do systems work?

They only work if their criteria result in value (ie that their selections' odds are overpriced by the market).

do you think its possible to look at past data, analyse the data and then find a system that works on it?
Yes

then check back results and if it proves profitable does the system really work or is this a fluke?
Pass

i know betfair odds will never be the same but if its proving profitable enough then surely it will work and there wont be that much of a difference between odds?
Pass

as for the premium charge problem at small enough stakes this isnt a problem, always the option of moving to the purple place...
Yes

finally do you think "systems" actually work and im not talking about bots maybe trading in and out at 2 clicks lower etc..
Don't know

i had a system that made me plenty for 3 years....
profits in 4th year where little but still profit........wasnt enough for me so i had to go back to the drawing board!!!!!

well done freddie but was that by looking at results finding something that seemed to always be winning or losing and checked back to see if it was profitable?  if do then you would agree with me that its possible to work something out in this way??

reason i ask is i think i have got something that works (nothing mega but profitable).. but before i put too much money into it i want to see if people think it may work, i dont have enough knowledge on horses but checking back results for my characteristics of the odds etc seemed to prove profiable
im just trying to see if there is any reason that checking those results back and seeing that it proves profitable would not work when put into action on betfair?

Just find urself 1 lay a day, todays was bertolini, never gonna get home at york

systems always need tweaked

your system for example......id say its based on a certain type of race........whether the horse has run once or never........etc etc



how many rules has your system....3,5,7,10??????

6 months ago I would have said yes.

Back the favourite if it is at least 3.0 and at least 1.0 clear of second favourite 30 minutes before the race. Data analysis showed a clear profit over 3 months. Tick for the trial. Use a bot. Worked for 1st month but could not sustain the eventual and inevitable run of losses. Also bot cant just stop at a single point in time.

In UK dogs back the 2nd, 3rd and 4th favourites in any race unless favourite is less than 3.0. same stake with a recovery plan over 2 cycles. Fantastic profit on 1st two months for the trial using stop loss and stop profit, but after that relentlessy lost the profit and the bank.

Systems like this can be heartbraking and also very satisfying.

The only one I do now is every evening I turn on the bot to lay US 2nd Favs to a max liability of £20 per race, with a max of 2 races unsettled. I stop loss at £40 and stop win at £50. Its a bit random and a game but I have never yet lost over a week. I never run it sat and sun nights.

If you knw any I will always try.

several freddie mate, but thats not important i doubt..
if its worked in 2009, AND 2008 surely it works for 2010? how wrong could i be? haha

Know your horses and form, no short cuts im afraid, you gotta put the hours in.

Cricketer thats fair enough things may work for a month then fail however what if its proved over a year or 2, surely thats more trustworthy and proves its more likely to work?

yes they do, but be careful not to 'back fit' when looking at past results

also ensure your test sample is adequate enough, depending on how many selections were provided 2 years results may not be enough to confirm

Statistics don't give answers they just provoke more questions.

Stevie, If I wasnt working maybe I could do this. You can gert historical data. I got a guy to download betfair data for 2 full years and did the analysis. Its just that I cannot sit in front of a computer 30 minutes before every race for example. And as its a 33% hit rate the chances are you might miss the winners if you do not bet in every applicable race.

as an example.

You need one that requires a press button and fcuk off to work. Bots are good but apart from the US one I havent quite got there on one that just creams a bit of money off each race.

what do you mean backfit mate? (maybe stupid question) haha

im convinced ive worked something out and have been using it with small stakes and its worked so far, scared to increase stakes incase it fails..

i tried it once before with a small bank and stupid staking.. on the first day i lost it all haha the reason being was i didnt stake in consideration the amount i could lose on that day... also turned out to be the only losing day of that month and the rest was very profitable haha

stevie can you live with your system having a bad day.....a bad week....or even a bad month?   
take for example this week......how many lays bets did ure system tell u on monday? was it a profitable day?
tuesday?what is a profitable day?
wednesday?was it a profitable day?

freddie

monday - profitable
tuesday - loss
wedensday - profitable


but that loss was only half of what i made on monday...

should be able to tell if your system is backfitting by your criteria for a yes bet. Lets assume the following has to be met:
Horse C & D winner
Break of between 20-30 days
Jockey won on horse
Trainer having at least 2 winners in last 14 days
Horse has the word 'diamond' in name
Jockey wears yellow socks

Then that's backfitting because of the last 2 criteria and any successful results would be coincidence, if you take the last bits off though then it could be classed as a successful system (pending large sample to confirm success).

stevies i remember u saying on the forum a while back your system had a bad month last year. i think it was september or october i think. how bad.....could u live with that

P-K-W

How would you see the socks?

