Forums

General Betting

Welcome to Live View – Take the tour to learn more
Start Tour
There is currently 1 person viewing this thread.
system99
24 Aug 16 21:36
Joined:
Date Joined: 23 Nov 07
| Topic/replies: 1,933 | Blogger: system99's blog
This isn't a "look at me" post, well maybe a little ha , but am genuinely interested from anyone experienced in systems how my figures hold up and how long before u would feel confident in playing it or any system for decent stakes? 

All Pre Race Lays, max 29/1, place only if over but again 29/1 max

Running 48 days

Horses 1299

Bets 2598 (1 point win 1 point place)

P/L + 493 points

13 losing days, max 23 points

35 winning days, best 53

Any other info important to access it's merits? have a records of horses so let me know

any input welcome and NO i am not planning to "release it" and after peoples cash, I've only actually been real money playing it two weeks and at min stakes for them, add to that i had lets say , teething problems with my staking and proper screwed the first week, should tell you i'm not after anything.
Pause Switch to Standard View How long untill your "confident" in a...
Show More
Loading...
Report system99 August 24, 2016 10:47 PM BST
correction - the number of horses/bets is only last 28 days, so i'd estimate all in more like 400 bets
Report system99 August 24, 2016 10:48 PM BST
4000 tut sry
Report Andriy August 25, 2016 10:31 AM BST
When you have a p-value of 3 start to feel smug, and if it reaches 4, buy a bottle of champagne.
Report Andriy August 25, 2016 10:31 AM BST
aargh i meant t-stat not pvalue
Report system99 August 25, 2016 11:39 AM BST
just tried to read what t-stats are all about, often kid myself that i'm ok at math, not true! reading one paragraph reminded me....I am DUMB as fook.
gis a clue, with the limited info available, what guesstimate would u have my t-stat at?
Report Andriy August 25, 2016 11:47 AM BST
To be honest, my reply was trying to act a bit too intellectual; you're effectively backing at prices down to 1.03, and would need to calculate an average back price.

A simpler answer would be that if you have to ask yourself the question "do i have an edge?", then it's very unlikely that you do.
Report longbridge August 25, 2016 1:13 PM BST
So here's one way of doing what you want to do - or at least to do what I'd want to do, which is determine if you're taking value bets or not.

(I'm describing this for back bets because it's simpler, you'll have no trouble inverting it).

Divide your bets into price bands of equal slices of implied probability:

0 to 5% = 1000 -> 20 \ mid point 2.5%
5 to 10% = 20 -> 10 \ mid point 7.5%

etc.

Now count the % of bets in each price band win.  If you're scoring more winners than the 2.5%, 7.5% etc centrepoint of the range, then you've found value (horses winning more frequently than the implied probability from their price).

Caveats: if you use narrower price bands, your mid-point average is more valid, but you'll have fewer bets in each bucket so sample error will hit you hardest.

I have a 7+ year database of UK&Ire horseracing so for most ideas I have I can find a big enough sample set to backtest; I don't know your exact strategy, if it's not one that lends itself to automated back-testing it'll be a long while before you have enough data to know if you have a working strategy or not.
Report fixed August 26, 2016 2:57 AM BST
after that you have to add 17.25 equal slices of implied probability bands and then multiply your backtest data results with the probability (zero) of it reoccurring in the future
Report system99 August 26, 2016 5:11 PM BST
wish i never posted this for two reasons

1- had my worst day the next day - 26.5 pts

2- keep popping back on and looking blankly/dumbly at replies, which really could be Swahili to me!
Report fixed August 28, 2016 1:06 AM BST
dont worry your short term results are meaningless.....much more important: you created new t-shaped data that will make enough noise to extract a bunch of systems from a fresh set of quantitative analysts
Report eight ball August 28, 2016 2:31 AM BST
Famous expression stood the test of time.

