Clever Punts Explains Stall Draw Bias

The Racing Stall Draw Analyser uses past race results to show whether horses starting from low, middle or high stalls have performed better or worse than expected at different courses and distances. It compares actual results with betting-market expectations and highlights possible draw advantages, while helping you distinguish reliable patterns from results caused by small samples or occasional big-priced winners.

Finding Draw Bias with the Racing Stall Draw Analyser

How to read the heat map, compare gap-adjusted bands, and understand Diff, A/E and SP P/L

The Racing Stall Draw Analyser is designed to answer a familiar racing question: does being drawn low, middle or high make a measurable difference?

Rather than judging a course from a handful of winners, the tool brings together historic flat-race results, groups runners by stall and distance, and compares what actually happened with what the betting market expected to happen. Its heat map gives a quick visual summary, while the profile and Stalls views let us investigate the pattern in more detail.

Reading the low, middle and high heat map

The heat map divides the draw into three bands:

• Low — the lowest third of the draw
• Middle — the middle third
• High — the highest third

The number in each bubble is controlled by the Profile metric selector. Green indicates a result above the metric’s neutral point and red indicates one below it. A stronger colour means the result is farther from neutral.

Bubble size can be fixed, which makes the values easy to compare, or based on the number of runners. When By runners is selected, larger bubbles represent larger samples. This is useful because an exciting result from 20 runners should not normally carry the same weight as a similar result from 2,000 runners.

The distance headings and the Low, Middle and High headings can also be used to sort the heat map. That makes it easier to spot the distances at which a particular part of the draw has performed best or worst.

Gap-adjusted versus Stall / runners

The Band method determines how a runner is assigned to the Low, Middle or High group.

Stall / runners

This method divides the runner’s printed stall number by the number of runners with a numeric draw:

The resulting percentage is placed in the bottom, middle or top third.

For example, imagine a six-runner race in which the occupied stalls are 1, 2, 4, 5, 7 and 8. Stall 4 is calculated as 4 ÷ 6 = 66.7%, so it sits at the top of the Middle band. Stall 8 is 8 ÷ 6 = 133.3% and is classified High.

This method preserves the published stall numbers. However, it can be distorted when non-runners leave gaps. In the example, six horses occupy a draw that stretches to stall 8.

Gap-adjusted

The gap-adjusted method first puts the occupied stalls in order and gives them ranks from 1 to the number of runners:

The same occupied stalls—1, 2, 4, 5, 7 and 8—therefore receive ranks 1, 2, 3, 4, 5 and 6. Stall 4 is the third occupied stall, so its adjusted position is 3 ÷ 6 = 50%, placing it in the Middle band. Stall 8 is ranked sixth, or 100%, and remains High.

Gap adjustment answers the question, “Where was this horse positioned relative to the runners that actually took part?”. It prevents empty stalls from making the field look wider or shifting too many runners towards the High band.

Neither method changes the individual stall statistics in the Stalls tab. It only changes how runners are allocated to the three broad draw bands.

Understanding the Profile metric

The selected Profile metric drives both the heat-map bubbles and the Draw Bias Profile chart.

Diff

Diff is the actual win rate minus the average starting-price chance:

Suppose a draw band supplied 100 runners and 15 winners. Its win rate is 15%. If those runners had an average SP-implied chance of 11%, the Diff is:

A positive Diff means the group won more often than its average market price suggested. A negative Diff means it won less often.

Diff should be read as a percentage-point difference, not as a percentage improvement. A result of +4% on the chart means four percentage points above the SP expectation.

A/E

A/E means Actual over Expected:

The expected total is built by converting each valid decimal SP into an implied chance (1 ÷ decimal SP) and adding those chances together.

• 00 is neutral: actual winners matched the market expectation.
• Above 1.00 means more winners than the prices implied.
• Below 1.00 means fewer winners than the prices implied.

If a band produced 12 winners when its runners’ prices implied 10 winners, its A/E would be 12 ÷ 10 = 1.20.

Diff and A/E examine similar territory but from different angles. Diff expresses the gap in percentage points; A/E expresses actual wins as a ratio of expected wins. A/E can look volatile when the expected number of winners is small, so it should always be read alongside Runners and Wins.

SP P/L

SP P/L is the level-stakes profit or loss from backing every runner in the group for one unit at its decimal starting price.

• A loser costs one unit.
• A winner returns its decimal SP, giving a profit of decimal SP − 1.

For example, ten one-unit bets cost ten units. If the only winner is returned at decimal odds of 12.0, the overall result is a profit of two units.

A positive SP P/L is attractive, but it is also the metric most easily moved by one big-priced winner. It should not be considered in isolation. Check the number of runners, the number of winners, A/E and whether the pattern repeats across distances.

The Draw Bias Profile chart

The profile chart plots Low, Middle and High as separate lines across distance. It uses the same selected metric as the heat map.

Its neutral reference line changes with the metric:

• Diff is neutral at 0
• A/E is neutral at 00
• SP P/L is neutral at 0

This chart is useful for distinguishing a broad course tendency from a distance-specific effect. A low-draw advantage may be strong over five furlongs, disappear over six, and reverse over longer trips. A line that remains on the favourable side of neutral across several distances is more persuasive than a single isolated spike.

Distances can be switched on and off. The selection applies to the heat map, profile chart and combined total, allowing like-for-like comparisons.

Reading the Stalls chart

The Stalls tab moves from broad thirds of the draw to each individual stall.

The default win-rate chart shows:

• the stall number;
• its win rate;
• the number of runners recorded from that stall; and
• a bar representing the win rate.

The Max stall control can restrict the chart—for example, to stalls 1 to 10—so that very high, sparsely used stall numbers do not dominate the display.

The bars show raw win rate. They do not adjust for the prices of the horses or the sizes of their fields. Stall 1 might have a higher win rate simply because it contained more well-fancied runners. For that reason, the chart is best used to locate a pattern before checking the data table for market-adjusted context.

A practical approach

Start with the gap-adjusted heat map and Diff. Look for a coherent Low, Middle or High pattern rather than the single brightest bubble. Switch the metric to A/E to see whether the same part of the draw also beat market expectation. Then check SP P/L, while remembering that profit can be driven by outliers.

Next, use the profile chart to see whether the effect persists across distances. Compare the result with Stall / runners: a major change warns that missing stalls are shaping the result. Finally, open the Stalls chart and table to see whether the broad band effect is spread across neighbouring stalls or depends on one exceptional stall.

The most credible draw-bias case is not simply the highest win rate. It is a pattern supported by a healthy number of runners, repeated across relevant samples, stronger than both field-size and market expectations, and not dependent on one large-priced winner.

That is where the analyser is most useful: it turns “low numbers seem favoured” into a set of claims that can be examined, compared and challenged.