Database, linemaking etc.

Search

New member
Joined
Sep 21, 2002
Messages
2,954
Tokens
I ve come to the long and hard conclusion that a database for my main betting interest, world soccer (football), is long overdue. I ve always found it very important to go through several quantitative factors before placing a bet, together with any qualitative appreciation and "gut feeling", but I ve never really had enough time to devote to built a proper database. Nor do I have now, lol. But at least I 've set myself this side project because I know despite the long hours I ll have to, bit by bit, devote to it, it's going to be a time and money saver, a huge one in the long run. And it's the only way I can statistically test several empirical conlusions I have come upon during my betting days, and use these as a basis for a statistical prediction model I have in mind.

At the moment, besides having brushed up my university level statistics, and gotten a few more advanced books to be looking at during the next few months, god willing, I am also on page 257 of an 800 page book on database design and management, not of any particular one just a general one (although I ve figured out that sql is the "it" thing when it comes to databases) I will also be brushing up my programming skills because it's really not much use having a database if you got to input everything by hand.

Anyone of you guys use a database prediction model for betting, I d appreciate it if you d give me the heads up on the dos and don'ts. Or share your general experience with such methods.

Bol. Jack.
 

Oh boy!
Joined
Mar 21, 2004
Messages
38,362
Tokens
I'm going to do the same thing for NBA Jack. Except I'm going to buy a database if I can find one cheaply enough. Then I'll apply statistical analysis to trends to see if they are viable.

I went through a losing streak last week and decided it's time to stop experimenting with my money by betting my hunches. I'm glad I went through that learning process as it allows me to find know what to look for. I just need to go out and look for it quantitatively.

Let me know how you do.
 

New member
Joined
Sep 21, 2002
Messages
2,954
Tokens
It's good to see we 've drawn similar conclusions here quantum. If you feel like it, drop me a line too at jackdee@mail.gr. I just wish guys that are been there done that could give us some advice here, not their profitable senarios obviously but how one goes about putting together and using such a beast.

When you say you are going to buy one, I suppose you mean some proper access implementation relational model, not just the tabulation in an excel or .csv file. That could be well worth it if it is reasonably priced, I haven't found anything similar to buy in soccer, I 've already bought results and odds in an excel tabulation which could serve as the raw data. My problem with a premade database is that you are not the one setting the schema (he,he, here I am after two hundred or so pages getting all jargony...), the design for it, so it's like buying a house, if you design it yourself you are going to fine tune it exactly to your needs, if you buy it as is, it migh suit your needs just fine or it might not be exactly what you have in mind. And from what I have been reading it makes a great difference of how you define and relate the various entities to each other.

"I'm glad I went through that learning process as it allows me to find know what to look for. I just need to go out and look for it quantitatively."

Same here.

I ll let you know how it's going but like i said it's going to take some time for me, trying to squize in the few hours I can every week.

That's why I am asking the rest of the guys here for some pointers such the software they use, if they have a separate statistics package (probably not), how they automate the update process etc. etc.
 

Oh boy!
Joined
Mar 21, 2004
Messages
38,362
Tokens
I used to teach a class in using the database functions in Excel so I'll probably just go that route. I've used Access but I'm more comfortable seeing all the information on spreadsheets rather than having it hidden away in many tables.

I'll send you an email from my Bill Green email account some time on Saturday. Get it? Green bills, Bill Green. I use it to fill out forms that require an address. I put my address down as 20 E. Greenway Road. LOL!
 

acw

Member
Joined
Sep 21, 2004
Messages
1,313
Tokens
Jack,

To encourage you, yes, a database has given me some edges.
To discourage you, you will make a lot more being able to lay 20 cent lines to some squares, which you can lay off at Even Money with line shopping than I have managed to make betting myself.
 

New member
Joined
Oct 16, 2005
Messages
528
Tokens
I've gone almost entirely the data analysis route since starting betting.

There's not much that I can really tell you about it...I've used combinations of C++/XML/SQL/PHP/Excel, but whatever you're most comfortable with is fine.

I haven't used any outside statistics packages, but I've read a couple of stat books and papers on binomial distributions. If it's been awhile, I'd definitely recommend checking one out.

