3 ms·

Mathematically speaking, this analysis is almost worthless. The inputs are not quantative in any way - he assigns values on a 1 to 5 scale, based on how he feel

Mathematically speaking, this analysis is almost worthless. The inputs are not quantative in any way - he assigns values on a 1 to 5 scale, based on how he feels company to company rather than taking some representative metrics. Beyond that, he has for some reason assumed that the error here is normally distributed (which it plainly is not as data is discrete) and applied OLS techniques to find his coefficients. There is no discussion on statistical significance, on model construction, or on any subject which would allow this to be called quantative analysis. Even his r^2, the fallback number for people who don't understand stats to demonstrate how 'good' their model is, is woefully bad. There is no insight here that would not have been given by expressing the same sentiments about pitch/investor fit without recourse to pseudo-statistics.

einhverfr 14y ago
I think they are meaningful. Basically what he is doing is identifying subjective impressions and approximately how they affected the final result. This is pretty helpful because what it shows to me is that if you are pitching to him, you want to emphasize team and market potential, and then touch on product and traction. Where it would be less helpful is in trying to use it to measure performance of the VC's funds since these things are subjectively rather than objectively measured.
joshu 14y ago
I have made money on much worse rsq. That said, I agree. There is no Team score, it is really BlueRun's Estimator Of Team Score, which has its own error, etc etc etc. 200 samples is not enough to back out 20 variables. A better model might be constructed around a decision tree mechanism but I have less familiarity with that. Finally it probably doesn't include statistics around which companies actually got money from some VC not just this VC. I suspect this is MBA statistics -- only statistics that can be done in Excel. Anyone want to find his bio?
njx 14y ago
The other thing to note "all funded pitches were mobile centric"

einhverfr 14y ago

I think they are meaningful. Basically what he is doing is identifying subjective impressions and approximately how they affected the final result. This is pretty helpful because what it shows to me is that if you are pitching to him, you want to emphasize team and market potential, and then touch on product and traction. Where it would be less helpful is in trying to use it to measure performance of the VC's funds since these things are subjectively rather than objectively measured.

joshu 14y ago

I have made money on much worse rsq. That said, I agree. There is no Team score, it is really BlueRun's Estimator Of Team Score, which has its own error, etc etc etc. 200 samples is not enough to back out 20 variables. A better model might be constructed around a decision tree mechanism but I have less familiarity with that. Finally it probably doesn't include statistics around which companies actually got money from some VC not just this VC. I suspect this is MBA statistics -- only statistics that can be done in Excel. Anyone want to find his bio?

njx 14y ago

The other thing to note "all funded pitches were mobile centric"