I am going to add the probability that the experiment will prove the life extension effect of the metformin, something like 0.5. However, even if the experiment fails, the start of the experiment will open the ways to other similar experiments, some of which will probably reach success. The main role of the TAME is that it is icebreaker precedent for the future research.
If you calculate expected value with numbers than it makes sense to use numbers that don’t seem fishy. Currently, those in your abstract don’t look good.
I reread Gwern and it was a year and not less a year but it was on correlational evidence and most of the time clinical trials find that drugs do less well then expected.
Your paper talks about how Gwern estimates 1 year and you estimate 3 years. It’s not clear to me why you use 3 years.
In your article you speak about the cost for a Metformin trials is estimated at $65 million. That’s in conflict with the claim in the abstract that it only costs $60 million.
If I remember correctly you also need two trials and not one to get a drug approved by the FDA. If the $65 million is actually for two trials, say so, otherwise you should calculate more costs for getting Metformin approved than the $65 million.
What’s the outside view of phrase III trials showing that a drug does what the creator wants? I would guess that it’s less than 0.5, so those seem optimistic as well.
GiveWell generally tries to use conservative numbers and I think your case would benefit from also using more conservative numbers even if it means that the cost per live saved rise by an order of magnitude.
3 years was the estimation of the expected effect by the main experimenter behind TAME - Nick Barzilai. This is what he expects to see. But I think more conservative will be to use Gwern’s estimation of 1 year, and I will turn to this estimate.
Theres is, in fact, another smaller trial of metformin for life extension called MILES, but it is not longitudinal study, just 3 weeks, but it has funding and preliminary FDA approval. Thus if TAME will happen, there will be two studies.
I am going to add the probability of success in the next version of the estimation and update the price of study (they told that they collected part of the money).
I am going to add the probability that the experiment will prove the life extension effect of the metformin, something like 0.5. However, even if the experiment fails, the start of the experiment will open the ways to other similar experiments, some of which will probably reach success. The main role of the TAME is that it is icebreaker precedent for the future research.
If you calculate expected value with numbers than it makes sense to use numbers that don’t seem fishy. Currently, those in your abstract don’t look good.
I reread Gwern and it was a year and not less a year but it was on correlational evidence and most of the time clinical trials find that drugs do less well then expected.
Your paper talks about how Gwern estimates 1 year and you estimate 3 years. It’s not clear to me why you use 3 years.
In your article you speak about the cost for a Metformin trials is estimated at $65 million. That’s in conflict with the claim in the abstract that it only costs $60 million.
If I remember correctly you also need two trials and not one to get a drug approved by the FDA. If the $65 million is actually for two trials, say so, otherwise you should calculate more costs for getting Metformin approved than the $65 million.
What’s the outside view of phrase III trials showing that a drug does what the creator wants? I would guess that it’s less than 0.5, so those seem optimistic as well.
GiveWell generally tries to use conservative numbers and I think your case would benefit from also using more conservative numbers even if it means that the cost per live saved rise by an order of magnitude.
3 years was the estimation of the expected effect by the main experimenter behind TAME - Nick Barzilai. This is what he expects to see. But I think more conservative will be to use Gwern’s estimation of 1 year, and I will turn to this estimate.
Theres is, in fact, another smaller trial of metformin for life extension called MILES, but it is not longitudinal study, just 3 weeks, but it has funding and preliminary FDA approval. Thus if TAME will happen, there will be two studies.
I am going to add the probability of success in the next version of the estimation and update the price of study (they told that they collected part of the money).