a redo of a post that was deleted. I generously assume that it was deleted due to a board moderator not understanding how Base Rate Fallacy applied to PPHM - so in this redo I was crystal clear how it applies.

[Robert C Jonson]

I don't know why iwfal's posts keep getting deleted, but a base rate fallacy seems to me to be something to be considered. What I infer from reading his post and looking at base rate fallacy in wikipedia is that one cannot make an accurate prediction about the success of Bavi in its phIII 2nd line NSCLC trial. Having said that, I still think it's likely to succeed.

Just for the record, I don't think iwfal's post should have been deleted, as it is not biased against PPHM and is a consideration for us all, even though it is not a favorable outlook for this stock. Is our board really only going to devolve to a forum only for cheerleaders?