General Betting
There is currently 1 person viewing this thread.
These 116 comments are related to the topic:
On the Chase
On the Chase
sort by:
Show
per page
Replies: 116
|
By:
Lex,
The article relates to volatility, which, as is well-known, is serially dependent and therefore (potentially) mean reverting. Series of sports events, though, are obviously independent and therefore the results are not mean reverting. |
|
By:
of course they are, why else would you get the average price of horse favourites being consistant month on month ? up and down they go... and arrive at the same mean.
Or am I wrong ? (careful now! ) |
|
By:
Someone posted that 60% probability equated to "odds" at 2.5.
P=60/100 Odds=60/[100-60]=60/40=6/4 in favour ,therefore the odds are on,i e 4/6. |
|
By:
Ssssh it was the same chap that said
Series of sports events, though, are obviously independent and therefore the results are not mean reverting. So by the same logic 60% = 2.5  I thought it was 40% by hey ho.... |
|
By:
Yes, sorry, a silly mistake.
|
|
By:
DFC hit the nail on the head.
It's not whether results will revert to the mean ( they will over time), it's how they will revert. For a simple example ( just for discuussion purposes). You might observe a hisrorical situation where in 100 events, only 30 % of even chance odds came up. So you then say thus must revert to 50%. It will in due time, but it is probably more likely to do it by travelling at or near the mean over an even longer future time period. For example, if it now hits 470 on out of the next 900 events I ( that is 52.22 % ), then it will have reverted fully as assessed over these particular 1000 events. But at this strike rate you will not win on even chances over these future 900 events, as you will be betting most probably on the bid side always ( which will be just below fair odds) and you will be paying commission on winning bets. And this is not to even discuss things as to whether your observed historical observation is either correct ( after all it was probably done on paper) and also the fact that it itself is only part of unlimited other possible historical observations of even greater length and accuracy all of which will show variances ( big and small) in all sorts of conttradicting directions, and raise the basic question, which one do you chose to attempt meanrsion on ? Contrarian is absolutely right on all this and anybody doubting him is dangerously kidding themselves. |
|
By:
In case anybody was googling " meanrsion" at the end of my post, it was only a typo of mean reversion.
Btw Lex is a well known humourist on here. And so perhaps it is simpler just to quote Monty Python to hm. " That parrot is dead sir ". |
|
By:
Lex : Ive got the gold part of the puzzle sussed, its the perpetual thingy that giving me jip.
Haha, the lump of rock we're living on is pretty close to being in perpetual motion. |
|
By:
Ta Froggy the fine haired one. - a civil reply - well done !
Ive been sifting through some very interesting stats. Once Ive sorted them and if they make sense then I will post up. - it appears that the 'mean' (for horses) is reached quicker than you would have thought ... Cheers All. Ex parrot. |
|
By:
The "mean" is irrelevant....because, if you are backing horses that are say 2-1 SP based on the mean of such odds horses winning 1 in 3 races, instead of backing it because it has a strong chance of winning, and its odds should be shorter......ie VALUE IN ODDS.....then you are just gambling. The "edge" is why you might win....not the fact that horses are 2-1 win 1 in 3 races.
|
|
By:
Thanks DFC I know that cheers, but its not what Im getting at.
Its the type of horse eg clear favourite which you back or lay at Betfair sp. You get prices up to say 7.0 and as low as say 1.2 - the mean is historically 2.85 the average win rate is 33% so you need average odds of at least 3.0 just to stand still. The win rate ranges between say 50% and 25%. Over x races (which I am looking into) the average win rate exceeds say 3.0 ( or whatever is needed to overcome commission etc and make a profit) . The frequency of when this happens is ahorter than you would expect, so a staking plan can be devised. When the win rate is less than the mean you lay when greater, you back. The figures are rough but the principle looks interesting. Thanks for the input guys. |
|
By:
Have you checked out "THE RETIREMENT STAKING PLAN" ?
It takes "strike-rate" in to account etc.....and the AVERAGE ODDS you generally win with etc Might suit you to a tee  http://www.grandstand.com.au/retirementstakingplan/ |
|
By:
Yes thank you. Using a divisor based on average odds is sensible imo.
