To hedge or not to hedge that is the question?

This may seem like a stupid question but regarding laying early prices and then hedging for a profit later.

I have been keeping records of these trades, and where they are successful (which seems to be around 70% of the time), and the horse has drifted out to greater than 1.3 times the lay price struck earlier, they very rarely win the race.

It struck me that perhaps hedging these drifters with low probability of success was wasting hedge money. Are there any maths geniuses out there that can work out a formula for hedging versus un hedged lays for these drifters.

Clearly the drifters do go on to win, but looking at the stats from Betfair;  horses in hcaps with a bsp between 12 and 13 are averaging a win ratio of 6.9% from over 2100 bets in 2009.  But if I lay them early at 12.5, on average and they drift to say an average price of 25 when they are hedged, the real probability of the win is the same as a 25.00 chance which on Betfair is 3.96% not the 6.9% lay price.

Therefore assuming the average lay is 12.5 and the average hedge is at 25 and the chances of the horse winning is a mere 3.96% would it not be best to leave the bet un hedged, safe in the knowledge that at worse only around 4 horses in a 100 will win? The stats from my own live betting show that from 154 bets where the lay price was averaging 12.5, only 6 of these horses won with an average sp of 28 which I had layed at 12.6 average.

I have spent hours on this so if anyone can put me out of my misery agonising over the long term outcome please let me know. 
Thanks a lot guys…….

This is interisting!

Here's the breakdown!

First and foremost 125 bets it's not sufficient to work out a strike rate for horses odds between 12 & 25. You need more like 600 bets for an idea of a true strike rate.
Ok, that aside looks like you are on to a betfair winner (just had to do that) and the bottom line will lie in your mentality.

hedging is nice safe steady profits which should see your bank go higher and higher everyday, slowly.
Un-hedged you will have lot of days of phenomenal growth, but might come unstuck for a day here and there maybe even the odd week as you get two or three in close succession. The key will be to stick with it.

I know what I am (un-hedged ... see sig) ... but do you know which you are??

If you don't know ... My advise is to do a trial run .. say you are laying $10 .. then only hedge out $5 or $7.5 (whichever you are comfortable with) ... do this as long as possible (150 - 300 bets/days) and then decide.
Keep accurate records everyday and have a look at the bottom line (running totals) ..
day ..... hedge...vs...un-hedged
D1  ........ $3..............$8
D2  .........$5 .............$-20
D200 ......$350 .........$600

Looking back after this (maybe even during) trial run you will realize which you prefer.

Looking back it's easy to say the more profitable route is the right one .. but had the six losses come over a 3 week period resulting in possibly no profits (maybe even a losing run) over that period would you have had the stomach for it??? Or would you have possibly scrapped a profitable system??

Many thanks for your input on this subject 1oser

Very good advice as ever,  and all taken on board. Happily though some of your advice has already been done in live, as trying to trade in test is much harder to get matched at the prices i need. As you know test bets  are not displayed to potential punters in test.

You are absolutely correct about the sample size as well. I would never pay much attention to any sample of less than 2000. This is why I sampled the Betfair returns of over 120000 results from 2008/9  in order to get the win ratios in each racing code for each BSP.

I assumed rightly or wrongly that the drifting lays from say 12 to 20 would return a similar win ratio to 20/1 winners returned at bsp in 2008/9 IN THE LONGER TERM and not likely to be much worse.

I now have over 22000 bets in live play 95% or which are hedged, but by using the excel spreadsheet to highlight certain categories like the big drifters, and seeing the potential profit from those lays the temptation is there to leave the hedge alone!
But as you can imagine trying to calculate a definitive profit /loss ratio based on the 2 different scenarios (hedge or not to hedge) is proving very tricky, for 22000 samples!

As you correctly highlighted those un hedged lays can jump up and give you a hiding.  On my very first day of attempting this approach I had 2 such results go against me costing me about 7 days of normal profits.  I then succumbed to the gamblers worst enemy of “fear” and reverted to the hedge approach whilst licking my wounds. Since then not a single drifting lay has won! I am very much a stats man so with the stats in my favour i would definately have carried on, but i cannot get my head around it.

Many thanks once again for taking the time for your input it is much appreciated.

I think your last paragraph said it well .. you got stung twice and bottled .. so you are are hedge man.