LOLLLLLL PKW awesome, nah its far from that actually! haha

its got few characteristics that doesnt take me long to work out each day what races ill be focusing on and it seems to prove profitable over last year..

freddie:
march profit loss        -251
thats my losing month... while the average winning month was: 3406.181818
haha

what  stakes
is that too?

My system selected in the 310 York:

BALCARSE NOV   PLACE/win [;)]

G/luck !! 

freddie that is to make 100 each race, risking which ever amount i need to make it

do systems work?

Erm, in the short term, yes.

In the longer term NO, because the 'system' becomes part of the knowledge out there and becomes a consideration in the price. Betting is a long term thing, so consequently all systems do not ultimately work. 
 
The only interesting time is the middle term, after the system has been tweaked enough to make it work and not long enough to become part of the 'general' knowledge in forming a price.

Be very careful when backtesting. It is easy to "find" false profit within a dataset by tweaking variables until you get a nice straight profit line. Best to use a very large dataset, and not to use too many variables/criteria that may co-vary.

stevie ill give you 5 grand
u email me every bet for that day before they run so i know ure not messing
after 6 months u give me back 50% of profits
after 12 months you give me 50% of profits plus my 5grand

take a look at flatracingsystems.com

bluenose - well done on that pick

the 1st reply is the only correct one as you have to be obtaining value otherwise the sums say you will lose

just an interest bluenose, what price did you back BN on the place market?

I'm pretty new to this, but I have a system which I built on historical data.  I took some advice from some of the financial traders who do similar stuff.  They strongly advised two things. Firstly, that I break my historical data into two sets.  Perfect the system on one set and then test it on the other.  If it holds up on the test set you may be onto something.  Secondly, be very very aware that the market changes its characteristics constantly.  If you have something that works expect it to decay after time as the market will adjust to effectively take into account what you are doing.

I've only been going a few months, and so far so good, but I think I already detect an element of decay.

Good luck.

freddiewilliams  13 May 10 15:13   

stevie ill give you 5 grand
u email me every bet for that day before they run so i know ure not messing
after 6 months u give me back 50% of profits
after 12 months you give me 50% of profits plus my 5grand


freddie both me and you know i would NEVER get that 5k and the fact is i wouldnt give my system to anyone... cause WHEN i get to higher money i hope it would lead to maybe increased odds etc
you would use those selections to try and work out my system without giving me a penny.. lol so it isnt worth it to me

Some do but mostly fail by people and discipline

Discipine is the key

I find it very hard to believe that the market adjusts itself automatically over a period of time to turn a winning system into a losing one

"............ then check back results and if it proves profitable does the system really work or is this a fluke?"

trouble is, you can't collect on past results.

systems seem to have a habit of breaking down, can work fine in theory
on a set of past results......but when applied in practice doesn't seem to work out the same.

several reasons why this might be, but basically horses or greyhounds
just don't fit neatly into systems.

As already mentioned on here, that horrible word 'discipline'
and you have to put in the hours to study. No short-cuts....

ST - Like you I was rather sceptical that markets would adjust over time.  The way it was explained to me was: don't under estimate the level of sophistication of some of the software tools and analysis that some guys are capable of doing.  If you've spotted something, chances are that someone, far better equipped than you has also seen it.  There are tools monitoring and catching vast amounts of data off the exchanges all the time.  Apparently these things can smell ant fart at 1,000 paces.