I havnt looked at divisors yet tbh. |
|
By:
Apparently there is software for this....I've no idea how this works....but might interest you?
http://www.soft82.com/download/windows/the-staking-machine/ |
|
By:
I need to drill down further into past results but so it looks like the revert back to the average happens roughly every 120 races. so during that time period the winning rate goes through and passes the mean average. On that basis then, taking into account the last peak, the favourites will be scoring well (around 40%) over the next 2 days.
we shall see.  |
|
By:
I honestly don't know whether you're joking or not!
|
|
By:
if this works you could apply it to almost any type of betting.... even roulette!
|
|
By:
I do understand that youre sceptical. So am I. But its the logic here. At some stage in the cycle of events, the resultls must be performing better than normal (mean) else all results would be 'flat'
According to the stats we are in a 'trough' and due a 'peak'. over the next 2 days we should see favourites perform better than 'average'. Lets see if they do. If they dont, then clearly I am mad as a box of frogs (sorry FAFH ) |
|
By:
According to the stats we are in a 'trough' and due a 'peak
But you do accept that the events you are describing are independent of each other? |
|
By:
for the millionth time (almost) forget roulette ! it works on sports (bookmaker markets)
 |
|
By:
Lex,
Are the events independent of each other or not? |
|
By:
Oh gawd, you dont get it do you.. they are NOT independant that is the point.
Roulette is independant, sports is not - it is cyclic and 'trendy' We'll resume this once the results come in. You should see the horsey favs performing better than normal over the next two days. If they do, then we know we are onto something, cos Lex can predict when to back or lay. and it costs nothing to wait an see.. |
|
By:
So you believe that the probability of whether X number of favourites win today is somehow affected by the number of favourites that won yesterday?
|
|
By:
I have my fingers crossed for the faves...
|
|
By:
hmmm good question. I dont know.
It could be a few things. a)bookmakers are aware of trends and price accordingly. b)there is a finite number of good horses running that happen to be favourites c)track considitions and other physical characteristics change. But the fact the industry SP is so solid along with the stats prompted this idea in the first place. Even in roullete the wheel gets changed because it becomes biased. |
|
By:
lex, the success with martingale depends on ur finacial ability to take big losses. i am sure u now this, but somehow u want to ignore it.
|
|
By:
why do u want to ignore it?
|
|
By:
I said a martingale type system
eg one that gets back your losses because things are in your favour but that doesnt mean stoopid mad double yer money type staking does it ? sheeeesh |
|
By:
so far so good, the favourites are performing better than expected so far today. But we've a way to go yet. Interestingly according to the stats if the next 16:30 Newton Abbot has a clear fav it should lose. But I think it should win (Blue Hills). Oh the dilemma ...
|
|
By:
me right; stats boll0x
|
|
By:
know
|
|
By:
yesterday I said that favourites will do well over the next 2 days with over 40% winners.
The cycle is around 120 UK races 11-06-2012 17/34 =50%, average SP 2.91 Today it wont be so high but still above the average. My comment re Blue Hills was because it is rare to get 5 consecutive winning favs but I miss counted. The 5th one did lose. 6th conseq is v.rare, 7 happens twice a year and 8 never happens. So if you had 4 winning favs start laying them. If you havnt had a fav win in 10 races start backing them although a few years back 3 days went by without a fav winning - 54 of them. Then over the next 2 days they hit gold and the stat for the week was the same as the previous week ! Digging still deeper into this jockeys and trainers stats (amongst many) are also cyclic. an example that springs to mind is AP McCoy. Bad few days then... boom. His Martin Pipe David Johnson days where extraordinary; shame I missed the boat by not stumbling on this earlier  Im learning excel now so I can graph this lot. Ive got results going back to 1988 ! So you disbelievers - it NOT like roulette ! Its BETTER ! |
|
By:
Why don't you just backtest this idea against historical data?
If you do it in real time, it will take months to get a meaningful sample size. |
|
By:
Daft begger !
Of course Ive tested it against 'back data' - since 1988 ! Thats why I can say 40% or better favs today.. I must be typing in a foreign language or somefing. Luv ya |
|
By:
You've backtest this system and it works! I suggest you've made an error in your calculations.
|
|
By:
Also, if you've backtested it and it works, then why the f*ck are you advertising it on the forum. Keep it to yourself!
|
|
By:
how much £ would you have made per year?
|
|
By:
Im stumped how to reply to that - I know your name is contrarian but that takes the biscuit - so you dont want me to share this with you ?
Ok I wont. JL, a staking and monitoring plan that avoids racking up losses due to losing runs is the next step. alternatively backing and laying according to where I am in the cycle. |
|
By:
Lex- this not an edge.....it is sheer madness.
Each race should be weighed up on its own merit.....using edge/s that you know "work" on that type of race at the distance of race and on the going for that race......and also taking every horse's form etc into account. Each favourite is either a value selection, or about rightly priced, or poor value .....taking all factors into account for each race. The previous day results re favourites are irrelevant to the next day races......as are the previous race result. Each bet should be based on the amount of value on such bet....this is the hardest part to weigh up. STAKING should not be based on previous losses.....but THE RETIREMENT STAKING PLAN does take strike rate/average odds / bank size etc into account, and probably is the "best" way for you to go ....based on what you have posted. GL with your trial. |