4 ms·

(I'm a VC.) I'm skeptical of simulations like this one because they typically assume that every startup has the same statistical distribution of returns, and wh

(I'm a VC.) I'm skeptical of simulations like this one because they typically assume that every startup has the same statistical distribution of returns, and whether you pick 15 or 500 startups, each startup will have the same expected value and variance. But in practice it doesn't work like that. A full time VC might see 2000 pitch decks per year, meet with a few hundred of those companies, and end up investing in 5. So they are picking what they perceive to be the top .25% of all pitches they see. If they invested in 20 companies, it would be the top 1%. If they invested in 200 companies, it would be the top 10%. But because these companies are ranked, the expected outcome distribution of the VC's #1 company should be a little better than the #10 company, and a lot better than the #200 company. The further down the list you go, the worse the expected value -- even though some of the lower ranked companies may turn out to be amazing. So you can't just 40x the number of investments you make and say "expected value is the same but variance is now much lower!!1!" because you would be lowering your bar a lot by expanding the # of investments. And your variance would go down considerably, but so would your expected value.

zzleeper 6y ago
(PhD Finance here) Even classical 70s-era finance is against this article: Rich investors don't need their VC to be super diversified. They can just invest into multiple VC funds and get diversification that way. Also, it doesn't matter how much risk is in a given VC fund, what matters is how investing part of my porfolio in that VC fund affects the risk of my portfolio!
- Tomminn 6y ago
  This was the first thing that occurred to me. Why do they need the VC firms to change from an investment perspective? Just buy the asset class as a whole. The lesson is VC funds are more like individual stocks than index funds from an investment perspective. Which is fine. In fact VC funds as a class are like a great class of individual stocks, from an investment perspective. The problem isn't from an investment perspective at all. It's from the VC perspective. VC's which invest in small numbers of companies are trusting the future of their fund on the role of a dice. That is risk they might want to eliminate for their own good.
erispoe 6y ago
Your explanation makes sense but is predicated on the VC being able to reliably rank these companies along expected returns. The perfect VC with perfect foresight might be able to pick just 1 investment and get the best returns from that. If returns are distributed on a power-law distribution (with the top performer returning a multiple of the second one and so on), then any deviation between the VC ranking and the outcome distribution of returns is very costly. What if I only invest in my predicted top 5, but the real top 1 is ranked 6 in my estimate? If what matters is to reliably capture the top performers, then the perfect VC would get the top 5 with only 5 investments. But for imperfect VCs, it might be worth it to invest in 50 just to approach 100% chance of capturing these 5 top performers, the 45 others are just the cost of doing that.
- lpolovets 6y ago
  > But for imperfect VCs, it might be worth it to invest in 50 just to approach 100% chance of capturing these 5 top performers, the 45 others are just the cost of doing that. Yep! This is exactly how most funds operate. They invest in 30 or 40 companies over 3-4 years with the hope that 30-40 is enough to hit 1-2 big winners. The more confident a VC is in their picking skills, the more likely they are to invest in 15 or 25 companies in a fund, instead of 30+.
cinntaile 6y ago
How do you decide the number of companies you invest in each year? Is this simply a function of the available capital or do you take future capital requirements of the companies you want to invest in into account? Let's say you only pick the top 0.25%, do you do any statistical analysis afterwards to see if this percentage gave the best ROI? Because I would assume that the top 1% at least gets funded by other VCs so it should be possible to do this? Although at the same time I wonder if outsized returns can really be made if there is a VC consensus of the top 1%, my understanding is that the best returns are made in firms that people disagree on.
- lpolovets 6y ago
  Good returns can come both from consensus and nonconsensus companies. For example, Robinhood has publicly spoken about the large number of investors that passed on their seed round, while Flexport has been inundated with interest during every funding round. One big constraint on the # of investments per year for a fund is the amount of time that the fund's partners have. Investors want to differentiate themselves with the help they can provide, and help takes time. As a result, a Series A investor/board member can manage 1-2 new investments per year, while a seed investor might be able to manage 3-8 new investments per year. So # of partners * ~5 investments per year is a good estimate of how many investments a hands on seed fund will do. There are funds that invest in a lot more companies, but they tend to write small checks and are less hands on.
jariel 6y ago
