The room I miss most is the one with the trading desk on Monday morning. Not for the markets — for the meeting. Eight people around a table, each with a position, each with a thesis, and one rule: nobody is allowed to say how convinced they are without saying how much they have on the trade.
The Monday morning equivalent in a commercial committee tends to involve eight people, each with a position, each with a thesis, and a slide deck. Conviction is expressed in adjectives. The size of the bet is rarely named, and almost never compared to the trader’s edge on the call.
Two committees, same problem
Trading desks and commercial committees are doing the same job. They are taking positions, under uncertainty, with someone else’s capital, against a future they cannot observe. The desk is sized in basis points and the committee is sized in budget lines, but the mathematical problem is identical: how much to put on the trade, given how confident you actually are.
The desk uses some version of the Kelly criterion or its sober cousins to translate edge into position size. Crudely: if you believe the trade has a positive expected value of x with a standard deviation of σ, the right position size is proportional to x/σ2. The implication is that conviction without volatility is what justifies size; conviction without an understanding of volatility justifies the opposite. A trade you are fairly sure about, with low variance in the outcome, is a large position. A trade you are fairly sure about, with enormous variance in the outcome, is a small position.
The commercial equivalent of Kelly is the same equation. Almost nobody runs it.
Commercial committees usually do something else. They size the bet to the ambition of the recommendation. The bigger the strategic rhetoric, the bigger the budget. If the deck says “market-leading position”, the budget goes up; if the deck says “defensive consolidation”, the budget goes down. The variance of the outcome is rarely a parameter at all.
What this looks like in the wild
Three patterns we see most often. Each one is a sized-to-ambition mistake. Each one would be caught by sizing-to-edge.
The hero launch. A new line or new market that is genuinely high-expected-value, but with a wide variance because the team has never operated in the geography or category before. The committee sizes the bet to the strategic ambition and ships £20m of capital into a launch with a 95% confidence interval spanning a factor of four. Half the time it’s a great trade. Half the time it’s a slow car crash. Either way, the bet is mis-sized: it should have been a £5m wedge with a kill criterion at month nine, not a £20m all-in.
The defended share. A core category that’s losing share to a structurally cheaper competitor. The expected value of the defence is genuinely negative — the share is going regardless — but the committee sizes the defensive budget to the size of the share that is leaving, rather than to the share that can actually be defended. Most defensive budgets are sized to ambition, not edge.
The pet partnership. A bilateral deal a senior leader is personally close to. The expected value may be fine, but the variance is high because the counterparty is small and the contract is loose. A trader would size this as a 25-basis-point position. The committee sizes it as a flagship initiative.
The fix
The fix is not to import the maths of Kelly into commercial committees. It is to import the discipline. Three small moves do most of the work.
First, before the budget conversation, force the room to name the variance of the outcome alongside the expected value. Not a confidence interval to a decimal place — just a sense of whether the variance is small, moderate or large. Half the bets in any given quarter are mis-sized because nobody asked.
Second, attach a kill criterion to every commercial bet over a threshold. A pre-committed trigger that, if hit, ends the bet. Trading desks live and die by stop-losses; commercial committees rarely commit to one, which is why most under-performing initiatives are still on the books two years after they should have been killed.
Third, separate the people who make the call from the people who get to revise it. A trader does not get to argue with their own stop-loss when it triggers. A commercial committee usually does. Don’t.
None of this is news on a trading floor. None of this is standard practice on a commercial committee. The gap between the two is where most of the value of this practice lives.