Pricing: Building Your Line
Turning ratings into fair odds
Here it is, the whole course in one line: the question is never "will this horse win?" — it is "is this horse correctly priced?"
Theory: how the market works
Before you can spot a wrong price, you need to know how prices are made.
Pari-mutuel (tote) pools every bet, removes the takeout, and divides the rest among the winning tickets. The odds you see are a live snapshot of where the money is — and they can drift as others bet, so your final payout isn't locked when you bet. You're competing against other bettors, not a house, and sharp players can't be banned.
Fixed-odds (bookmaker) locks your price when you bet. The book builds in a margin (the "overround") and — this matters — quietly limits or closes accounts that win.
Exchanges (Betfair, Betdaq) are peer-to-peer: you back or lay, paying commission only on net winnings. Prices converge to the sharpest pre-race numbers with the lowest effective cost, and exchanges don't restrict winners.
World Pool — the MWP user's structural advantage
World Pool is the HKJC's commingled international tote pool: money from many jurisdictions flows into one combined pool for major meetings — Royal Ascot, the Derby, Dubai World Cup day, the Breeders' Cup, the Hong Kong International Races and dozens more. It matters for four reasons: deep liquidity (a large bet barely moves the price), a global mixed-skill crowd (more soft money to exploit, especially in exotics), no winner restriction (a tote can't ban you for being good), and prices that on average beat fixed-odds books, particularly on longer-priced winners.
Takeout — the hurdle you must clear
Takeout isn't just a fee, it's a hurdle rate. World Pool standard rates: Win/Place 17.5%, Exacta 19.5%, Trifecta 22%, Pick 3 25%. The brutal arithmetic: a 20% takeout doesn't mean you must be 20% better than the field — you must be 25% better, because you only keep 80p of every pound (1 / 0.80 − 1 = 25%). This is why high-takeout pools are unwinnable for casual players and a goldmine for sharp ones.
Theory: value
A horse with a 60% chance of winning is a bad bet at 1/2 (implied 67%). A horse with a 10% chance is a great bet at 15/1 (implied 6.25%). The first is "likely to win" and a losing proposition; the second is "unlikely to win" and profitable over time. Picking winners and picking value are different skills. The second one pays.
The value equation
Expected Value = (your probability × decimal odds) − 1
- Probability 25%, odds 5.00 (4/1): EV = (0.25 × 5.00) − 1 = +0.25 → positive EV, good bet
- Probability 25%, odds 3.50 (5/2): EV = (0.25 × 3.50) − 1 = −0.125 → negative EV, bad bet
Same horse, same chance, different price. One is a professional bet. The other is a donation.
Building your odds line
- Handicap the race — apply everything from Chapters 1–11.
- Estimate each contender's win probability — these sum to ~100% (or ~110% for a margin of safety).
- Convert to fair odds — decimal odds = 1 / probability. 25% = 4.00, 10% = 10.00, 40% = 2.50.
- Compare to the market. Market 6/1, your line 4/1 → a 50% overlay.
- Bet only when the overlay clears your threshold — 20% minimum is a reasonable floor.
What you are betting on is the gap between your line and the market — horses the market is mispricing relative to your line — not the horse most likely to win.
The MWP process
Open the racecard → eliminate non-contenders → adjust expected ratings → run the Handicapper → read projected margins → convert margins to probabilities → write down your fair odds → wait for the market → bet only overlays. Ten to fifteen minutes per race once fluent.
[QUOTE: "Build your fair-odds line BEFORE you look at the tote. If you look first, you anchor to the market's opinion instead of forming your own."]
Practical: price the race
The first case opens with results hidden — set your line before you see what happened. The second is a free build: use the Pricing tab, the 🎲 Guesstimate baseline, and ⚖ Adjust to 100% to produce a full line.
Train the habit: build a line on four real races
Treat the pricing exercise as a habit, not a one-off. Switch between the four races below, set probabilities for every runner, and watch your line emerge. Each race is a different shape: a clear favourite to size correctly, an overlay hiding behind two weak runs, a pace play, a class drop.
More overlays in the wild
Three real spots where the market got the price wrong — an overpriced class drop, a Dubai turf overlay, and a quiet class-drop angle.
Reading our form going in (Lady's Choice): build your line first and she tops it — a real, recent 143.2 peak, 8+ clear of the field — yet she's 26.0 because her last two runs look awful. The market anchored on the fresh duds and binned the genuine peak. She's inconsistent, so fair odds aren't 3/1; but 26/1 on the field's best recent figure is far too big.
Reading our form going in (Soul Rush): rank on figure and he's the clear second-best (154.4), behind only the 1.3 superstar — yet priced as fifth choice at 10.0. When one star dominates the market, the genuine number two gets underbet. Rank on form, not on names.
Reading our form going in (Tourbillon Prince): his last lines read 6th-13th-4th, so he's 11.9 — but those were all in Class 2, above his level. The figure that matters is the 141.2 he earned winning in Class 3, his correct grade, which makes him the form top here. Read the class of the recent runs before trusting the positions.
The shape of the market: where the value lives
Before you hunt for a wrong price, learn the one bias that runs through every market. We measured it across the full MWP database by starting-price band — both what a flat bet actually returns, and how each band does against the crowd's margin-free estimate (the SP POT measure from Chapter 5):
| Starting price | Win rate | Flat-bet return | vs the margin-free market |
|---|---|---|---|
| Odds-on (≤2.0) | 52% | −12% | +7.6% |
| 2.0–3.5 | 31% | −13% | +5.7% |
| 3.5–6 | 18% | −17% | +1.3% |
| 6–11 | 10% | −20% | −2.5% |
| 11–21 | 5% | −23% | −5.7% |
| 21+ | 1.5% | −41% | −21% |
Read both columns, because they say different things. The flat-bet return is negative everywhere — that is the takeout, and it is why betting blind is a losing game at any price. But the last column strips the bookmaker's margin out and asks the sharper question: relative to the crowd's true estimate, who is overbet and who is underbet? There the favourite-longshot bias appears — one of the oldest and most durable facts in betting. The crowd overbets longshots (the lottery-ticket dream) and underbets the short end: a 21/1 shot is overbet by about 21% against its real chance, an odds-on favourite slightly underbet.
Two things follow. First, "take the longshot for the value" is backwards — the longshot is where the crowd is most wrong against you, not for you. Second, this is why the MWP rating's top pick is a real starting edge: the highest-rated horse in each field wins about 20% and beats the crowd's margin-free estimate by roughly +6% (+8% at a short price). It still loses to a flat bet — you pay the takeout like everyone — but it gives back most of that takeout before you have applied a single judgement. Your job in this chapter is to find the spots where your own estimate clears the margin: bet only the overlays, into the deepest, lowest-margin pool you can reach.
Closing Line Value
The closing line — the final market price — is the best public estimate of each horse's chance. If you consistently bet at prices higher than the closing line, you are by definition pricing more accurately than the market. CLV is the single most reliable measure of whether you have an edge — it cuts through variance in dozens of bets, where win rate and P&L stay noisy for hundreds. Track it on every bet: average CLV above zero over 100+ bets means the results will come; negative CLV means something in your process is wrong.
Quiz
Up next: betting itself — stakes, pools, and turning a priced line into placed bets.