If you're like me, you'll want to see this for yourself...but there are very few things that one can exploit from the data that are profitable over the long term. I certainly haven't looked at everything, but I've tried a LOT of different things that did not work. I would highly recommend it as a source of knowledge which can be put to very good use, but there's a very small chance of directly finding any exploitable situation.

All of that being said, there's no better way to do it. Let me know if you have specific questions, and I'll answer what I can.
 

LA Clippers Junkie
Joined
May 14, 2005
Messages
11,323
Tokens
acw said:
Jack,

To encourage you, yes, a database has given me some edges.
To discourage you, you will make a lot more being able to lay 20 cent lines to some squares, which you can lay off at Even Money with line shopping than I have managed to make betting myself.

Great post.
 

Oh boy!
Joined
Mar 21, 2004
Messages
38,362
Tokens
One thing I would like to add is that when you are using a database, especially one that covers many years, it should be used as a guide, not an automatic betting calculator. What I mean by this is that past numbers are not always accurate for the short-term. Things such as long road trips, injuries and other intangibles must be taken into consideration for these will have a large influence on the line.
 

Oh boy!
Joined
Mar 21, 2004
Messages
38,362
Tokens
Salain said:
I've gone almost entirely the data analysis route since starting betting.

There's not much that I can really tell you about it...I've used combinations of C++/XML/SQL/PHP/Excel, but whatever you're most comfortable with is fine.

I haven't used any outside statistics packages, but I've read a couple of stat books and papers on binomial distributions. If it's been awhile, I'd definitely recommend checking one out.

If you're like me, you'll want to see this for yourself...but there are very few things that one can exploit from the data that are profitable over the long term. I certainly haven't looked at everything, but I've tried a LOT of different things that did not work. I would highly recommend it as a source of knowledge which can be put to very good use, but there's a very small chance of directly finding any exploitable situation.

All of that being said, there's no better way to do it. Let me know if you have specific questions, and I'll answer what I can.

Salain:

You mentioned in another thread that you may have found a system that hit 70% of the time. I wouldn't expect you to tell us what it is. But can you tell us what method you used to find it? Have you found any significant systems using a database?

I'm just trying to find the best way to use existing data to find a significant trend.

Here's the thread I was talking about-
http://forum.therx.com/showthread.php?t=357615
 

New member
Joined
Sep 21, 2002
Messages
2,954
Tokens
@acw
"To encourage you, yes, a database has given me some edges. To discourage you, you will make a lot more being able to lay 20 cent lines to some squares, which you can lay off at Even Money with line shopping than I have managed to make betting myself."

yeah but I d have to set up a small shop to do, because what with the competition at the exchanges, although I have not tried this, it's hard to see how I could take them on at a 20 cent line, probably at higher odds taking advantage of the favourite longshot bias would be easier but the risk would be greater too. Acw, do you still have an active database in soccer? I was thinking along the lines of something including all the basics, team line ups, budgets, leagues, ages, playing distances, goas etc. etc. plus such things as shots on goals, ball possesion, corners, fouls, even localised graphs of shots on goals there are some good resources for these, refs etc. and then maybe adding in few other factors (if I can, which for most I am sure I can find) such as weather, wind, humidity, attendances etc. etc. You got my email get in touch if you want to talk about the particulars.

@Salain

Thanks for your input buddy, much appreciated.

".I've used combinations of C++/XML/SQL/PHP/Excel, but whatever you're most comfortable with is fine."
Well for the time being excel would be it, but for reasons other than my betting interests I should soon be profficient (well, at least good enough) in the rest too (I ve just resumed a cs major from a 3 year hiatus I d gone to work)

"I haven't used any outside statistics packages, but I've read a couple of stat books and papers on binomial distributions. If it's been awhile, I'd definitely recommend checking one out."

I 've brushed up on the basics, and I ve taken a few relevant statistical papers about forecasting (wrt gambling and financial markets too) and a couple of more advanced books.

