Bookmakers, and one firm in particular, appear to be systematically conning punters through a marketing strategy originally designed with polar opposite intentions.

My own exposure to this started out as a perception, after a series of bets struck early succumbed to swingeing Rule 4 deductions. And, off the back of a frustrating triple whammy price cut in a single race, I decided to dig a little deeper.

Background

The apparent con relates to early prices offered by most big name bookmakers. These odds are generally made more attractive to smaller staking punters and those who like to bet the night before (perhaps due to time constraints or work commitments), by the provision of a Best Odds Guarantee, or BOG.

This 'BOG' concession means that punters will receive the better of the price they take and the returned starting price. So far, so good. But if a horse is withdrawn from the race between the time the bet is struck and the time the race goes off, a 'Rule 4 deduction' is applied.

Tattersalls Rule 4 - and all its other rules - can be viewed in this pdf document.

Rule 4 (R4) deductions are calculated based on the odds of the non-runner at the time of withdrawal, and involve a deduction from the bettor's stake. The R4 table is shown below:

1/9 or shorter = 90p in the £
2/11 to 2/17 = 85p in the £
1/4 to 1/5 = 80p in the £
3/10 to 2/7 = 75p in the £
2/5 to 1/3 = 70p in the £
8/15 to 4/9 = 65p in the £
4/5 to 4/6 = 55p in the £
20/21 to 5/6 = 50p in the £
Evens to 6/5 = 45p in the £
5/4 to 6/4 = 40p in the £
13/8 to 7/4 = 35p in the £
15/8 to 9/4 = 30p in the £
5/2 to 3/1 = 25p in the £
10/3 to 4/1 = 20p in the £
9/2 to 11/2 = 15p in the £
6/1 to 9/1 = 10p in the £
10/1 to 14/1 = 5p in the £
Over 14/1 = No Deductions

 

A good example of how to calculate Rule 4 on your bet can be found here.

The Rule 4 deduction chart is fixed, and seems entirely fair. After all, if for instance the withdrawn horse is only a 4/1 shot, it must have had reasonable claims in the race; so the chance of my selection is materially enhanced by his absence.

But a key clause in the Rule 4 documentation is this:

Where a bet has been placed and a price taken on the day of the race and there is subsequently an official notification that a horse has been withdrawn or has been declared ‘not to have started’, the liability of a layer against any horse remaining in the race, win or place, will be reduced in accordance with the following scale depending on the odds current against the withdrawn horse at the time of such official notification: [scale as per above table]

To labour the key phrase, the bookie liability shall be reduced against the scale based on the odds of the withdrawn horse "at the time of such official notification".

The Problem

If there is no problem with the Rule 4 system, then where exactly does the issue lie? Take a look at the image below [click on the image to open fill size in a new tab], which shows the odds movements of three horses from the same race on odds comparison site, oddschecker:

 

At the intersection of each vertical and horizontal red box, there are two items - a number in blue, and the letters 'SUSP'. The blue number represents the revised odds of the horse in question, which have shortened (pink equates to lengthening odds revisions). SUSP means betting has been suspended on that horse.

Now note the times to the left of each horizontal red box.

The proximity of the blue numbers - i.e. shortening odds - to the suspension of betting is interesting. Actually, it's downright concerning.

In the first two cases, the difference between the shortening odds and that horse being suspended in the betting was five minutes. In the last case, it was a single minute.

In all three cases, the time the horse was shortened was later than the time the horse was declared a non-runner on the official BHA website.

In other words, it was possible to know beyond doubt that the horses shown above were not running before shortening their odds.

Why is this important? Because if a horse is 14/1 or shorter and becomes a non-runner, bets placed on every other horse in the race prior to this horse's market suspension will incur a R4 deduction.

It remains possible that the triple example above was merely coincidental. Suspicion grows, though, as it becomes clear that two of the three non-runners were shortened into the next R4 deduction bracket.

In other words, it could be inferred that prices were deliberately managed to mean a multiple deduction on any bet left standing on any other horse in that race.

 

A Closer Look

Now of course a single race - no matter how frustrating in 'bookie tax' terms - hardly constitutes a body of evidence. Nor either does the anecdotal evidence that was its precursor.