This is a good point, but I think the point being, you'd 5x your staff and 5x the size of your funnel, thereby ostensibly keeping the same ratio of 'good deals'. Obviously, at some scale, you literally tap out of opportunities ... but not at just 2K pitches.
rorykoehler 6y ago
This was my initial thought too but it is tempered by the "deals we passed on" meme that is prevalent in VC circles. A VCs perception of the best companies is not based on reality. Like a gambler a VC has no idea whether their bet is the right one until much later. Sure you can de-risk it based on previous exits of founder etc but the reality is every VC probably passes on more great companies than they invest in.
- lpolovets 6y ago
  Agreed that VCs pass on way more good companies than they invest in. But if a VC invests in .5% of companies they look at, and there are 3 great companies in that .5% and 15 great companies in the other 99.5%, that suggests the VC's ranking is pretty good on a relative basis.
  - rorykoehler 6y ago
    It's impossible to know as getting VC investment increases the odds of success or at least surviving to the point where you can even attempt to have a go. Clearly some are better than others at choosing winners. Part of me thinks that this could be driven by network more than competence though.
nyrulez 6y ago
They could've done an analysis on the variance of returns categorized by # of investments per fund. That would illuminate if # of investments is a big correlational/causative factor as they claim. A simulation alone isn't going to cut it as it involves too many assumptions though it can give rise to reasonable hypothesis (not conclusions).
georgeecollins 6y ago
>> I'm skeptical of simulations Yeah, no kidding. Once you see a monte carlo simulation your eyes should roll. This is like all financial predictions that have a bell curve outcome hidden in them. Dice are easy to predict when you roll them many times. Finance is just like dice, just ask the Nobel Prize winners at Long Term Capital Management.
- mmmbn 6y ago
  LTCM didn’t end quite so well.
mikekoscinski 6y ago
Right. It seems commonly misunderstood that the universe of investable startups represents a zero-sum game for investors. It isn't as simple as "invest in more companies"; one fund's investment represents another fund's forgone investment. (Not many good companies want a five-fund Series A where each fund buys 20%.) Having the capital and desire to make an investment is the easy part of investing. Actually earning that allocation is difficult. Yet, most conceptions of investing seem to be shaped by public markets and their ubiquity. That is to say, public market float is taken for granted - when in reality it is missing from every asset class (practically or literally) save for G7 (+China) secondary-market public equities (and now cryptocurrencies).
throwaway2245 6y ago
Even if we agree that a VC is able to rank investments by high EV and low (non-zero) variance, it's still the case that a small number of investments poses a higher risk than optimal.
whoisnnamdi 6y ago
Leo - as other folks have noted it is a major assumption to think that VCs are good at calculating the expected values of companies in which they make investments. The fundamental randomness is so high here that any ranking a VC made would likely be garbage - the sample size of investments is not nearly high enough to come to a precise viewpoint on expected value of one good deals vs another. As an analogy, if this was a regression the standard errors would be massive, so much so that each investment would not be statistically significantly different from one another in terms of expected return. I think it’s plausible that they’d be able to distinguish best from absolute worst, but best from slightly less good? No way. I think a more realistic “model” of a VCs behavior is one of thresholding - set the bar somewhere and invest in any company that seems to be above that quality bar, with some stochasticity as to which deals you actually win in the end.
- texasbigdata 6y ago
  Go backwards. How many IPOs have not had VC backing? Obviously maybe they grow slower, or maybe IPO bound tech firms are oversubscribed so it’s only an individual VCs prediction ability that’s wrong not the asset class as a whole, but still.
- lpolovets 6y ago
  Hey Namdi! > it is a major assumption to think that VCs are good at calculating the expected values of companies in which they make investments. It's not quite the same as proving VCs can calculate expected values well, but VC is one of the few asset classes with persistence of returns -- meaning having a top quartile fund is correlated with the next fund also being top quartile. In most investment classes, being top quartile across funds is basically uncorrelated. So good VCs tend to be reliably good and not good just because of pure chance. Source: https://www.morganstanley.com/im/publication/insights/articles/articles_publictoprivateequityintheusalongtermlook_us.pdf?1596549853128 https://www.morganstanley.com/im/publication/insights/articl... (p 51) > The fundamental randomness is so high here that any ranking a VC made would likely be garbage A ranking doesn't have to be perfect, just better than random. E.g. let's say you give me a bunch of companies and ask me to rank them, and I put 60% of the actual top 1% and 40% of the next 1% into my top 1%. I'm making mistakes, but that's still pretty good. As long as I'm directionally right on average, I should do well on the investing side.

