DC Market Intelligence Report –Sep 12, 2026

01 The Market Verdict

Yesterday’s strong-steamer cohort was a net underperformer versus its own closing-price expectation. The group returned 9 winners from 83 qualifiers (10.8% win rate) against 10.54 expected winners, producing a performance index of 85.4. That matters because it tells us the late, decisive money (15%+ contraction from 10am to final) was not, as a set, accurately predicting the final outcome on the day; the market “knew” these horses were more likely than they turned out to be.

The profit picture reinforces the same point from two angles. At 10am the theoretical level-stakes return was only slightly negative (-3.62 pts), but by SP it was heavily negative (-19.92 pts). In other words: the day’s late compression did not translate into realised edge; if anything, the crowd paid up for the story as the day went on. The average move also hints at where the pressure arrived: Open→10am was only a mild average contraction (-1.80pp), while 10am→final was a material average shortening (3.40pp), consistent with most of the information arriving late.

Crucially, this was not an isolated oddity when placed against the broader baseline. Over the last 30 days, strong steamers show a similar underperformance profile: 306 wins from 2166 (14.1%) versus 349.89 expected, for a performance index of 87.5. The longer sample also carries a stark split in theoretical P/L: large positive at 10am (509.67 pts) but deeply negative at SP (-337.51 pts). That combination is a warning flag about how sensitive any apparent “value” is to price capture and how punitive late over-shortening can be.

02 Market Beaters vs Money That Misfired

03 Where the Money Went

04 The Practical Read

1. Anchor on expectation, not anecdotes. Yesterday’s strong steamers produced 9 winners versus 10.54 expected (PI 85.4), and the 30-day PI (87.5) says the same thing: closing price has not been a reliable “edge” measure lately, even when the move is strong.
2. Separate “identified contender” from “mispriced winner”. The place count (23 places; 27.7%) shows many steamers ran credibly. Treat close losers and placed horses as evidence the money can still locate competitiveness, but also evidence that the price often ends up too short for the true win chance.
3. Be forensic about where the move happens. The day’s big action was 10am→final (average 3.40pp) rather than Open→10am (-1.80pp). Late-only moves are more vulnerable to being crowded trades; early-plus-late agreement (as seen with some winners like Huscal) is structurally more reassuring than a single late lurch.
4. Watch the relationship between “final” and SP. Cases where SP is longer than final (for example She’s Too Kool) imply the peak price pressure occurred before the off; that can matter if you’re benchmarking whether the crowd kept buying right into the start or whether enthusiasm cooled late.
5. Respect sample size and avoid reading too much into repeated names from one day. Trainer/jockey clusters tell you where money concentrated yesterday, not that a connection has become a systematic “steam signal” on its own.
6. No filter-supplied qualifiers were available at generation time today. That increases the importance of confirmation in the Live Market: require both meaningful percentage contraction and a coherent price narrative (not just a flash move) before treating any runner as informed money.