3 ms·

Probabilistic model for measuring success

4 points by nurall 20y ago

rfrey 20y ago
P(S) = i/A + j(N/3) + k(1/Q)^2 + x((1/GT)^3) A = Average age of founders N = Number of founders Q = Quality of idea, 0-1 G = Graham Quotient: How hard you'll work, 0-1 T = Taleb Quotient: How lucky you'll get i, j, k, x: Adjust to fit the sample of 5 famous startups in the news this week Repeat as required
- nurall 20y ago
  Good one!!! Did you model this? :-).. interesting... you are saying that if P(S) is greater than 1, P(S) doesn't apply to that particular startup.. its meaningless.. and they should stop wasting their time.. is dat a valid argument to why your metric could be greater than 1?
nurall 20y ago
I am sure there are a lot of management books out there that define SUCCESS, and I am sure they have their respective models to describe the probability of SUCCESS. I was just thinking about modeling the probability from a startup perspective. Following are the assumptions this model makes - 1. It takes 'n' number of iterations before an idea is a HIT, where n is greater or equal to 1 2. There is a probablity of success associated with each iteration, which implies that there is also a probablity of failure associated with each iteration.. duh!! Model 1: Using geometric distribution This is the most conservative and simplistic approach to model success p(HIT) = Probability of success per iteration P(SUCCESS) = p(HIT) x (1 - p(HIT))^n This is simple in the sense that each new iteration does not carry the benefits from the previous iteration Model 2: Bernoulli's trials based model This probably is more realistic of the two models p(HITn) = Probability of success in the nth iteration = P(HITn-1) +/- deltaP Similarly, p(~HITn) = Probability of NO success in the nth iteration = P(~HITn-1) +/- (1-deltaP) where deltaP = Probability of iteration (n-1) increasing/decreasing the probability of success in the nth iteration P(SUCCESS) = p(HIT0) x p(HIT1) x ... x p(HITn-1) x p(HITn) = 1 - p(~HIT0) x p(~HIT1) x ... x p(~HITn-1) x p(~HITn) where p(HITn) tends to 1 AND p(~HITn) tends to 0, We need to be deterministic if we have to be optimistic... ;-) Any thoughts on this??? Do you think this model is right/wrong?? Do you have your own models??? Please do share!!!

4 points by nurall 20y ago

rfrey 20y ago

P(S) = i/A + j(N/3) + k(1/Q)^2 + x((1/GT)^3) A = Average age of founders N = Number of founders Q = Quality of idea, 0-1 G = Graham Quotient: How hard you'll work, 0-1 T = Taleb Quotient: How lucky you'll get i, j, k, x: Adjust to fit the sample of 5 famous startups in the news this week Repeat as required

nurall 20y ago
Good one!!! Did you model this? :-).. interesting... you are saying that if P(S) is greater than 1, P(S) doesn't apply to that particular startup.. its meaningless.. and they should stop wasting their time.. is dat a valid argument to why your metric could be greater than 1?

nurall 20y ago

I am sure there are a lot of management books out there that define SUCCESS, and I am sure they have their respective models to describe the probability of SUCCESS. I was just thinking about modeling the probability from a startup perspective. Following are the assumptions this model makes - 1. It takes 'n' number of iterations before an idea is a HIT, where n is greater or equal to 1 2. There is a probablity of success associated with each iteration, which implies that there is also a probablity of failure associated with each iteration.. duh!! Model 1: Using geometric distribution This is the most conservative and simplistic approach to model success p(HIT) = Probability of success per iteration P(SUCCESS) = p(HIT) x (1 - p(HIT))^n This is simple in the sense that each new iteration does not carry the benefits from the previous iteration Model 2: Bernoulli's trials based model This probably is more realistic of the two models p(HITn) = Probability of success in the nth iteration = P(HITn-1) +/- deltaP Similarly, p(~HITn) = Probability of NO success in the nth iteration = P(~HITn-1) +/- (1-deltaP) where deltaP = Probability of iteration (n-1) increasing/decreasing the probability of success in the nth iteration P(SUCCESS) = p(HIT0) x p(HIT1) x ... x p(HITn-1) x p(HITn) = 1 - p(~HIT0) x p(~HIT1) x ... x p(~HITn-1) x p(~HITn) where p(HITn) tends to 1 AND p(~HITn) tends to 0, We need to be deterministic if we have to be optimistic... ;-) Any thoughts on this??? Do you think this model is right/wrong?? Do you have your own models??? Please do share!!!