Hard to believe, but perhaps plausible.

thanks lads hopefully it works, only time will tell

Just use the diddles system. Post the winners after the race and charge some mugs £5 a pop for your tips (Which are spoof selections if they loose. Genuine if they win)

FWIW as someone who has played with various systems over the years I think they can work however I believe that you need to be very cautious if using historic data to create them as it is all to easy to add extra rules that result in a profit - particularly rules around starting price, over the timeframe being researched but fall apart very quickly in real time.
Another area to be particularly worried about imo is where the majority of the profits of the testing are produced by one or two big odds winners(+20/1+) with a lot of shorter (5/1 or less)losers.

I've been creating and scrapping systems for years.

Firstly looking at historical data is back fitting and never works because you start to put money on, just when its run is about to end. The invention of staking plans,(increasing your stake during a losing sequence) is meant to keep you in profit but in practice can break you.

Lets remember that a system can be as simple as backing number 2 in every race. The reason people lose on the horses is because they swap from one system to the next. If you look on Adrian Masseys site, you can find plenty of systems that break even or make a very slight loss.

The fact is people lose because they don't stick with one system, but start backing it before the good run starts not when it is about to finish.

be very cautious if using historic data to create them as it is all to easy to add extra rules that result in a profit

Another area to be particularly worried about imo is where the majority of the profits of the testing are produced by one or two big odds winners(+20/1+) with a lot of shorter (5/1 or less)losers.

all the odds are below a certain value so that isnt a problem with big odd winners

as for using historical data, i started off and had 2 criteria checked results and found that it was not a sucessful as would resulted in a loss (small loss) then i changed it slightly and found that it would result in profit
you are saying adding extra rules that result in profit however i only changed 1 of the rules then checked results for the years again and it proved useful.. i hope ive got something i can work on here but i'm beginning to doubt it even though the results tell me it works

morning stevie .whats the highest odds ure layin?

the very highest would be around 7s

i thought ure system was layin favs

you never seen a race where favs odds have been near 7s? and you been here since 2006....

ive been here since the start of betfair stevie......
not many favs start at 7s or above...some but not many


u said yesterday tuesday was a losing day for u.....

how many system bets did u have and how many won and lost?

freddie, by asking these questions and giving you the answers all you are trying to do is work out my system which i am not going to give you enough detail to do so, you tried this before dont forget

ure not so slow



i thought i had it but i dont know how u had a losing day tuesday....back to the drawing board again......


keeps me busy working out systems

the thing is i hardly post selections, and some selections i post on here are not to do with my system at all, they are my own personal choice
so keep trying, there is a number of ways it could of been a losing day yesterday such as 1 bet and a loss, 2 bets 2 losses etc

are there days your system has no selections

freddie yes AND no

work that one out

how many lays had ure system on tuesday?

150

given up yet?

a lot of favs won on tuesday but they where in handicaps.....

ive run it though my program again ...and there is a way u could have had a losing day......

bought a system back in the 80's,from fineform (I thk),last 2 runs:
winner 5 pts.
2nd    2 pts.
3rd    1 pt.
course winner 2 pts distance winner 1 pt (or other way round,well it has been 25 years !! ) add the totals up horse with highest total wins race !! (well sometimes )
MAX BET was last two runs a winner + c/d winner 13 POINTS !!

[;)]

haha bluenose you running that system still?

doubt you would be if your saying it on this thrread

freddie forget it, im not going to tell you ANYTHING towards helping you work it out.. however if you are trying to work it out must make you believe that systems work.. thanks for your answer

of course i believe that systems work......
otherwise i wouldnt spent my time working on systems and searching out new ones...ive got several that work very well....profits every week and profits every month....but im still searching for one that has profits every day....im getting close though.......

freddiewilliams your systems are probably bots that back and lay at certain odds yes? taking maybe 2% here and there where as the one i worked out doesnt involve trading at all lol

working out systems is what i also have been doing recently i find it tough to do but worth it when you find 1 that you think works

i dont use bots...i told u that a few months ago.....dont trade either.....boring as fluck


did u buy that bot you where on here asking about a few months ago?

betbot pro? no mate

i signed up for the free trial, then my internet died for few days and by time i had internet again my trial code period had ended so never even got to try it