Australian
There is currently 1 person viewing this thread.
These 89 comments are related to the topic:
lay betting suggestions
lay betting suggestions
sort by:
Show
per page
Replies: 89
|
By:
what r your inefficencies in AFL head to head, Dark. the only inefficiencies i can see is that if the oods shorten or drift is that the line very rarely keeps adjusting to the selling odds
|
|
By:
Guys - this is nothing even close to a system.
Loggy - of course I respectfully disagree. Take a simple variable, "distanced traveled" or "home team". The impact of these variables has hardly changed over the course of the last 10 years. There are many more predictive variables you can use that haven't changed over the years. Sure air travel is easier these days, but the effect of travelling and the amount of time the home team wins barely has. |
|
By:
Loggy - the market (even on betfair) is heavily biased toward the home team for start.
|
|
By:
I have to make this point here because I feel it is grossly misunderstood:
EVERYTHING is a system. The difference between the type of system we are talking about and race ratings in their various forms is that a "system" attempts to circumvent the vast amount of work required for a more traditional rating. Ratings are, however, still a systematic means of processing and interpreting data. So can we say: "We don't subscribe to the value of shortcut systems." |
|
By:
not really bias towards home teams but to home fav's.
|
|
By:
Not just faves, Loggy. I think Dark has a point.
|
|
By:
no short cuts crazyone, just watch every game week to week
|
|
By:
I would say a system is something with a set of rules?
|
|
By:
Isn't a horse rating system exactly that?
|
|
By:
How can something so simple and so obvious like home teams be biased on a supposedly efficient market like Betfair? It's staring everyone in the face yet its still there...
There are no doubt many more but it's not worth my while at the moment to investigate further without all the data i need  |
|
By:
I sidetracked a bit...
I would prefer not to debate that too much. I don't think anyone will be changing anyone else's mind any time soon. |
|
By:
no rules, unless you r ZR and even then rules are always broken i'm talking sports not horse racing then i would agree to a set of rules/system
|
|
By:
I can tell you why, Dark....
Because there is fck all volume unless we put it there. Most guys don't show up for a bet til 2 hours before the game. |
|
By:
Yes a system is, mine doesn't have any rules.
It has many variables that a regression algorithm decides how much weight to give each depending on circumstances before producing the final probability. |
|
By:
I can be a pedant but I won't. I would prefer to tackle the more interesting facets of this one.
|
|
By:
crazy u were going to say rules=system.......system=rules
werent u |
|
By:
Regression needs a whole stack of data points.
So then you are talking about general trends rather than specific data points, being a horse's form. |
|
By:
Loggy
 |
|
By:
ie. so generalised trends for horses with a particular set of characteristics, rather than specific ones?
|
|
By:
This is a fascinating approach! I would never have thought to tackle it in that way, if that in fact is what you are doing...
|
|
By:
Yes that's closer to the mark.
At some point i'll probably just buy the historical time stamped data from Fracsoft but for the time being its on the shelf for other things. |
|
By:
Crazy thats my point - I wouldn't imagine many are doing it this way
|
|
By:
My mind is officially blown
 I need a break lol |
|
By:
If I say more than 3 "wow"s in a 2 minute period I walk away from my desk...
See ya. |
|
By:
me too, i'm going to give birth to a little mrben
|
|
By:
LOL!
I was just thinking i need to take the kids to the pool (no inference to mrben on my part) |
|
By:
Ok, wait I had to come back. My mind is racing...
So let me get this straight...I can make profits if I can create a set of principles such that the vast majority of horse perform to these "rules" So for example, if I know the percentage of horses that win first up after a 6 week break I can include that in my calcs? Or not that kind of thing? |
|
By:
Ok, kids come first, we can pick this up later.
|
|
By:
Ok on the original subject at hand...
I would suggest an idea that yields -5% on only 4 rules has potential. A tweak here or there might get you over the line. Eeven break even could be useful. If you're a PC player it might be a way to normalise your profits and reduce the PC you pay. That alone might make you 20k per annum! |
|
By:
Temlee if you go to racing and sports a good website they have a customised form sheet which may be helpful to your needs its not hard to work out and i find it very handy i understand a database is not what your looking for.
|
|
By:
Sort of Crazy. A bloke by the name of Michael Nunamaker wrote a book on what you are suggesting, i.e. how often horse win off a 6 week break. I believe his method was to give each impact value (don't use strike rates they aren't accurate) a weight and combine them together. I believe he was succesful but its seriously flawed...
Take a very basic example of, a horse off an inside barrier who loves the wet. Two very good variables on their own, but what about when you combine them together on a track where the inside is off when its heavy? You have no way of correlating the 2. Now take logistic regression (a vastly superior option). One of the most predictive variables Benter and Woods (the best in Hong Kong) had in their model was how many starts a horse has. On its own it really shouldn't give much indication to their winning chances... But lets say you add in the horses average speedrating with its total starts. When you combine the 2, the regression algorithm can determine the correlation between them. It knows that a horse with 2 starts and an average speedrating of 100 going up against a horse with 100 start and an average speedrating of 100, the horse with 2 starts is more likely to win because it can tell that as the less starts a horse has had, the more likely it is to improve. |
|
By:
Thanks clarkie55 it took 70 replies but finally someone came up with a suggestion
|
|
By:
My database tells me that Richmond won the flag in 1980 so if they took the same side into this weekends game I would be able to sell my database to that fella who spends 15K setting them up, and back them would'nt that be lovely.
|
|
By:
Ok, Dark, well I get that point...
I guess there was a lot of trial and error for you to come up with a working model. Ironically, you HAVE to use past results as an indicator of future performance, something which itself has the potential to be deeply flawed. |
|
By:
WHEN YOU GET A DATA BASE THAT CAN TELL YOU WHAT IS GOING TO WIN OR LOSE TOMORROW POST IT ON THIS TREAD STOP DREAMING HOW THE PAST IS GOING TO TELL YOU WHAT WILL HAPPEN TOMORROW YOU ARE JUST A COMPUTERHEAD NOT A PUNTER.
|
|
By:
Temlee - I have no interest in what horse is going to win tomorrow, my only interest is in getting my win probabilities more accurate than the rest
|
|
By:
temlee,
If I'm not a punter then what the hell have I been doing the past 10 years? |
|
By:
Let's be serious for a moment. IF I had a successful method/system that I could share with you, why the hell would I? You're obnoxious!
|
|
By:
Looking at your database thats what your been doing, by the way when did you last have a bet if you haven't then get your database and get of here
|
|
By:
you nor your database has nothing to do with lay betting or lay betting systems or any suggestions concerning the subjet so just go to bed.
|