And, in fairness, what follows is less than a categorical damnation; but it does further illustrate the issue and focus a beam on one particular operator who could be said to take greater advantage of this perceived con trick than most. (Note, this operator is NOT the only one perpetrating the practice).

The bookmaker in question is PaddyPower, and the following data relate to that firm's prices as represented by oddschecker.com.

I looked at all of the non-runners from Monday 13th July up until 4pm GMT (or 4.01pm as it turned out), and noted the following:

- Official time of withdrawal of each horse (NR Time)
- The penultimate betting show for each horse (Show -1)
- The time on oddschecker of the penultimate betting show (S-1 Time)
- The R4 associated with the penultimate betting show (Assoc R4-1)
- The final betting show prior to withdrawal (Last show)
- The time on oddschecker of the final betting show (LS Time)
- The R4 associated with the final betting show (Assoc R4)
- The difference in R4 deductions between the last two betting shows (R4 diff)
- The time the horse was suspended by Pad Pow for betting purposes (Susp Time)
- The difference between time of suspension and last show time (Susp-LS Time)
- The difference between time of official withdrawal and last show time (LS-NR Time)

In the below table, I've used the following presentation formats to highlight a couple of things:

Grey Italics - No change in the horse's odds all day
Green bold - Favourable R4 movement between last two shows
Red bold - Unfavourable R4 movement between last two shows

Here are the data in race time order:

Although the above is not that easy to read, a few things become apparent. First, there is a lot more red than green, meaning that - on this day at least - punters came off much worse 'nett'  when comparing positive and negative R4 deductions. Specifically, there were eight R4 deduction extensions in the final shows, and three R4 deduction reductions (if you catch my drift).

Second, note the comparison between the official withdrawal times and last show times (right hand column). In all but one case, the last show was after the horse was officially withdrawn. And, therefore, in direct contravention of the stipulations of Tattersalls Rule 4 (C).

Mr Power's published rules with regards to Rule 4 can be seen here, and are in line with Tattersalls Rule 4 (C). Thus there is a disconnect between Power's stated rules, and the company's application of them.

*

Here's another interesting thing: let's sort the data by the size of the last show Rule 4 (Assoc R4)...

 

Notice how that raft of red has gravitated towards the top of the table?

This means that, in the admittedly small sample of non-runners, those highest in their respective markets were most susceptible to what I'll cautiously suggest could be interpreted as manipulation.

Fridge Kid, the sole green entry at the top of the list, was the only horse in the table (excluding the grey italicized entries, whose prices did not move all day) that did not change within at most five minutes of the suspension time.

Now, it's important that I caveat this: it is possible that this is a 'feature' of the way Padd.y's data feed talks to oddschecker. Possible, but unlikely, in my opinion.

 

Conclusions

There are reasons why it is dangerous to be categorical about what the data appear to suggest, starting with the fact that I don't know enough about how oddschecker receives and processes data from the PaddyPower feed.

I do however know something about how this data is consumed by geegeez's own odds comparison engine, and it is reasonable to assume similar processing happens on oddschecker.

Also, there is insufficient data in the sample to conclude anything stronger than a tendency towards 'aggressive book management', as opposed to a less equivocal and more absolute contention.

Unfortunately, the information aggregation process is time consuming and not something I'm able to commit to longer term. However, I will 'spot check' this from time to time over the coming weeks, as I believe it is more than mere coincidence.

However, despite all the caveats and the limited dataset, it is clear that P.P. are consistently revising their prices after the official notification, and using the revised price as a basis for Rule 4 deductions from pre-existing wagers.

I contacted PaddyPower's twitter team, hoping to get a comment from the trading room on the matter but, as yet, have not had a reply.

In lieu of a response, it seems sensible to tread carefully when taking an early price with PaddyPower. Personally, where another bookmaker is offering the same price, and the same BOG concession, they will be my chosen layer.

Ultimately, if this is indeed deliberate - as appears likely - the notion of giving with one hand, via the availability of early prices, and the Best Odds Guaranteed concession; only to take with the other, by manipulating the odds of known non-runners, is a fiercely sharp practice. And bookmakers who deploy it have no place in punters' considerations. You hear me, Padddy?

Matt