Windsor Staking Plans: Flat Stake, Percentage and Level

Three Staking Methods, One Evening Card: Which One Fits?

A staking plan is the system you use to decide how much to bet on each selection. It is not the same as a bankroll — your bankroll is the total amount available; the staking plan determines how that total is allocated across individual bets. The difference matters, because two punters with identical selection skills and identical bankrolls will produce different returns over a season if one uses a disciplined staking method and the other bets by feel.

Windsor’s six-race evening cards make staking decisions both simpler and more consequential than at bigger meetings. Simpler, because the number of betting opportunities per session is small and manageable. More consequential, because each bet represents a larger fraction of the evening’s action. The BHA’s 2025 Racing Report recorded a 4.3% decline in overall betting turnover on British racing — a reminder that the market environment rewards discipline over volume. This guide covers three practical staking methods, each illustrated with a worked example across a typical Windsor card.

Flat Staking: Fixed Units, Simple Maths

Flat staking means betting the same amount on every selection, regardless of your confidence level or the odds. If your unit is £10, every bet is £10 — whether the horse is 2/1 or 12/1, whether you rate it as a strong selection or a marginal one.

Worked example for a Windsor evening. Bankroll: £300. Session budget: £50 (roughly 17% of bankroll). Planned bets: four selections from six races. Unit stake: £12.50 (£50 ÷ 4). Results: Race 1, loser (−£12.50). Race 3, winner at 7/2 (profit +£43.75). Race 4, loser (−£12.50). Race 6, winner at 3/1 (profit +£37.50). Session result: +£56.25 on a £50 outlay. The two winners more than covered the two losers because the prices were long enough to compensate.

The strength of flat staking is its simplicity. There are no calculations between races, no adjustments based on confidence, and no temptation to increase stakes after a losing run. You set the unit before the first race and you keep it through to the last. For punters who find their judgement deteriorates when they start adjusting stakes — chasing losses with bigger bets or going small after a winner — flat staking removes the decision entirely.

The weakness is that it treats all bets equally. A selection you rate at 40% confidence receives the same stake as one you rate at 70%. Over a long sample, this means you are under-staking your best bets and over-staking your weakest ones. Flat staking is optimal for punters who struggle with discipline; it is suboptimal for punters who can accurately assess their own confidence. For a punter attending Windsor’s 22 flat meetings per season and betting on three or four races per card, flat staking produces around 70-90 data points per year — enough to identify whether your selection method is profitable, but not enough to optimise stake size per bet with statistical confidence.

Percentage Staking: Your Bankroll Adjusts Automatically

Percentage staking sets each bet as a fixed percentage of your current bankroll — not your starting bankroll, but the live figure that rises with winners and falls with losers. The most common percentage is 2-5% of bankroll per bet.

Worked example. Starting bankroll: £300. Percentage: 3%. First bet: £9.00 (3% of £300). Result: loser. Bankroll: £291. Second bet: £8.73 (3% of £291). Result: winner at 4/1. Profit: £34.92. Bankroll: £325.92. Third bet: £9.78 (3% of £325.92). Result: loser. Bankroll: £316.14. Fourth bet: £9.48 (3% of £316.14). Result: winner at 5/2. Profit: £23.70. Bankroll: £339.84.

The key feature of percentage staking is automatic bankroll protection. As your bankroll shrinks, so does each individual bet — you cannot lose your entire bankroll because each subsequent stake is a fraction of a declining number. Conversely, as your bankroll grows, your stakes increase, allowing profitable periods to compound. Over a season of Windsor meetings, this self-adjusting mechanism smooths out the variance that flat staking does not address.

The downside is complexity. Calculating 3% of £316.14 between races at Windsor requires a phone calculator and a few seconds of maths. It is not difficult, but it adds a step that flat staking does not require. Some punters round to the nearest pound to simplify; others automate the calculation on a spreadsheet updated after each bet. Either approach works, provided the percentage is applied consistently.

Level Stake Profit and the Kelly Approach

Level stake profit (LSP) — the metric used throughout Windsor’s trainer and jockey statistics — is calculated by placing one unit on every qualifying selection and measuring the net return. It is not, strictly speaking, a staking plan in itself; it is a measurement tool. But it provides the data that fuels the most sophisticated staking method: the Kelly criterion.

The Kelly criterion calculates the optimal stake based on your perceived edge and the odds. The formula is: stake = (bp − q) / b, where b is the decimal odds minus 1 (the net payout), p is your assessed probability of winning, and q is the probability of losing (1 − p). If you assess a horse’s win probability at 25% and the odds are 5/1 (b = 5), Kelly recommends: (5 × 0.25 − 0.75) / 5 = (1.25 − 0.75) / 5 = 0.10 — meaning 10% of your bankroll.

The data from OLBG’s five-year Windsor stats illustrates why this matters. William Buick’s LSP of +15.57 from 32 wins tells you that backing Buick at Windsor at level stakes would have returned a profit. The Kelly criterion would go further: it would calculate the optimal stake for each individual Buick ride based on the available odds and the implied edge from his historical strike rate. In theory, Kelly maximises long-term growth. In practice, full Kelly staking produces volatile swings, which is why most practitioners use fractional Kelly — typically half or quarter Kelly — to reduce variance at the cost of slightly slower growth.

For most Windsor punters, full Kelly is impractical because it requires accurate probability estimates for every selection — a skill that takes years to develop. Fractional Kelly at 25% (quarter Kelly) is more forgiving of estimation errors and still outperforms flat staking over a large enough sample. If you have the discipline to estimate probabilities honestly and the patience to apply the formula consistently, quarter Kelly is the most mathematically efficient staking method available for a Windsor evening card.

Pick a System, Stick to It for 20 Races, Then Review

The worst staking plan is the one you change every week. Flat staking, percentage staking, and Kelly all have strengths and weaknesses — but all of them outperform no system at all. Pick the method that matches your temperament: flat staking if you value simplicity, percentage if you want automatic bankroll protection, Kelly if you trust your probability estimates. Then apply it for at least 20 bets — roughly four or five Windsor meetings — before evaluating results. Twenty bets is the minimum sample to separate the system’s performance from short-term variance. Trust the numbers, review after 20, and adjust only if the data warrants it.