The two bars
Each horse gets an EARLY bar (how fast it runs to the first call) and a LATE bar (how strongly it finishes), shown as a percentile against the other horses in this race, with its rank in the field. A horse at the 94th early percentile is not "a 94 horse" in any absolute sense — it is faster early than almost everything it is facing today. Every number on the card is relative to the field, because races are.
Next to the bars sits the horse's running-style label: E (early — wants the lead), E/P (early/presser — can lead or sit just off it), P (presser/stalker), S (sustained — a closer). Keep the labels in mind; they matter more than most readers think, and less than the bars suggest — a point we will come back to with data.
A live read: Parx Race 2, October 20, 2026
This was the day's free race — 6 furlongs, dirt, claiming, nine horses. Here is the card exactly as it published in the morning:
Parx Racing — Race 2 · 6 Furlongs · Dirt · Claiming
The quick read says: We Miss Neil (#6) tops the early graph at 94 — the speed of the speed. The careless read stops there and calls him lone speed. He is not, and the card itself tells you why, if you run the checks:
- Top of the early graph? Yes — #1 of 9.
- A real gap to the second horse? Barely. Filled With Desire (#2) sits at 88, six points back, and she is a committed E style who will go with him.
- How many other early styles are in the field? Count them: three more E or E/P horses (Basin Bay, Dirty Cash, Duke of Suggins) plus the pressers. This is a crowd at the front, not an empty road.
- Who owns the late graph if the front end gets busy? Filled With Desire again — late #1 at 96 — with the closer Chilling Time (#5, late #2) waiting behind her. The horse most likely to benefit from a duel is the same horse most likely to cause one.
- What do the early-speed points say? Points — the count of how often and how strongly a horse has made the early lead in its recent races — are a second, independent read on who actually goes. When the points and the early graph disagree, believe that the race is less settled than the graph looks. (This check earned its place the hard way; it is the subject of the Keeneland race below.)
Lone speed is a strong claim. It means the top of the early graph, by a gap, with no other horse in the field whose job is also to lead — and a late graph that says nobody is sitting close enough to punish a soft trip anyway. Most races do not have one. The honest read of Parx Race 2 is "contested early, decided late" — which is a different race entirely.
When the obvious read falls apart: Keeneland Race 7, October 17, 2026
Three days earlier, the Perryville (Grade 3) at Keeneland — 7 furlongs on dirt — gave us the cleanest lesson in our files. Before the race, our pace sketch projected the early leader by blending early-pace figures with early-speed points into one score. The projection was not close:
| Horse | ML | Style | Projected early score | Early-speed points | Actual position at the first call |
|---|---|---|---|---|---|
| #11 Obliteration | 9/2 | E/P | 2.85 | 3 | 8th |
| #4 Roscoe's Brother | 15/1 | E | 0.70 | 6 | 1st — led |
| #5 American Man | 20/1 | E/P | 0.64 | 5 | 2nd |
| #8 I C Light | 30/1 | E | 0.32 | — | 4th |
| #12 Deep Flame | 7/5 | E/P | 0.12 | 5 | 5th |
| #7 Mo Koko | 30/1 | E/P | −0.14 | 3 | 3rd |
| #1 Solitude Dude | 6/1 | E/P | −0.40 | 5 | 6th |
| #6 Dr. Kapur | 12/1 | E/P | −0.71 | — | 9th |
| #3 Chief Wallabee | 8/1 | E/P | −0.71 | — | 7th |
| #9 Street Beast | 20/1 | E | −0.91 | 1 | 10th |
Starters only; two entered horses scratched. First-call positions and finish from the official Equibase chart.
Obliteration topped the projection by 2.15 points over the next horse — one of the widest projected gaps in the study we later ran: only two of 207 graded races projected a leader more strongly. On the morning card he also shared the top early rank in the field. It looked like his lead to lose. But the second look was sitting right there in the same table: the early-speed points — the simple count of demonstrated early speed — made Roscoe's Brother the field's most committed early horse, 6 points to Obliteration's 3, and two more committed E styles (#4 and #8) were in the field with him. Obliteration was never going to be alone; the only question was who would actually go.
The chart answered it. Roscoe's Brother led at the first call. Obliteration was 8th at the first call — he rallied to finish 3rd, paying $3.26 to show, but the lead the sketch handed him never existed. Our pick that day, Deep Flame (#12, the 7/5 favorite, who led both graphs on the published card), was 5th at the first call and finished out of the money. And the race went to the horse the early argument ignored completely: Dr. Kapur, 9th at the first call, won and paid $50.24.
This was not one bad beat we chose to remember. We graded the question properly — 207 settled races, four ways of projecting the first-call leader — and the simple early-speed points named the actual first-call leader in 41.2% of gradable races, against 32.4% for the blended score that produced the Obliteration projection. The sketch on our card now projects from the points alone. The full study, including the parts that came back against us, is the companion guide: What 207 Races Taught Us About Projecting the Lead.
The short version
- Read both bars: early says who can lead, late says who is still running when it matters.
- "Tops the early graph" is not lone speed. Lone speed needs a gap and an empty road.
- Count the other E and E/P styles before you believe any lead projection.
- Check the early-speed points against the figures. Disagreement is information.
- Always know, before the gate opens, who the race falls to if the lead never happens.