Tomasz Tunguz is a partner at Redpoint Ventures
Tunguz gives us a refresher on the Kelly criterion, named after John Kelly, a researcher at Bell Labs. The principle has two goals:
- Avoid total loss: The investment strategy should never result in a 0 or negative balance.
- Grow the value of the investment as quickly as possible.
First, this requires abandoning the arithmetic mean for its geometric equivalent. An example: Three equally likely investment opportunities have expected returns of 100%, 50%, and 0%. The arithmetic mean, and thus expected value, is 50%: (100+50+0)/3. The geometric mean is 0% (the cubic root of the three percent multiplied together). The Kelly principle says not to invest, while a normal average would say the opposite.
Second, the Kelly criterion tells investors how much to invest. This is expressed as KC = W - (1-W)/R — the probability of winning (W) minus the probability of losing (1-W), divided by the win/loss ratio (R). Investors tend to overestimate their chance of success, meaning that they bet half or three-quarters of the criterion. As this runs the risk of bankruptcy, it is always better to remain conservative.
Jason Lemkin is the founder of the firm SaaStr
Since Lemkin started investing in 2013, he has noticed that one key rule for founders, if they want to win, is to avoid building a VC syndicate where at least two parties do not trust each other.
Breaking this rule can lead to the following:
- Alignment in future rounds will be impossible to build unless the startup is a rocketship that buries all antagonism. This will make closing the round more difficult.
- VCs that have differing opinions on burn rates will provide conflicting advice, making a founder’s job harder.
- Table games, such as firms secretly trying to buy out ex-employees or other investors out at love prices, will proliferate. While not fatal, this can lead to lawsuits.
- If the board and investors have no way of de-escalating, issues will simply not get resolved. When lines of communication are present, most issues can be easily talked through, as they tend to not be anything new anyway.
To improve trust:
- Founders should meet with the late-stage investors that early investors recommend, even if no deals are made.
- Founders need to make room for pro-rata and existing investors in future rounds, even if not 100% participate, and be aware of the trade-offs of bringing new investors on board. A ‘sharp elbow’ round, where existing investors are not consulted, will immediately result in broken trust.
- Investors who do not even reach out to current investors are to be avoided. Those who are looking to play on a team will make this courtesy call.
- Board meetings should include invitations for all investors, not just one big investor.
Did you enjoy reading my summary of Tunguz's and Lemkin's pieces? Normally, VC Masterclasses are only available to premium subscribers. Upgrade to see all my Inside VC Premium features for either $10/month or $100 billed annually! Previous Masterclass features have included pieces by Elad Gil, Tomasz Tunguz, Daniel Li, Sree Kolli, Benjamin Beltzer, Steve Blank, Sammy Abdullah, Bill Gurley, and others. Click here to upgrade to premium!