CSF / Exacta

[radders]

Hi All,

I know this has been up for discussion before now, but it was raised again this evening in my local.

Q.

How is the CSF dividend arrived at?

I had responded to earlier thread with the time tested 'rule of thumb' calculation based on the simple method of multoplying the odds together to arrive at a reasoable approximation of the CSF .. e.g winner at 3/1, 2nd @4/1 would usually pay approx £12 to £1.

However, yesterdays 100/1 & 50/1 result at Wolves thru this out and a lawyer mate who had been there, re-started the debate on what the official guidelines / calculation methods are for this type of bet..... the exacta is based on the Tote pool (I think?) ... what 'computer' calculates the CSF?

Q.

Has anyone noticed an increased dividend level on CSF since exactas were identified as paying more in some cases than the CSF, and are there any stats on the cumulative payouts if you followed one method or the other?

Regards
Radders

[GlosRFC]

The CSF is a complicated formula that takes into account the winning SP's of the first two horses AND the number of horses running - it is deliberately biased against small fields in the event that short-priced out-and-out favourite is pulled up or falls. It also takes into account other anomalies such as draw bias on certain tracks.

As you say, the Exacta is based on the amount paid into the pool so it's generally a bigger payout to £1. However the CSF £1 return is fixed so if you stake £10 you will get 10 x the CSF return. Likewise £100 on the CSF would give you 100 times the return. With the Exacta your remaining 9 winning bets will significantly lower the returns therefore the overall payout will be less (and even less still if you'd staked £100).

One final thing to remember is that the Exacta is not always won so the "dividend" declared is the entire pool. That's why the Exacta initially looks like it gives a higher return.

With the 5.00 Wolves example there probably wasn't a single winner so the entire pool was declared as a "dividend" but, if we assume there was one, they would've received about £1545 (£2033.80 less the 24% the Tote extracts) for their £1 bet. If another 9 punters had placed a £1 bet on the winning forecast then that original punter would only have received £154.

With the CSF all 10 winning punters would've been paid £2425.03

[radders]

Thanks Glos,

It's the 'complicated formula' mentioned in first part of your reply that I'm interested in.

Is it a fixed formula and is it written down anywhere within the industry or is it a set of variables that may or may not be included dependant on circumstances?

Thanks & Regards
Radders

[GlosRFC]

It's fixed...and then again, it isn't.

Essentially the formula is fixed for a given race at a given course with a given number of runners and given a certain combination of SPs and, no doubt, given certain other variables. So if any of these variables change, the formula produces different results.

However the formula itself has been changed over the years as the bookies attempt to eliminate the occurrence of some anomalies which has exposed them to greater risk. These changes include the examples I previously mentioned about accounting for very small fields and perceived draw bias at certain tracks.

I've no doubt that the formula, if it doesn't already do so, will be changed again in due course to take into account the impact of banded racing - in particular the greater percentage of higher priced combinations of horses filling the first two places.

Precisely what the formula is, how it's derived, who calculates the return after each race, etc. appears to be a closely guarded secret and it's doubtful if even your local bookie or betting shop manager would have a clue. An enquiry to the head office of one of the leading bookmaking firms might possibly give you more information.

[GlosRFC]

Right, after some further research, what appears to be happening in a CSF is that the winning horse is effectively ignored - after all the original SP takes into account some of the anomalies I've previously mentioned and the SP should also be an accurate reflection of the weight of money wagered on that beast and on the race as a whole. This should also, in theory, be an accurate reflection of the incidence with which that particular animal occurs in a forecast bet, i.e. an odds-on horse is likely to figure in far more CSFs than a longshot.

The real issue is the horse that comes second so, for all intents and purposes, the SP of that horse is altered by discounting the winning horse and recalculating all SPs to generate a more accurate overround (i.e. profit) on that horse for the bookmaker. A second adjustment is then made to account for the size of the field to reduce the risk of two high-priced horses occupying the first two places, which is why the Wolverhampton example returned £2425 instead of the expected £5151.
A further adjustment is then made to reflect the overall weight of betting on the race that would've been made if the winning horse was removed from the equation. The purpose of this is to ensure that the overround that the bookie expects to make on the entire race is consistent.
Yet another adjustment is made to reflect the type of race (different courses, flat/hurdles/chases/all-weather, draw bias, etc.) which takes into account statistical anomalies and reduces the risk to bookmakers.
Finally I believe that there is then some kind of "ceiling" imposed on the result, again designed to further reduce the risk to those poor old bookies.

The formula itself is revised and drawn up each year by a team of mathematicians and statistical analysts employed by the Betting Office Licencees Assocation (a cartel of the leading bookmaking chains) and you can request your own copy of it by writing to BOLA, 3A Lower James Street, London W1 or giving them a call on 020 7434 2111.

Hope this is useful.

[vegyjones]

Wow, that's good stuff Glos...
Well done for finding all that out!

It would be far easier to multiply the odds of the horse that won, by the odds in second would have been if the horse that won was a non-runner. But as you say... probably wouldn't have been in the interest of the bookies as the other day's 100-1 followed by 50-1 would have been 5000-1, taking into account that horse 2's odds probably wouldn't have been affected by horse 1 being a non-runner.

If that makes sense! :doh

[GlosRFC]

Wow, that's good stuff Glos...
Well done for finding all that out!

No problem Vegy, I needed to do something to generate a few reputation points :D

[vegyjones]

:laugh

I'll stick to my begging tactics rather than go through all that lark Glos! :D

[radders]

What can I say Glos?

Research of the highest order!

I'll share this with the pub tonight.

Thanks again,
Radders