I would recommend you stick to hedging for a good while, because Murphy's Law ... you will get stung again when you go un-hedged.

Hi betfair winner,

If your stats of your first post are correct then you would be better off to hedge.

You are saying that 6.9% of lay bets lose at an average price of 12.5. That is indeed value, as statistically8%  are supposed to lose.

If you don't hedge your profit, after 1000 bets (bet size 10) is:

win bets: =931*10*.95 = 8844.50
lose bets: = 69*10*11.5 = 7935
profit: 909.50

with hedging:

win bets: =971*5*0.95 = 4612.25 (hedge for equal profit)
lose bets = 29*10+11.5 = 3335
profit: 1277.25

That is if really 40 out of the 69 winners will drift to a minimum price of 25 first before price is shortening again and the horse eventually wins the race.

Based on your statistics you could even enhance profit if you don't hedge for even profit but just cover your potential losses (i.e. placing a back bet of 10 instead of 5).

But better check your statistics again. You really only have to look at those runners that have drifted to at least 25 from former 12.5 and not at all runners that were priced 25. I believe there is a catch here as it is a very rare scenario that you can make better profit with hedging compared to no hedging.

If it is that just average priced horses at 12.5 (BFSP) are winning in 6.9% of the races and horses priced at 25 (BFSP) are winning in 4% of all relevant races then the horse that is priced at 25 cannot provide you any value and it is better not to hedge (but at those long odds you need a sufficient bank and especially good nerves!).

If you want to enhance profit with hedging you require two "value bets". In most cases a hedge bet will eat up the value that has been secured with the original bet. That's why I'm usually against hedging (especially during in running) and I'm looking for straight forward value bets.

But if you have all data in Excel, then Excel is a wonderful tool to calculate any potential profit and loss scenario.

Cheers,
bamthwok


P.S.: Do you have a database of BFSP, lowest BF in running prices and highest BF in running price for the winner??? I was looking into the possibility to extract them from the Betfair result page but did not find the time yet to do so.

Hi Bamthwok

I Love the appraisal of the scenario and your workings out are spot on. Interesting that you also that you prefer to adopt a value approach to you positions and press this when a horse is drifting.
Indeed my stats back this up. From the calculations you have done it would appear that the figure of 1.3 times the lay price taken has been omitted from the sums?

Below this figure the gambles that frequently take place show conclusively that leaving these un hedged is the route to the poor house, so much so that I adopted some time ago your suggestion of backing these steamers that were layed at bigger prices in addition to the hedge. However the ones that drift are the ones that really concern me for leaving un hedged.

The problem with excel calculating the profit and loss scenario for the category of the 1.3 times the lay price taken is that I cannot compare the hedged profit to the un hedged profit.

I have got all the BFSP prices, so to filter all the 1.3 and above lay price Versus BFSP is easy. Doing this for August and September reveals
1562 lay bets where the BFSP was at least 1.3 times the lay price taken
Of these just 83 went on to actually win the race at an average BFSP price of 24.4
The average price of all the lay bets of the winners was 11.9.
The winners horses unhedged took -£1477 from the bank meanwhile the un hedged lays that lost the race added +£2403 this left a profit of £925.00 (before commission) But how can you calculate the hedged profit or loss on this?
And here is the problem because you cannot highlight a column and get the average lay price and average BFSP and calculate profit and loss.
 
You can have a lay at 12 and a hedge at 24 three times and then a lay at 16 and a hedge at 4 just once and your average tells you are in profit. The average lay is 13 and average hedge is 19 hence on paper you make money. However this is not the case since £1 on each would give you 50p on the first 3 bets, but a £3 loss on the 4th bet. So each profit and loss for each bet need to be calculated separately, and with all the other bets in the spreadsheet and the problems of identifying which trigger fired which bet this is near on impossible to calculate en mass

Anyhow to answer your question re the stats from Betfair I spent around 2 months
gathering these 1 day at a time for 2008,2009 and up till May 2010 until I finally lost the will to survive any longer. They do however show in running lows and highs. Hope that helps.

Thanks again for taking the trouble to reply to this thread and and your very helpful observations. I hope I come up with the definitive answer soon.

Cheers for now

P.S.: Do you have a database of BFSP, lowest BF in running prices and highest BF in running price for the winner??? I was looking into the possibility to extract them from the Betfair result page but did not find the time yet to do so.