zzleeper 6y ago

(PhD Finance here) Even classical 70s-era finance is against this article: Rich investors don't need their VC to be super diversified. They can just invest into multiple VC funds and get diversification that way. Also, it doesn't matter how much risk is in a given VC fund, what matters is how investing part of my porfolio in that VC fund affects the risk of my portfolio!

Tomminn 6y ago
This was the first thing that occurred to me. Why do they need the VC firms to change from an investment perspective? Just buy the asset class as a whole. The lesson is VC funds are more like individual stocks than index funds from an investment perspective. Which is fine. In fact VC funds as a class are like a great class of individual stocks, from an investment perspective. The problem isn't from an investment perspective at all. It's from the VC perspective. VC's which invest in small numbers of companies are trusting the future of their fund on the role of a dice. That is risk they might want to eliminate for their own good.

erispoe 6y ago

Your explanation makes sense but is predicated on the VC being able to reliably rank these companies along expected returns. The perfect VC with perfect foresight might be able to pick just 1 investment and get the best returns from that. If returns are distributed on a power-law distribution (with the top performer returning a multiple of the second one and so on), then any deviation between the VC ranking and the outcome distribution of returns is very costly. What if I only invest in my predicted top 5, but the real top 1 is ranked 6 in my estimate? If what matters is to reliably capture the top performers, then the perfect VC would get the top 5 with only 5 investments. But for imperfect VCs, it might be worth it to invest in 50 just to approach 100% chance of capturing these 5 top performers, the 45 others are just the cost of doing that.

lpolovets 6y ago
> But for imperfect VCs, it might be worth it to invest in 50 just to approach 100% chance of capturing these 5 top performers, the 45 others are just the cost of doing that. Yep! This is exactly how most funds operate. They invest in 30 or 40 companies over 3-4 years with the hope that 30-40 is enough to hit 1-2 big winners. The more confident a VC is in their picking skills, the more likely they are to invest in 15 or 25 companies in a fund, instead of 30+.

cinntaile 6y ago

How do you decide the number of companies you invest in each year? Is this simply a function of the available capital or do you take future capital requirements of the companies you want to invest in into account? Let's say you only pick the top 0.25%, do you do any statistical analysis afterwards to see if this percentage gave the best ROI? Because I would assume that the top 1% at least gets funded by other VCs so it should be possible to do this? Although at the same time I wonder if outsized returns can really be made if there is a VC consensus of the top 1%, my understanding is that the best returns are made in firms that people disagree on.

lpolovets 6y ago
Good returns can come both from consensus and nonconsensus companies. For example, Robinhood has publicly spoken about the large number of investors that passed on their seed round, while Flexport has been inundated with interest during every funding round. One big constraint on the # of investments per year for a fund is the amount of time that the fund's partners have. Investors want to differentiate themselves with the help they can provide, and help takes time. As a result, a Series A investor/board member can manage 1-2 new investments per year, while a seed investor might be able to manage 3-8 new investments per year. So # of partners * ~5 investments per year is a good estimate of how many investments a hands on seed fund will do. There are funds that invest in a lot more companies, but they tend to write small checks and are less hands on.

jariel 6y ago

This is a good point, but I think the point being, you'd 5x your staff and 5x the size of your funnel, thereby ostensibly keeping the same ratio of 'good deals'. Obviously, at some scale, you literally tap out of opportunities ... but not at just 2K pitches.

rorykoehler 6y ago

This was my initial thought too but it is tempered by the "deals we passed on" meme that is prevalent in VC circles. A VCs perception of the best companies is not based on reality. Like a gambler a VC has no idea whether their bet is the right one until much later. Sure you can de-risk it based on previous exits of founder etc but the reality is every VC probably passes on more great companies than they invest in.

