Replies to post #128852 on Biotech Values

[iwfal]

I tested positive, and know my chances of having the disease....

There is a 70% chance I have the disease, and a 30% chance that I don't.

This is incorrect and is the source error that almost everyone makes - that Test False Positive+ Test True Positive=1. Or, put another way - that Test False Positive Proportion equals Population False Population Proportion. These are both false assumptions. See Base Rate Fallacy and False Positive Paradox .

More particularly the real formula for the Risk-that-if-you-test-positive-you-really-have-the-disease is I*(1-FN)/{(1-I)*FP+(1-FN)*I}

So, why does this matter in biotech investing? Because:

1) Humans are STRONGLY predisposed to misjudge the odds when the incidence rate is lower than the false positive rate. In particular they will vastly over-estimate the odds that a positive test means they have the real thing.

2) In the case of biotech investing the incidence rate (the percent of drugs entering into ph iii that turn out to have good efficacy) is very low - and the rate of false positives is very high (almost all those drugs entering ph iii had some kind of 'positive' indicator).

3) Therefore experienced biotech investors learn, over time, that they need to hugely decrease the False Positive rate. I.e. they are much much more pessimistic than a newbie would think justified. 0.03 on a subgroup sounds great to a newbie - but still has way too high a false positive rate for an experience biotech investor.

[DewDiligence]

One more wrinkle on iwfal’s problem...

In #msg-68160499 you said:

…the false negative rate doesn't matter. I tested positive.

In #msg-68161042 you said

You're right. I forgot to add the 0.9% to the denominator.

Actually, the 0.9% quantity you forgot to incorporate when calculating the positive predictive value of the test depends on the false-negative rate as well as the false-positive rate—see the second paragraph of #msg-68160330 where the 0.9% figure is derived. This is yet another counter-intuitive aspect of the kind of data analysis that is important to biotech investors. Regards, Dew