Why a Tipster Needs Proofing and a Casino Game Doesn’t

The best part of a tipster advert is the small print. The one I keep going back to is a press advert in the Racing Post for a service called Isiris Racing, the one the Advertising Standards Authority ruled on in June 2015. The headline said £89,270 profit in five months, with every bet proofed to the Racing Post before each race. The small print said total turnover was £527,250, that the betting bank needed to acquire that profit was “just £500”, and that the average bet was £249 a horse. Put those numbers next to each other and you have a service staking half its bank on every selection, more than two thousand times over, and, on its own account, never going bust.

A member of the Starting Price Regulatory Commission complained, in a personal capacity, and the ASA upheld the complaint. The claim being wrong is not the point. What the ASA found when it asked for the proofing is, because it explains why proofing exists, and it explains why nobody has ever bothered to proof a slot.

What a Timestamp Is For

A tip is a forecast. Until the race is run it is worth nothing, and after the race it is worth exactly what the result says, and the gap between those two moments is where every bad tipster lives. A record can be tidied once the results are in. A losing bet gets remembered at a smaller stake. Each-way becomes a win single. A 7/2 that was 5/2 by the off stays 7/2 on the spreadsheet. None of that is possible if somebody independent received the selection, the stake, and the price before the race. That is all proofing is. Someone who does not care whether you win had the bet before the off.

Isiris had, on paper, the best proofing in the business. When the ASA sampled the bets at random, there was no proofing documentation at all for two of them, each-way terms recorded in the spreadsheet were not visible in the screenshots, and some of the odds in the spreadsheet did not match the odds in the proofing. The spreadsheet listed 2,122 bets and £533,400 staked, which is not the £527,250 in the advert. The last finding was the one that mattered. In 95 per cent of the sampled bets the starting price was worse than the price on the spreadsheet, and at SP the same bets would have produced 55 per cent of the claimed profit. The ad was found to breach the rules on misleading advertising, substantiation, qualification and exaggeration, and was told not to appear again in that form.

The way I run a proofing period here, three months at a minimum, real money, every stake and price recorded against what was actually available, is written up on the How We Review Tipsters page, so I will not go through it twice. Proofing exists for one reason, which is that the seller can rewrite the record afterwards if nobody else holds a copy.

Why Nobody Proofs a Slot

A casino game has no judgment in it. Its return is fixed in the design before anyone plays it, and in Britain it cannot be put in front of the public until it has been through an approved test house. The Gambling Commission’s testing strategy puts it in one sentence: all new games and random number generators can only be released once the testing has been completed and the report provided to the Commission. The Commission keeps the list of test houses allowed to do that work. The technical standards then say the return to player percentage and the house edge must be easily available to the customer before they gamble. A tipster will never tell you his long-run edge. A slot has to, it is negative, and a test house checked it before the game existed as far as you are concerned.

Mega casino is casino only, with no sportsbook attached, which is why I picked it for the example. There is nothing on the site anyone could forecast. Every game on it was tested before it went live, so a write-up of the site can only describe what is there, and the same goes for any licensed casino-only site in the country. A tipster’s record has to be collected after the fact, bet by bet, by someone who was not paid to collect it. A casino game’s record was written before the first spin and nobody can edit it afterwards, the operator included.

Two Kinds of Sample

Both are samples, and that is where the trouble starts. A tipster’s record is a sample of a person, and samples of people are slow. Fifty bets tells you almost nothing about a tipster, however good the fifty look, because a sound service backing at 6/1 can run cold for a month without anyone having done anything wrong, and a hopeless one can land four in a row. I have written before about why several hundred proofed bets is the floor, and nothing in the Isiris file argues against that. The spreadsheet had 2,122 bets on it. The sample size was never the problem there; the prices were.

That is the other half of what makes a tipster’s sample usable, and it gets less attention than the count. A record proofed at prices that had gone by the time anyone could get on is a sample of something, but not of the service you would be paying for. The ASA’s remedy in the Isiris case said as much: future profit claims had to rest on odds most customers could actually get. Two thousand bets at prices that existed for ninety seconds on one firm’s board is a smaller sample than two hundred at the price on the exchange at the off, and the second one is the only one a subscriber will ever see.

A casino session is also a sample, but of a return that was printed before you sat down. A good week on a slot changes nothing about that return and tells you nothing about the next week. The person who proofs a tipster for fifty bets and calls it evidence and the person who has a hot fortnight on a slot and calls it form are making the same mistake from opposite ends. One of them has not collected enough. The other has collected something that cannot tell him anything, and has usually paid for the privilege.

Two Columns

What follows from that is a filing rule, and it is the only advice in this piece. Casino spend goes in the entertainment column, where the expected cost was known before you paid it, the same column as the cinema and the away day. The betting record is the only ledger you keep where an edge could ever show up, and the moment it has casino sessions in it, it cannot be read. A profitable month with a tipster and a bad week on a slot net off to nothing on the bank statement and to two entirely different facts in the two ledgers, and only the second version tells you whether to keep paying the subscription.

The sheet I keep for tipsters has the service, the date, the selection, the advised price, the price I actually got and the result, and nothing else goes on it. A night on the slots is not a bet in that sense, because there was no price to get and no judgment to test, so it has no business on that sheet. It goes on the other one, with the takeaway and the taxi. That sounds fussy until the first time you try to work out whether a service made money over six months and find a run of casino evenings sitting in the middle of the column, each one a loss with no selection next to it.

It cuts the other way as well. If a tipster’s proofed profit turns out to have casino results in it, or if a service sells a sports record and a casino method off the same page, that is the flag itself. A service with a real edge does not need to pad it with something that has none.

It is two columns instead of one, and most people run one. That is a large part of why most people cannot tell you whether their tipster is any good, and why the people selling the tips would rather you kept it that way.

Back to the small print. A £500 bank that turned over £527,250 in five months at £249 a go is a forecast presented as a record, and the presentation was good enough to get into the Racing Post. The ASA’s remedy was that the ad must not appear again in its current form, and that any future profit claim had to rest on odds most customers could actually get. The proofing, when somebody finally asked for it, was missing for two bets, wrong on the terms for others, and at the prices available on the day worth a little over half the profit claimed. Eleven years on, the adverts have moved to Telegram and the small print has got shorter, and the question has not moved at all. If a service will not send its selections to someone independent before the off, there is no record, whatever the screenshot says. If it will, the record is worth reading, and the first thing to read is the price.

Don't Follow a Tipster Until You Check This

Know which tipsters are actually performing — before you pay to follow them. Every month, I track the latest ROI, profit, verified results and ranking changes across the tipsters we monitor.

You'll see:
📈 Who's moving up
📉 Who's falling down
💰 Who's generating the best verified returns
🔎 Which tipsters are worth watching

 Free to join. No spam. Unsubscribe anytime

We respect your privacy. 100% free to join

Leave a comment

Don't Miss Todays Winners!

Get today's free horse racing and football tips now!

Simply enter your email address below.

We respect your privacy. Over 18's only.