The public record, and why it is public

Why a tipping record only means something if it's complete

A record with the losers taken out isn't a record, it's an advert. Anyone can look sharp if they only show you the bets that came in. The only figure worth trusting is one built from every single selection that was made public and settled, win or lose, dated as it happened rather than tidied up afterwards.

That's the whole point of publishing everything. The Racing Bot's public record currently stands at 15,265 settled picks across 423 days, with 25% winners and 50% placed. Those numbers include every dud alongside every winner, because a record that only remembers the good days tells you nothing about how a method actually performs over time.

Horse racing is a game of variance. Favourites win roughly a third of the time, long shots come in occasionally, and even a sound method will have losing runs simply because that's how probability works. A record spanning hundreds of days and thousands of picks smooths out the noise of any one lucky or unlucky week and gives you something closer to the truth.

What to actually check before trusting anyone's numbers

Start with the sample size and the timeframe. A handful of good months proves very little in racing; too few picks and too short a run can be explained by chance alone. Ask how many selections sit behind the headline figure and how many days it covers, because both matters just as much as the win rate itself.

Next, check whether losers are visible at all. If a service only ever tweets or posts about the winners, there's no way to know the true strike rate, and no way to hold it accountable. A genuinely open record lets you scroll back and see the picks that didn't work out, dated and sitting alongside the ones that did.

Finally, look at how the result was priced. A winner quoted at odds that were never actually available when the bet was made is a winner on paper only. Fair reporting uses prices that could realistically have been taken at the time, whether that's an early price, a starting price, or something clearly stated and consistent across every single entry.

The traps to watch for

Cherry-picked windows are the oldest trick going. A tipster having a purple patch will happily show you "last month" or "this meeting" while quietly ignoring the quieter stretches either side of it. Always ask what the record looks like over the full run, not just the flattering slice.

Unrecorded losers are the quieter version of the same problem. Some services delete or simply never mention selections that lose, so the visible history drifts further and further from reality the longer it goes on. If you can't find a losing pick from three months ago as easily as a winning one from last week, that's worth noticing.

Then there's the price problem. Results quoted at odds nobody could have got, best-price boasts built from the one bookmaker who briefly offered a bigger number, or prices taken well after the tip was issued, all inflate figures without changing anything about the actual advice. Consistency in how prices are recorded matters more than any single big-priced result.

This is why The Racing Bot scores every runner across 15 measurable signals and settles every pick against a clear, stated price, then leaves the full history open to check. With 15,265 settled selections and 423 days behind it, the 25% winners and 50% placed figures are what they are, not filtered, not rounded up, and not a promise about what happens next.

Reading a record like a punter, not a marketer

A 25% strike rate sounds modest next to the bold claims some services make, but it needs context. Placed rate matters too, and a 50% placed figure over more than 15,000 picks shows a method that's finding a reasonable number of horses running well, even when they don't win outright.

No record, however honestly kept, can tell you what's going to happen in the next race. Racing carries genuine uncertainty, and past performance over 423 days is a description of history, not a guarantee of anything to come. What a full public record can do is show you how a method has actually behaved, warts and all, so you can judge it on real evidence rather than a highlight reel.

The simplest test for any tipster or tool is this: can you find their losers as easily as their winners, and does the whole run add up to a stated total you can check? If the answer is no, treat the claims with caution. If the answer is yes, you at least know you're looking at the real picture, not the polished one.