Anyhow to answer your question re the stats from Betfair I spent around 2 months
gathering these 1 day at a time for 2008,2009 and up till May 2010 until I finally lost the will to survive any longer

Hi,
Are you guys aware of this website that has Betfair SP's available for free download going back to 2008?

http://promo.betfair.com/betfairsp/SP_history.html

OT

Hi Obias Trader,

I did not know about this BF service. Thanks very much for highlighting it. Very useful!

Cheers,
bamthwok

Indeed my stats back this up. From the calculations you have done it would appear that the figure of 1.3 times the lay price taken has been omitted from the sums?

Yes right, I did not consider the figure of 1.3 times the lay price but only calculated the concrete numbers given

Hi Bamthwok


 
You can have a lay at 12 and a hedge at 24 three times and then a lay at 16 and a hedge at 4 just once and your average tells you are in profit. The average lay is 13 and average hedge is 19 hence on paper you make money. However this is not the case since £1 on each would give you 50p on the first 3 bets, but a £3 loss on the 4th bet. So each profit and loss for each bet need to be calculated separately, and with all the other bets in the spreadsheet and the problems of identifying which trigger fired which bet this is near on impossible to calculate en mass

Why not you add columns with P/L and calculate P/L for each individual runner. For example you can filter all runners with BFSP 12 and then check how many of them reached at least 12 x 1.3 = 15.6 (or 16). Then you hedge them at imaginary price of 16 and calculate a profit of 2.5 (for a stake of 10).

You do this calculation for each individual BFSP and later you can group price ranges together.

But I'm not sure what exactly you want to calculate. Can you give me very clear scenario of what you want to calculate? I will try to help you out here.

Cheers,
bamthwok

I am being very dim but I have re-read this thread twice and although I understand and can follow all the maths issues what I can't understand is what your lay criteria is. Am I missing something you have noted on another thread?

I am being very dim but I have re-read this thread twice and although I understand and can follow all the maths issues what I can't understand is what your lay criteria is. Am I missing something you have noted on another thread?

There is no lay criteria!

post a profitable scenario in a public forum and you can kiss the profits bye bye

oh 1oser
You do make me laugh. But interestingly re "kiss the profits bye bye", Not necessarily.....................

I made an almighty cock up yesterday. Having been gradually increasing my lay
staking for a couple of weeks. I changed all the trigger conditions to lay matched is less than £4, from the £3 it was previously.
 
Imagine my surprise after coming home in the morning having walked the dogs, to see a horse at Kempton laid at 13,12.5, 11.5, 11 and 10.5 for £44 in total, with a liability of £528!

I had inadvertently punched in 44 instead of 4 and not realised. So it may be that there is plenty of liquidity in the triggers after all for a dozen or so people to run the same block. Happily the horse drifted to 14’s before I lost my bottle and hedged before the auto hedge kicked in. It did actually drift moments later to 16's
These are the ones that Alfaman and i have been working on ever since we won the trigger competition earlier in the year.

Hi Obias-Trader
Thanks for posting the link to the down load from Betfair.
This is where i went to get all the results. Sadly though you do have to take it 1 day at a time.

Hi Obias-Trader
Thanks for posting the link to the down load from Betfair.
This is where i went to get all the results. Sadly though you do have to take it 1 day at a time.

Not really! After I got the link to the results, I quickly started to work out how to get all these data into one single database.

It's definitely possible. First you have to develop a VBA routine to download and safe all data behind the hyperlinks and after that you have to merge all csv files to one single csv file. There is a simple dos command (command prompt) for it.

I'm working on it right now.....

Cheers,
bamthwok

Hi Bamthwok
Many thanks for your suggestion. This is exactly what I have done.
It works well and is suggesting a clear advantage for the un hedged bet longer term, as i suspected.

Once I have built up my nerve again I shall have another go at leaving the drifters un hedged.

Good news also that you are able to get all the data from Betfair without spending hours doing each day by hand. I cannot bear to think the number of hours i spent on this.

Broadly speaking however 208/9/10 have been reflecting each other in the areas i have interest in, and I am not sure how much more benefit I would get from additional years. I already have around 100000 rows so I guess that is enough to be going on with!

Thanks again for your help
BW