v-zero
-
Posts
337 -
Joined
-
Last visited
Posts posted by v-zero
-
-
Re: February NBA Picks Technically not February any more, however the money line (and spread) in the Dallas vs NYK match tonight is a joke: Take Dallas Mavericks over New York Knicks on the money line at 1.79 on Betfair.
-
Re: February NBA Picks Couple of strangely priced games tonight on the money line. Hawks @ 2.56 over Magic Nuggets @ 2.48 over Spurs Both at Betfair.
-
Re: Bairn's Value Away Wins
I'm not going to ask for an insight into your methods, because I wouldn't give you one into mine, I'd just like to say: :clapA few stats... (these only account for the proofed results on this forum and were done quickly so might be a bit off) Number of bets... 143 Number of winners... 47 Strike rate... 32.87% Total staked... £1,242.50 Profit... £585.84 ROI... 47% -
Re: Sophocles' PR System
I believe, Mathematically, that Eskisehirspor was a good pick.Hm... I'm currently backtesting laying the away likely looser - so pretty much the same. But I wouldn't bet on this game (Eskişehirspor - Antalyaspor). Antalyaspor's performance on away games are good (that is bad, but good for system - last 5 games = 4 looses and 1 draw), but Eskisehirspor has 2 looses, 1 draw and only 2 wins - 2 wins with 17th and 18th team (last teams). So although Antalyaspor is not likely to win, Eskisehirspor is not as clear winner. Draw is very possible. So by "points collected at home vs the points collected away'" you mean that it's 7:1 and therfore it's worth playing? Without checking with who those points were earned? -
Re: Software / System Builder I personally use Python for every part of my betting automation. I have it using the Betfair API 24/7 to find and place bets. The automation is top to bottom and requires no interaction on my part - it tweets me when a bet is placed with the details. Basically my point is you can do it all with Python, and it's an easy language to pick up...
-
Re: Kelly criterion vs Flat staking Tackling the problem of uncertain value can be done, and should be done, by reducing to half-kelly or less still. In fact it is suicidal to bet full-Kelly.
-
Re: A.I. Football There is also a free online NN available at zunzun.com, though it is limited on the number of inputs.
-
Re: SAFE MULTIBET system Try not to get so irate about your detractors, it won't help you concentrate on what you are trying to achieve. Personally I cannot draw any kind of conclusion on your system yet, all I know is that mathematically there is no money to be made from selections at prices of less than 1.5, but perhaps that is because the mathematics cannot hope to quantify those last few factors that make the difference between a selection being a 1.3 and a 1.25.... So good luck, I hope you succeed. :cheers
-
Re: When can a system be considered to have "proven" itself ?
It will not produce a result that is significantly different to an Archie value if the number of bets is large and the average strike rate is close to 0.5, but otherwise it will produce a result that is more specific. Moreover the simulation will not give an approximation, but will in fact render exactly the answer to your question.For those of us without a stats background, would you be good enough to clarify whether the "monte carlo simulation" generates a different result to that of the "Archie" probability value? And congrats on the results in your "close matches" thread. A 17% yield with your high volume of action and excellent strike rate is very impressive; and no surprise to see that it is already down to a 5% chance that you've simply been dancing with La Divina Fortuna. -
Re: When can a system be considered to have "proven" itself ? A Chi-sqaured test (pronounced kai, by the way) is somewhat suited to the task, but not perfectly so. A more useful test is to put a bound on how likely it would be to observe your results from your set of bets under the assumption that the odds received are fair. For instance, in my close games thread there is around a 5% chance that my bets would have achieved that yield or better over that many bets by mere chance. The easiest way to do this with accuracy is to do something called a monte carlo simulation. The ROI and number of bets are hugely important to this calculation, as are the odds at which the bets are placed.
