Betting analytics glossary
What is Return on Capital (ROC)?
ROC treats betting like any other capital investment: how much did the money you actually deployed earn you. Harder to compute than yield, more honest about whether the operation is worth your time.
The short definition
Return on Capital (ROC) is total net profit divided by the bankroll you deployed, expressed as a percentage. If you set aside 100u to bet with and you finish the season up 27u, your ROC is 27%. Simple as that.
This is the metric a finance-brained observer would actually care about. Not “how many units did each stake earn on average” (that’s yield) but “what was my percentage return on the pot I put at risk”. The number you could compare against an index fund or a savings account without embarrassing yourself.
The formula
ROC % = (Profit ÷ Bankroll Deployed) × 100Profit is net (returns minus stakes). Bankroll deployed is the tricky bit. Three defensible ways to define it, in rising order of honesty:
- Starting bankroll: whatever pot you opened the tracking period with. Easy but flatters you if the bankroll grew mid-period.
- Peak bankroll:the highest running total during the period. This is the “risk capital” interpretation and is what most serious trackers use. It penalises leverage on a hot streak, which is the correct incentive.
- Time-weighted average bankroll: the textbook finance definition. Most accurate for comparing against traditional investments, but rarely worth the calculation effort for a hobbyist.
Am I Upuses the peak-drawdown-adjusted risk capital by default (the smallest bankroll you’d have needed to survive your worst losing stretch) which is close to the peak-bankroll definition without rewarding lucky variance.
Worked example
Same bettor as the ROI page. 100u bankroll. 300 bets at 1.5u average stake, so total stake = 450u. Net profit = 27u.
- Peak bankroll: 100u (assume no top-ups)
- ROC = 27 ÷ 100 = 27%
Compare against yield, which comes out at 6% on the same season (27u profit ÷ 450u total stake). Same person, same trades, two very different headline numbers. The ROC-flavoured 27% is the number the bettor can defensibly compare against “what if I’d bought the S&P 500 that year”. The yield-flavoured 6% is the number a tipster shows off with.
Why serious bettors track ROC
- Yield can hide bad staking. A cautious punter with 10u average stakes on a 10u bankroll is running dangerously overleveraged even if their yield is +5%. ROC surfaces this immediately by dividing profit against the actual bankroll they were risking, not the accumulated stake volume.
- ROC lets you compare against other asset classes.An index fund returned 10% this year. Was your betting operation actually worth the time it took? ROC is the number that answers that. Yield can’t.
- ROC exposes churn. A bettor who turns their bankroll over 20 times a year with a 3% yield is generating a 60% ROC before variance. A bettor with the same yield turning over once a year is generating 3%. Same yield, wildly different capital efficiency. ROC catches this.
Where ROC gets misleading
- It rewards under-stated bankrolls.A tipster who reports a bankroll of 20u while regularly staking 5u per bet is claiming an ROC based on a bankroll they’d bust every second bad month. Always check whether the claimed bankroll would actually have survived the drawdowns.
- Bankroll size is squishy.Deposits, withdrawals, promotional credit, and interest all shift the number. Some tools smooth this by using a risk-capital-adjusted denominator (the smallest bankroll you’d have needed to weather your worst stretch), which is more honest but harder to explain in one line.
- Small samples still lie. ROC is a historical measure just like yield. A 40% ROC over 50 bets tells you very little about whether the bettor has an edge. Look at CLV for the sharper signal.
ROC benchmarks (1000+ bet samples)
Very approximate. Depends on staking style, sport, and turnover rate. Meant as a sanity check, not a target.
- Negative: the honest answer for the majority of bettors, including many who report positive yields on small samples.
- 0 — 10%: in the ballpark of what a savings account or index tracker returns. Worth asking whether the time investment is paying off.
- 10 — 30%: genuinely outperforming most passive alternatives. Where a skilled hobbyist with real edge lands.
- 30%+: professional territory. Also where you should start seriously questioning survival bias, stake-cap dodges, and whether the sample is big enough.