psilocybin Posted August 3, 2005 Share Posted August 3, 2005 Hi folks! I have a system that is able to group together 136 games from Serie B since 1993 (12 seasons) where the percentage of draws is 56%, equivalent to 1.79 in odds. The Home-Draw-Away results are as follows: 32-76-28. I'm curious to know, based on the information given, if punters think that a similar percentage can be expected over the next few seasons? All responses appreciated! Quote Link to comment Share on other sites More sharing options...
slapdash Posted August 3, 2005 Share Posted August 3, 2005 Re: A Draw Trend in Serie B Did you come up with system by looking at the data? That can be a dangerous thing to do. The problem is that if you look at the data and work hard enough you're almost bound to be able to cook up a rule that picks out an unusually large number of draws for that particular set of data, especially if you allow the rule to be complicated. And because the rule is designed for that set of data, there's less reason to think it will also work for a different set. A better way would be to divide the data into two halves, use one half to design the system, and then use the other half to test it. Quote Link to comment Share on other sites More sharing options...
MattyNeth Posted August 3, 2005 Share Posted August 3, 2005 Re: A Draw Trend in Serie B Serie B does have more than it's fair share of draws, BUT I am sure the odds reflect that. Pretty sure most games are priced up at well under 2/1 for each game for the draw Quote Link to comment Share on other sites More sharing options...
spreadman Posted August 3, 2005 Share Posted August 3, 2005 Re: A Draw Trend in Serie B Agree with Mr. S,..........It's called 'CHERRY PICKING',....looking for trends once the results are known. And as for there being a high per centage of draws....if this is the case you can bet your last lire, or even, euro, that the bookies will have noticed, and as Matty says, the odds available will reflect this fact. I'm off now to watch the fireworks in the Newcastle match....a bookings buy , methinks. :loon Quote Link to comment Share on other sites More sharing options...
psilocybin Posted August 3, 2005 Author Share Posted August 3, 2005 Re: A Draw Trend in Serie B Thanks slapdash! I used data produced from looking at previous results in Serie B. I do agree that it is surely likely that you will be able to dig out a high percentage of anything from so many results. However, it think it's more correct to say that the results(data) are within the rule and future results would fall within the rule. I'm simply curious to know if punters think that the trend will continue? If not, do punters think that a profit would be likely? Quote Link to comment Share on other sites More sharing options...
AussieRules Posted August 4, 2005 Share Posted August 4, 2005 Re: A Draw Trend in Serie B psilo best thing you can do is to work on a system with a random selection of the data (so-called sample) and use this to build you model / method. Then appkly your method to the remaining part of your original populoation and if the system holds up then you have the beginnings of a legitimate system. While the sample should technically be a random selection from all your data, you might want to consider finding a system from the data from 1993-2000 and then applying that system on the seasons 2001-2004 and see how the numbers compare. This will give a little more solid ground than just cherry picking. If you need Italian odds I can probably help you over the last few years, but I don't have a lot before that. Hope this helps. Oz Quote Link to comment Share on other sites More sharing options...
psilocybin Posted August 4, 2005 Author Share Posted August 4, 2005 Re: A Draw Trend in Serie B Thanks for that AussieRules! You really believe that if I take a random sample of data and use that to build the system, and then if I apply that to the whole, and if it holds true, then I have a worthy system? Very interesting! Did you think of this yourself or is this standard practice? How much random data are we talking about. Too little may be meaningless. Too much may spoil the point. 10-20% ok? I appreciate your point about comparing past data to recent data. Running a random set of data through the model may also be worthwhile. I'll put your words into practice and see what happens. Oh, well I be needing Java and MySql for this experiment? :D Cheers! Quote Link to comment Share on other sites More sharing options...
AJ Posted August 4, 2005 Share Posted August 4, 2005 Re: A Draw Trend in Serie B Serie B is famous for the high number of draws, I'd advise taking your system and paper testing it without real money for the first half of the season, if it seems to be working out ok, start placing real bets after Xmas. I assume you are doing this with level stakes ?? Quote Link to comment Share on other sites More sharing options...
psilocybin Posted August 4, 2005 Author Share Posted August 4, 2005 Re: A Draw Trend in Serie B Thanks for the reply Foolsgold99! I realise there is a lot of draws in Serie B. Paper testing would be a good idea, but I can't resist. I was thinking of using 5% of bank. Quote Link to comment Share on other sites More sharing options...
AussieRules Posted August 5, 2005 Share Posted August 5, 2005 Re: A Draw Trend in Serie B Psilo What I suggested is standard practice for building statistical models. You are correct about the correct amount of data to use in the model (too little vs. too much). I would advise using about 70% of your data to build the model and about 30% to test it (depending on how much data you have). For example with my Dow Jones system I used about 700-800 data points to build the model and about 350 to test it (and it held up). I then paper-tested for about 100 data points, and now it is in full on production in another thread. You will not need MySQL to do what I have suggested. A well worked Excel spreadsheet will suffice and/or Access if you have the know how. If you want EXACT odds for the last three years of SerieB I am pretty I have that. I will throw this out there, and I understand everyone's proprietary feelings about their own systems :) (me too!) But if you are not sure you know how to do it, I am willing to help you build the excel sheet that will allow you master your possible system. If you are interested drop me a line (ozcapper at hot mail) - I am sure you can work that one out. I know I am a "newbie" on this forum, but I am well known on some other forums (see the statfox forum under the same username). If this helps you believe I am not trying to just pry your system away from you. :) Honest guv' If nothing else you can see I am willing to give out the info with my picks here. Have a good weekend. Oz Quote Link to comment Share on other sites More sharing options...
AussieRules Posted August 5, 2005 Share Posted August 5, 2005 Re: A Draw Trend in Serie B Another thing I forgot to add. Level stakes is a good method. If it is an untested system you may want to cut back to 2-3% until you are sure it works, and set yourself a "bankrupt" level too where you cut the system dead should you ever hit that point. For example, if you were starting with $1000, you would cut it short when you hit $250 - that way you can cut your losses ina very literal way. If the system is analyzed correctly you should be able to use a kelly-betting type system that can maximize your profits too, but that will come with more development of the system. Hope this all helps, I know it is a lot. Oz Quote Link to comment Share on other sites More sharing options...
psilocybin Posted August 8, 2005 Author Share Posted August 8, 2005 Re: A Draw Trend in Serie B Thanks for that Oz. Interesting stuff! I'm curious about the figures 70-30. Did you come up with this as a rough estimate? Do you have any links to statistical models that refer to this method of testing? So, if the model holds true for the test data, I can say it is viable? And future predictions under the model should match the model? I'll have a go at creating a revised model based on what you said, and if it holds, I'll post the tips here! Will only be about 10-12 games this season. Also, thanks for the advice on banking. Quote Link to comment Share on other sites More sharing options...
AussieRules Posted August 8, 2005 Share Posted August 8, 2005 Re: A Draw Trend in Serie B Psilo Those are rule-of-thumb percentages and not set in stone, there are number of variables that come into play as to exactly how much data should be used in the model build/test, but the 70-30 rule will put you in the right neighbourhood. If doing this means revising your system then it means we just saved you some sheckels mate! :) Good luck with the system Oz Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.