Replies to post #9508 on Biotech Values

[poorgradstudent]

>How do you protect yourself from this kind of result, where the PII data looks good but turns out to be a stastical fluke?<

A new approach I've adopted is to remain skeptical of compounds that only have one phase II trial result.

Your chance of fluking out in one phase II is much higher than fluking out in two independent phase II trials run at different centers.

I think the other important concept to grasp is that the proverbial train almost never leaves the station. Many investors think that they need to be in a full position leading up to phase III results. I think most of the case examples in biotech will show that a relatively modest position ahead of phase III results (with put protection if you want) is the way to go. Stocks almost always revisit their pre-phase III prices.

The additional danger of having a large position ahead of phase III results is that it becomes very difficult to objectively assess an investment. The money that is at risk inevitably requires investors to rationalize away the uncertainties... I find that dangerous.

[DewDiligence]

Zebra's Law revisited:

When you posted about this last December (#msg-4890815), I instructed people to print out your post and hang it on a wall as a reminder. What impressed me was not just the discussion of survivorship bias (which I like to call program-survival bias to make clear that it has nothing to do with patient survival), but also your second point, which I christened Zebra’s Law.

Here are the relevant excerpts from your December post:

>>
1) Survivorship Bias: Take 1000 drug candidates and put them through a phase 1. Just by chance, 100 of these will show statistically significant results at p = 0.1. Take those 100 and do a phase 2. Again, just be random chance, 10 of these 100 will show statistically significant results at p= 0.1. Finally, put these 10 ‘survivors’ into a phase 3, and you will end up very disappointed with the trial results.

2) [Zebra’s Law]: Small, underfunded biotechs simply cannot afford to stop development of their lead drug candidate. Even if they know it is not working, even if they have something 100 times more powerful in the lab, they have to press forward. Otherwise they will not be able to raise more money. So they end up doing trials that should never be done.
<<