Ask a regular horse racing bettor what last season actually returned, and you will rarely get a number. Not out of carelessness, but because horse racing resists tracking. A spreadsheet that works perfectly well for football bets falls apart on the first Quinté+ played as a reduced field, and the history quietly goes wrong.
The problem is not the bettor's discipline. It is structural, and it comes down to three things.
Why an ordinary bet tracker fails on horse racing
The stake is not the one you think
A horse racing bet is almost never a single line. A Quinté+ reduced field with 2 bankers and 5 associates is not a €2 stake, it is C(5,3) = 10 combinations, so €20. A Multi with 6 horses is 15 combinations. If you record the unit stake in your spreadsheet, your ROI is mechanically inflated, sometimes tenfold.
The odds do not exist when you place the bet
That is the nature of pari-mutuel betting: the odds shown before the off are a probable dividend. They move with the pool right up to the start and can drift 30 to 40 % from the final figure. The only number that matters is the final dividend, published after the result. Recording the odds you saw when placing the bet means storing a value that will be wrong minutes later.
One ticket can win several times
On a field bet, several combinations can win at once. The return is not "odds × stake", it is "dividend × unit stake × number of winning lines, minus total stake". No won/lost column can model that.
These three biases do not cancel each other out, they stack. A hand-kept horse racing history almost always overstates real profitability.
Logging a turf bet from the official programme
The starting point changes everything. In BettingTracker, meetings and races are imported from the official PMU programme, runners included. You retype nothing: no racecourse, no race number, no horse names. You open the calendar, pick your meeting and your race, and tick your horses.
The main bet types are covered, from the simplest to the most combinatorial:
- Simple win and place
- Couplé win, place, exacta and win-place
- Trio, Trio ordre, Tiercé, Quarté+, Quinté+
- 2 sur 4, Multi, Mini Multi and Super 4
- Fixed-odds products: Top 4, Top 4 ordre, Top 5, Super Top 5
- Ze Show, Ze Couillon, Géant, Pick 5, Pick 6
Betting formulas are modelled, not approximated
This is the heart of the matter. Every bet accepts the formula you actually played:
- Reduced field (champ réduit): your bankers against your selected associates
- Full field (champ total): your bankers against every remaining runner
- Box (combiné): every combination of your selection
- Flexi 100 %, 50 % or 25 %: the unit stake is fractioned, and so are the returns
- On ordered bets: the all orders option and the field position selector
The combination count and the total stake are computed as you build the ticket. You see what your bet really costs before confirming, and your history records the correct stake.
Settlement runs on official dividends
Once the race is run, the bet settles without you. The system pulls the final dividend from the operator, PMU as well as Genybet, identifies the winning combinations on your ticket, and computes the exact return, multiple winning lines included.
Two practical consequences:
- Your bet's odds are corrected automatically. The probable dividend you were looking at is replaced by the real one. Your average-odds statistics finally mean something.
- Partial wins are handled. A field bet hitting three combinations out of ten is settled as such, not flattened into a binary won or lost.
Multi-race bets such as Pick 5, Pick 6 or Géant are not settled automatically: they depend on several races and on operator-specific rules. They stay on manual settlement rather than being resolved from uncertain data.
The statistics horse racing deserves
This is where tracking stops being bookkeeping and becomes a decision tool. Because every bet is linked to the real runners, your history slices along dimensions specific to racing:
- By discipline: harness, mounted trot, flat, and jumps broken down into hurdles, steeplechase and cross-country
- By field size: are your small 8-runner fields worth your big 18-runner ones?
- By distance, combinable with discipline
- By jockey, trainer, owner, breeder and horse
- By betting formula: is your full field profitable, or merely reassuring?
- By bet type: does your Quinté+ fund your Couplés, or the other way round?
Every entity gets its own page: how many races you backed it in, its win and place rates, its average finishing position, and above all what the tickets containing it actually returned. The two are not the same: a horse can sit in winning tickets without ever having carried the win.
The ranking cuts both ways. It does not just surface your best picks, it isolates your leaks: the trainer you have been backing out of habit for two years who costs you money, the formula that inflates your stakes without improving your strike rate. That is usually the most uncomfortable part, and the most profitable.
What it changes day to day
In practice, a race day looks like this: you build your tickets from the programme, you follow the results, and by the evening your figures are already correct. No dividends to key in, no returns to work out on a calculator, no doubt about the real stake of a field played at flexi 50 %.
After a few weeks you have something very few horse racing bettors own: a complete history, with the right stakes, the right dividends and the right returns. It is the only honest starting point for knowing whether you win, and more importantly where.
Getting started
Create a bankroll dedicated to horse racing, kept separate from your sports betting if you track both. Log your bets from the calendar rather than after the fact, so automatic settlement can do its work. Let it run three or four weeks before drawing conclusions: jockey and trainer statistics need a minimum of volume before they say anything.
Then open your rankings and look at the leaks first. It is rarely pleasant, but that is where the most immediate room for improvement sits.