-
Re: SAFE MULTIBET system
It is quite easy to calculate the likelihood of reaching a positive roi given a set of odds, and assuming they are fair.um,I am sorry Maybe isnt entering into debate anymore cos I am most intrigued to know why he doesnt also lay those "dangerous" odds-on shots his system identifies eg Liverpool(v Norwich).Surely they would be better value? Or at least post them. Also,Datapunter,you appear to have answered a question I asked on another thread a few weeks ago re.when exactly a system is said to be "proven". Could you please confirm that its proven after 500 games? Is this something that is generally agreed upon? If so,presumably that means that,at the present rate,we will have to wait til about 2016 to find out if this works before we all lump on!? I spose the bookie offering the best prices on the double that day will get a big surprise:)!! also,in november 2016,what roi will the bets have had to have made before its deemed proven? would 1% be enough or is there a minimum roi required after 500 games? How many people in the history of PL have actually achieved this 500 game proof and do we know what happened afterwards ie did the prices move on the week after the system was proven? If not,then does anyone have an explanation why not? many thanks,and good luck Maybe,see you in 2016(or earlier if you can step up the selections:)) -
Re: Best Statistical Analysis software to use
I have been considering posting a piece of here about how to, at least in terms of the mathematics, create a ratings system that intrinsically uses either a logistic or poisson approach. I'm not willing to give away my more complex methods, but I'm happy to help people get on their way with modelling. A precursor to this, though, is an ability to program. I think Python is the perfect language for this type of use, as it is quick and easy to hack bits onto. All of my models, and my bots, use Python exclusively.Thanks Plinge for the unprompted plug. I agree with the confusing use of the term "data mining" on here. I've never really heard it used in this way, the problem is more one of "backfitting" as you suggest but also "overfitting". It would appear BTL has already had a stab at SoccerSystemBuilder: To answer BTLs question, throughout my career I have used SPSS (Statistics & the "Data Mining" Workbench Modeler), SAS, KXEN & R. R is probably the only option here isn't it as it is free to individuals for personal use. I can't think of any good books on the topic, research papers are probably your best bet, as v-zero suggests. But if stats aint your thing then the research papers might blow your mind a little. -
Re: Best Statistical Analysis software to use
Fair enough. Apologies to the OP if this is getting off track but out of curiosity how do you define "data mining"? It seems to be in common currency here' date=' but isn't a term used in my neck of the woods in Australia. I've noticed that it often carries a pejorative implication, akin to what I'm used to calling "backfitting", ie. cherry picking criteria from a database to identify historically profitable angles with no logical or statistical validity. I certainly don't believe a resource like SSB has to be used that way, and I gather from you've written that you agree.[/quote'] In this instance, the key difference between statistical analysis and data mining is that data mining represents the search for historically favourable bets under simple circumstances, for instance recent form, or odds categories, whereas statistical analysis attempts to find what is significant to predicting future events by making assumptions and then testing them against genuine occurrences. What may seem significant in a data mining situation, will often turn out to be nothing more than random error, and not significant statistically. Equally, when it turns out something is significant, there is ample space for that opportunity to disappear, as it is not a case of some clever piece of modelling beating the bookies, but is rather just a quirk of their odds that if they notice they can make vanish - and bookies are very good at noticing long-term-profit makers. Data mining can be useful, it can find profit, but it will never tell you why something works. Statistical analysis, and modelling, are so much more powerful because they offer up insight in a very genuine way, and they are not subject to the abuse of the bookies. -
Re: Best Statistical Analysis software to use
Don't get me wrong, data mining software like SSB can be very useful, but the question is about statistical analysis, which isn't something SSB offers.I'm going to give SoccerSystemBuilder (the product linked to in my above post) a plug, so first let me aver that I'm not affiliated with it any way, except for the fact that since I've been a customer, the client support and openness to feedback from Matt, the designer, has been incentive enough to don a spruiker's hat on his behalf. I'm new to soccer punting, but when doing some work for a professional punter a few years ago, I had the opportunity to do extensive database research on Australian horse racing. The products I was using literally cost thousands of dollars and when used judiciously provided knowledge that would be impossible to glean without access to such resources. SSB might not be quite as sophisticated, but I have found it amazing for its price (and I think Matt might still be offering a free trial period). There may be no ultimate "short cut" to successful gambling on any sport, but as a newbie to soccer, I have absolutely no doubt that SSB has already given me insights that would have taken me years to glean without it, if I ever discovered them at all. -
Re: Best Statistical Analysis software to use There is, to be honest, no piece of software that will offer you any kind of worthwhile short cut. Find and read all of the PHD/Research work you can, get to grips with a nice programming language like Python or Java, and do a lot of reading on statistical models. There is no short cut to knowledge and understanding.
-
Re: UNDER 2.5 goals using previous results
How many bets, roughly, do you make a season, and at what stakes?I do take your point about not posting a yield figure and technically i suppose i should if following the rules of the Forum ..if any mod insist's then i'll gladly try to figure it out and keep it up to date but basically the only figures that matter's [imo] are the one's posted when time is called on each season . I've ran the system for 5 seasons now and had a starting bank of 100pts each time , the profit /loss for each is as follows - 06/07 - +105.96pts 07/08 - +375.18pts 08/09 - +184.41pts 09/10 - 100pts LOSS 10/11 - +115.11pts There are usually 1-3 bets each weekend . On the question of staking , it is 'generally ' accepted that around 3-5% of one's betting bank should be the the amount staked on each wager but in the case of this system where the strike-rate is [so far] fairly high i see no reason to lower the stakes . Surely your strike rate combined with the odds on offer and how you stake [ eg running doubles / trebles ] should dictate how much each individual should invest in each bet ? If you have a 'system' or strategy which is continuing to show profit why not maximise you returns by increasing your stakes to a level you think your bank could safely handle a losing run ? Appreciate your comments vzero :ok please feel free to comment further , anyone else hold an opinion about this subject ?? -
Re: UNDER 2.5 goals using previous results
I do, and I think that combined with your lack of a yield posted may mislead followers as to the success of this system. To use my own system as an example, I use a one point stake, and having started with 100 am up to 115.4, with a yield of 16% after 96 bets. To many, in comparison to your numbers this year, that may look poor, but were I to risk 20 points per bet that would be a profit of 308 points, but the yield would be the same. However, it is generally accepted that bets shouldn't be more than about 5% of your bank, with 1-3% being more reasonable.I take it you think that each initial bet of 20pts [20% 0f starting bank ] is too high ?? -
Re: UNDER 2.5 goals using previous results Why such unrealistic stakes considering the starting bank?
