The field-size problem. A raw draw number can't be compared across races: gate 8 is the widest in an 8-runner field but mid-pack in a 16-runner field. So we never judge draws by their number alone. Instead, for every historical race we rank the runners by their draw and split the field into three equal groups by relative position — the inside, middle and outside thirds. A "low" draw therefore means the same thing — bottom third of that field — whether 8 or 18 went to post. This is what makes results from different field sizes directly comparable.

Impact Value (IV). Within each third we compare how often it actually produced winners with how often it would if the draw made no difference at all — i.e. if winners were shared out purely in proportion to how many runners each third held (a third of the runners would take a third of the wins):
IV = (that third's share of all winners) ÷ (its share of all runners). An IV of 1.00 means that part of the draw won exactly its fair share — no bias. 1.30 means it won 30% more often than expected (a real edge); 0.80 means 20% fewer wins than expected (a disadvantage). The same calculation on top-3 finishes gives the Place IV.

Bias strength & the chart. The leaderboard ranks trips by bias strength — the gap between the best and worst third's win IV — so the most lopsided draws rise to the top. The bar chart breaks the same data down to each individual draw number; treat the highest draws cautiously, as fewer big fields means smaller samples there.

Worked example — Greyville 1000m Inside draws there win 13.9% of their starts — that is, of every runner drawn in the inside third, 13.9% went on to win. Divide that by the Win IV of 1.37 and you get ≈ 10.1%, the baseline a neutral draw would manage (and a hint that the average field is about ten runners, since roughly one winner emerges per ten starters). So inside draws win 37% more often than that baseline, while outside draws (7.5%, IV 0.74) win 26% fewer. The place line reads the same way: 35.9% of inside starters finish in the top three, and a Place IV of 1.18 means 18% more top-three finishes than a fair share.

Two things to keep in mind. First, the three win percentages are independent rates, not slices of one pie — each is "of the runners in this third, how many won" — so they won't add up to 100%; they sum to roughly three times the overall win rate. (What does add to 100% is the share of winners each third supplies, since every race has exactly one winner.) Second, the three groups don't each hold exactly a third of the runners when a field doesn't divide evenly by three (an 11-runner field splits roughly 4 / 4 / 3 by draw rank). Impact Value handles this — it divides by each third's actual share of runners, not an assumed 33% — which is why IV, rather than the raw win %, is the figure to trust when comparing the three groups.

Built from historical results from August 2020, counting every runner that took its place in the gates. Only track-and-distance-and-going combinations with enough races are shown, so a single freak meeting can't skew the picture. Polytrack races are identified by Standard going; all other goings are turf.