lpolovets 6y ago
Agreed that VCs pass on way more good companies than they invest in. But if a VC invests in .5% of companies they look at, and there are 3 great companies in that .5% and 15 great companies in the other 99.5%, that suggests the VC's ranking is pretty good on a relative basis.
- rorykoehler 6y ago
  It's impossible to know as getting VC investment increases the odds of success or at least surviving to the point where you can even attempt to have a go. Clearly some are better than others at choosing winners. Part of me thinks that this could be driven by network more than competence though.

nyrulez 6y ago

They could've done an analysis on the variance of returns categorized by # of investments per fund. That would illuminate if # of investments is a big correlational/causative factor as they claim. A simulation alone isn't going to cut it as it involves too many assumptions though it can give rise to reasonable hypothesis (not conclusions).

georgeecollins 6y ago

>> I'm skeptical of simulations Yeah, no kidding. Once you see a monte carlo simulation your eyes should roll. This is like all financial predictions that have a bell curve outcome hidden in them. Dice are easy to predict when you roll them many times. Finance is just like dice, just ask the Nobel Prize winners at Long Term Capital Management.

mmmbn 6y ago
LTCM didn’t end quite so well.

mikekoscinski 6y ago

Right. It seems commonly misunderstood that the universe of investable startups represents a zero-sum game for investors. It isn't as simple as "invest in more companies"; one fund's investment represents another fund's forgone investment. (Not many good companies want a five-fund Series A where each fund buys 20%.) Having the capital and desire to make an investment is the easy part of investing. Actually earning that allocation is difficult. Yet, most conceptions of investing seem to be shaped by public markets and their ubiquity. That is to say, public market float is taken for granted - when in reality it is missing from every asset class (practically or literally) save for G7 (+China) secondary-market public equities (and now cryptocurrencies).

throwaway2245 6y ago

Even if we agree that a VC is able to rank investments by high EV and low (non-zero) variance, it's still the case that a small number of investments poses a higher risk than optimal.

whoisnnamdi 6y ago

Leo - as other folks have noted it is a major assumption to think that VCs are good at calculating the expected values of companies in which they make investments. The fundamental randomness is so high here that any ranking a VC made would likely be garbage - the sample size of investments is not nearly high enough to come to a precise viewpoint on expected value of one good deals vs another. As an analogy, if this was a regression the standard errors would be massive, so much so that each investment would not be statistically significantly different from one another in terms of expected return. I think it’s plausible that they’d be able to distinguish best from absolute worst, but best from slightly less good? No way. I think a more realistic “model” of a VCs behavior is one of thresholding - set the bar somewhere and invest in any company that seems to be above that quality bar, with some stochasticity as to which deals you actually win in the end.

texasbigdata 6y ago
Go backwards. How many IPOs have not had VC backing? Obviously maybe they grow slower, or maybe IPO bound tech firms are oversubscribed so it’s only an individual VCs prediction ability that’s wrong not the asset class as a whole, but still.
lpolovets 6y ago
Hey Namdi! > it is a major assumption to think that VCs are good at calculating the expected values of companies in which they make investments. It's not quite the same as proving VCs can calculate expected values well, but VC is one of the few asset classes with persistence of returns -- meaning having a top quartile fund is correlated with the next fund also being top quartile. In most investment classes, being top quartile across funds is basically uncorrelated. So good VCs tend to be reliably good and not good just because of pure chance. Source: https://www.morganstanley.com/im/publication/insights/articles/articles_publictoprivateequityintheusalongtermlook_us.pdf?1596549853128 https://www.morganstanley.com/im/publication/insights/articl... (p 51) > The fundamental randomness is so high here that any ranking a VC made would likely be garbage A ranking doesn't have to be perfect, just better than random. E.g. let's say you give me a bunch of companies and ask me to rank them, and I put 60% of the actual top 1% and 40% of the next 1% into my top 1%. I'm making mistakes, but that's still pretty good. As long as I'm directionally right on average, I should do well on the investing side.