Suppose there is a consultancy that is trying to help many VCs. To keep things simple, it
Question:
Suppose there is a consultancy that is trying to help many VCs. To keep things simple, it classifies start-ups according to whether or not they show promise to eventually yield an operating profit. Over its lifetime, the consultancy has had a decent track record. Of the start-ups which ultimately showed a profit, it classified 75% as promising. Likewise, of the start-ups which never showed an operating profit, it classified 75% as not promising.
The "base rate" of eventual operating profitability is, of course, low. Suppose that in the sector in which the consultancy operates and with the level and stature of the start-ups it engages, this rate is less than 10%.
If the consultancy classifies a specific start-up as promising, roughly what probability might someone who engages in prediction by evaluation place on the start-up eventually showing an operating profit? How does this probability compare to the probability that would be derived from a proper statistical prediction?