Kelly-Style Risk Sizing
Size speculative bets so one bad run cannot wipe you out
- Difficulty
- Moderate
- Time to result
- ~days to results
- Steps
- 6
- Confidence
- 94%
Kelly-style risk sizing separates confidence in an idea from the amount of capital placed behind it. The investor estimates the probability and payoff of success, considers the loss if the thesis fails, and chooses a stake small enough to survive an unfavorable sequence. The mechanism recognizes that even a genuinely favorable wager can produce several losses in a row. Rather than going all in, the investor preserves most capital while retaining enough exposure to benefit from a large asymmetric payoff. In extreme scenarios, only a tiny position may be necessary: if a security could plausibly rise hundreds of times, a few shares can provide meaningful upside without putting a home, retirement fund, or financial security at risk.
Origin
Extracted from Masters in Business, where Spencer Jakab applies the Kelly betting criterion to meme-stock investors who risked their entire wealth on extreme price targets.
Core principles
- 01Confidence never eliminates uncertainty
- 02Survival matters more than maximizing one payoff
- 03A favorable edge can still produce a losing streak
- 04Exposure should reflect both probability and downside
- 05A small position can capture an asymmetric payoff
How to run it
- 1
State the thesis
Write down the event you expect, its estimated probability, and the evidence supporting that estimate. Distinguish genuine evidence from enthusiasm or group consensus.
Pro tip Express confidence as a probability range rather than a certainty.
Watch out A vivid story can feel much more probable than it is.
- 2
Map the payoff
Calculate the plausible gain if correct and the maximum loss if wrong. Include leverage, borrowing costs, taxes, and the possibility of total loss.
Pro tip Use conservative exit prices rather than the most exciting target.
Watch out Options and short positions can behave differently from ordinary stock ownership.
- 3
Stress-test losing streaks
Model several consecutive losses even when each individual wager appears favorable. Determine whether that sequence would impair essential savings or prevent future participation.
Pro tip Test longer losing runs than intuition suggests are likely.
Watch out A positive expected value does not prevent short-term ruin.
- 4
Set a survivable stake
Limit the position to a fraction of available risk capital so the modeled losing streak remains tolerable. Never use money required for housing, bills, or near-term obligations.
Pro tip Use less than the theoretical maximum when probability estimates are uncertain.
Watch out Mortgaging a home turns a speculative mistake into a life-changing loss.
- 5
Preserve asymmetric exposure
Buy only enough exposure to produce a meaningful outcome if the extreme thesis succeeds. When the claimed payoff is enormous, a very small position may be sufficient.
Pro tip Ask how few shares or contracts would still matter if the target were reached.
Watch out Owning more does not make the thesis more likely.
- 6
Reassess without doubling down
Review whether evidence, probability, or payoff has changed. Reduce or exit when the original assumptions fail instead of using losses as automatic justification to add capital.
Pro tip Schedule reviews before entering the position.
Watch out Online loyalty narratives can turn risk management into a perceived betrayal.
In the wild
An investor believes a low-priced stock could reach an extraordinary price but recognizes that the probability is small. Instead of mortgaging a home or committing an entire portfolio, the investor buys only two shares. If the extreme target occurs, the payoff is still meaningful; if the thesis fails, the loss is trivial.
→ The investor retains asymmetric upside without threatening financial security.
A bettor believes an independently tested model wins more often than it loses. Before staking heavily, the bettor models a run of eight or ten losses and caps each wager at a small percentage of capital.
→ The bettor survives normal variance and can continue using the edge.
Common mistakes
Confusing confidence with certainty
Strong evidence may improve the odds, but it does not eliminate the possibility of failure or a long adverse run.
Going all in on an asymmetric bet
An enormous potential payoff usually means only a small exposure is needed; risking everything adds ruin without proportionate utility.
Using essential assets as collateral
Borrowing against a home or required savings converts a bounded speculation into a threat to basic financial stability.
Is it for you?
Best for
It is best for investors considering concentrated, leveraged, or highly speculative positions.
Not ideal for
It is not ideal for people who cannot estimate probabilities or who will continually increase a losing position.
From the transcript
“if you are really sure, like you really have a high degree of confidence. Uh, Like I just went back to the Kelly betting criteria.”
“you don't bet all your wealth. These people are going all in.”
“you should never bet more than like four percent of your steak because you could have a run of even with weighted coin of like…”
From the episode
Spencer Jakab on the Death of Meme Stocks
Spencer Jakab