-
Re: OVER 2,5 GOALS With Statistics
Spot on. I have done a lot of work on predicting totals, and the fact is that in all the leagues I have considered with soccer there is next to zero correlation between past performance and future performance in terms of totals. What's more, the totals predictions from a Poisson distribution are wrong, whereas the 1X2 predictions from the same distribution can be very good. Unlike soccer, though, games such as Basketball and American Football have much more predictable totals.I'm just seeing this now...I tried something like this once, and it didn't win. There's simply no correlation between recent totals and a given game's totals, and you end up playing too many games to be successful. Also, a lot of leagues are hit and miss...average goals look normal, but when you see the games you see a wide range between 4-3 games and 1-0 games (Spain Segunda and Italy B are both like this. What I'd found was that in general, leagues in which there are few clean sheets are better...games between lower-table teams are better than top teams...and games with high odds are better because goals are generally random events. And the best leagues change from year to year...the only consistently high-scoring leagues are in Holland, the Bundesliga and lower leagues in Norway and Sweden. The best this past year at turning up high-odds winners has been, strangely, France Ligue 1. That never used to be the case. -
Re: Draws for profit...Pt. II
That's just small smaple error. Leagues will in general tend towards the mid to high twenties for the full season length.Hmmm!! I thought about doing something like this myself, although I was going to target the French and Italian premier leagues. The French league has 38.2% draws and the Italian has 32.9% draws. Even the Greek premier, Argentina premier, Austrian premier, French league 2 and German league 3 all have mid 30% draws. Soccerstats.com has all the info you could need. -
Re: NFL ATS Comp 2011/12 *Week 6 selections* Just plain forgot about this... I posted them on my blog: http://spredictions.wordpress.com/2011/10/12/nfl-week-6-predictions/ But I can understand if you don't feel that these can be included. They're also posted to ThePredictionTracker (though they have differing lines to the ones used for this thread) in case you want proof it was posted before the off. :unsure
-
Re: Is the market always right?
It is obvious that over/under markets odds derive from match odds. These markets have the same roots (the same model). I assume its poisson' date=' but it could be something more sophisticated. Anyway, no chances for arbitrage or hedging.[/quote'] They probably have their roots in the same model, but it won't be poisson. -
Re: Is the market always right?
If that were the case then making money on football would be easy, as a either the match odds, or over/unders odds would have to be wrong. Poisson modelling is not very good, in all honesty, it is only truly useful for creating in-play models.My own explanation is that in financials (equites' date=' CFDs, etc.) risk is more manageble. But in soccer one goal ruins the day and you can do nothing about that. Various hedging techniques simply does not work due to efficiency of the markets ([b']and the fact that all the odds for all related markets in any specific match come from one poisson model). -
Re: NFL ATS Comp 2011/12 *Week 5 selections* Date & Time Favorite Spread Underdog 10/9 1:00 ET @ Indianapolis -2.5 Kansas City 10/9 1:00 ET @ Minnesota -2.5 Arizona 10/9 1:00 ET Philadelphia -3 @ Buffalo 10/9 1:00 ET @ Houston -6 Oakland 10/9 1:00 ET New Orleans -6.5 @ Carolina 10/9 1:00 ET @ Jacksonville -2 Cincinnati 10/9 1:00 ET @ Pittsburgh -3 Tennessee 10/9 1:00 ET @ NY Giants -9.5 Seattle 10/9 4:05 ET @ San Francisco -3 Tampa Bay 10/9 4:15 ET @ New England -9 NY Jets 10/9 4:15 ET San Diego -4 @ Denver 10/9 8:30 ET Green Bay -6 @ Atlanta Monday Night Football Point Spread 10/10 8:35 ET @ Detroit -5 Chicago
February NBA Picks
in NBA & Basketball Predictions
Posted
Re: February NBA Picks There is value out there tonight, but nothing that is very likely to come off. However, as far as value goes, these are the best: Atlanta Hawks over Miami Heat at 10.0 on Betfair. New Jersey Nets over LA Clippers at 3.8 on Betfair.