Give me a man and his system
I'll have his money
and he'll still have his system.
Report xmoneyx August 28, 2016 10:39 AM BST
better off not doing e/w,looking back at bets over the year
Report Macsys November 25, 2016 10:04 PM GMT
how is it going now?
Report zooot December 3, 2016 12:54 AM GMT
The answer is that if it is not based on detailed form study you can never be confident long term  because the market can adjust at any time.

But to be a bit more helpful:

- It will be thousands of bets over an extended time.
- The longer the odds the more thousand or even tens of thousands of bets needed.
- It will need to span the cycles of manipulation that go on in many markets to drag the punters in particular directions and to then wrong foot them just as they think they have a system or pattern identified (the powerful big players and bookies vary their approaches over time to obscure their approaches and fleece the sheep faster).
- It will need to span and survive a number of vicious losing streaks that are part of any data set - there can be steady declines or sharp reversals that seem almost too extreme to be just a natural cycle of randomness but often are simply that.
- it will need to survive the impact of commission which impacts heavily in straight betting creating a leak in the bucket that mutes the winning periods and exaggerates the losing steaks impact on your bank.

- long term your system will need to survive the tendency for other punters to spot patterns that create value sufficient to be quite profitable - i.e. if you are looking across a largely flat desert searching for the oasis of value signaled by vegetation and changes in terrain, it is highly likely that a hundred or even a thousand other punters also either spot the same oasis or at least a part of it over time and start to move in.  If your data is what everyone else is looking at and your system is real, has volume and is persistent - other will spot it sooner or later.  Some will move in slowly and nibble away. Others can just hit it hard and kill it overnight.  I have seen numerous patterns persist for say 6 months and then almost over night shift into break even and stay there.  Some of this was chance but I am sure it was because what I had spotted was being watched by others and we all started to scale up our bets at the same time thinking we had hit something good and collectively killed it.

- if you find something, don't draw any attention to it - don't scale up your bets too fast or much.  Use random amounts and placement pattern.  Don't get greedy.  In low volume markets even more so (one small bet in a thin market can send the market running in the wrong direction for you.

A sustainable system almost needs to be marginal or low volume to work long term - it must stay hidden or sit in the "why bother" (such as two bets a week in obscure situation sort of thing category) so other punters overlook it or leave it alone to hit higher volume stuff. 

Or else it must operate on skinny margins and high volume so that it is hidden by the ups and downs of randomness to all but the far sighted, deep pocketed, patient and low commission payers.

The other type of system that will work long term is if your are the manipulator - the deep pocketed knowledgeable, low commission paying syndicate that combines deeply understood form and data based approaches with a layer of manipulation on top to shift markets in your favour consistently.

So in summary, it is all easy and peachy for every new punter to find a winning system and get rich - just a couple of speed bumps described above to overcome.

Excited
Report sean rua December 3, 2016 10:38 AM GMT
One of the best posts on the subject that I've ever read from zooot there!Happy

My quest for answers over several decades has failed, but not for the want of trying, so I think the OP will need to stick with his system for several years more.
However, by then, the once good thing may no longer be.

For reasons, see zooot's post; he seems to have covered most eventualities,
Report DFCIRONMAN December 7, 2016 7:54 PM GMT
No idea how you arrive at lays, and I am biased towards BACKING.

High odds winners do happen, especially in mixed age h/c races types.

If you have been laying in other race types you are bound to get long runs of successful lays, BUT if you don't have enough "edges" to take on such race types, then the losing run can and will happen ( BLACK SWAN day !).

Climate change is likely to cause havoc to "systems", as well as trainers branching out from FLAT to NH and AW racing : NH to FLAT and AW : from AW to FLAT and NH. So your "system" has to be capable of taking such factors into account, as well as multiple other factors.

Can your "system" cope with such factors ?
Post Your Reply
<CTRL+Enter> to submit
Please login to post a reply.

Wonder

Instance ID: 13539
www.betfair.com