"If you're like me, you'll want to see this for yourself...but there are very few things that one can exploit from the data that are profitable over the long term"

Hmm, it's the double sword...and yes I am like you, I like to look things for myself. I suppose you are refering to baseball though or basketball judging by your locale, right? I know for a fact that there are some people who have produced profitable prediction models in football who manage to do consistently well.

"I certainly haven't looked at everything, but I've tried a LOT of different things that did not work. I would highly recommend it as a source of knowledge which can be put to very good use, but there's a very small chance of directly finding any exploitable situation."

A can sense what you mean when you say a LOT, it really must mean a lot, I am at the moment feeling a bit like my hands are tied, I d prefer trying different things and failing, instead o havin to wonder what if and not being able to put a hypothesis to the test, which I am sure has been your motivation too. But what's the difference between a source of knowledge as you say being put to a good use, and directly finding an exploitable situation, it's hard for me to grasp the distinction...

"All of that being said, there's no better way to do it. Let me know if you have specific questions, and I'll answer what I can."

Thanks a lot, like I said much appreciated, won't have anything done the coming month or so, but I hope there'll be some steady progress.

@quantumleap

"I'll send you an email from my Bill Green email account some time on Saturday. Get it? Green bills, Bill Green. I use it to fill out forms that require an address. I put my address down as 20 E. Greenway Road. LOL!"

I ll be waiting for the cash then, he, he..

"What I mean by this is that past numbers are not always accurate for the short-term. Things such as long road trips, injuries and other intangibles must be taken into consideration for these will have a large influence on the line."

I was hoping to factor all that in too, first of all calibrate the extent to which a long trip say effects a result, or the relative weight of injuries on the line (this of course is much more qualitative, but I can't see why if one has some basic end season player stats and the games they missed during the year why they wont be able to correlate the two and arrive at some basic datum for the influnce), if one can find some back data.

"What I mean by this is that past numbers are not always accurate for the short-term."

Sure, I see what you mean.


Jack.
 

New member
Joined
Sep 21, 2002
Messages
2,954
Tokens
^ (wil, and other mods hope it's not too intrusive bumping this up for visibility.)
 

New member
Joined
Oct 16, 2005
Messages
528
Tokens
quantumleap said:
Salain:

You mentioned in another thread that you may have found a system that hit 70% of the time. I wouldn't expect you to tell us what it is. But can you tell us what method you used to find it? Have you found any significant systems using a database?

I'm just trying to find the best way to use existing data to find a significant trend.

Here's the thread I was talking about-
http://forum.therx.com/showthread.php?t=357615

I mentioned that I had a system that went 70-30 out of the gate, and that if you try many many ideas out you will get some that look extremely good for extended periods of time that are actually crap. The longer you look, the more likely you are to find things that look better for longer. That particular system has been profitable since then, but on a much slower pace.

I have found only a very small number of "situational" systems that are marginally profitable. I've also created some pretty sophisticated power ranking systems, but they are more complicated to describe. The simplest way to look for a system is to count the wins and losses of only the teams that fit a set of situational "if"s. A slightly more advanced version of this is to create a numeric rating system based on factor's of a team's performance...so a team gets +1 if it wins by 10 points, +2 if it wins by 20, and reset to 0 if it losses...then you bet (ie count the game in your results) if the team's rating is above a number. Can rework this in many ways, and it is well suited for combining multiple systems.

But I have 10 seasons of data for most college and pro sports (besides baseball which I am VERY laboriously collecting info for) and I have found that there are almost no unaccounted for situations that can be found from the majority of the historical statistical data - Dates, teams, closing lines for sides and totals, and final scores. Use the info that you gain to understand what is already being accounted for, and it will do 2 things for you. First, you'll know where you can stop looking. Second, you can look for systems that would only be valid if these things are true.

I know that is vague, but I can't even remember everything that I've tried. I have done almost nothing but this in my free time since October, and I have only within the last several months really found things that show promise when applied. I have made almost %400 on my starting BR, if that is encouraging to you.

Wish you the best of luck with it!
 

acw

Member
Joined
Sep 21, 2004
Messages
1,313
Tokens
JackDee said:
@acwyeah but I d have to set up a small shop to do, because what with the competition at the exchanges
The exchanges are so much fvcked that I dare to say that BetFair has less than 10 users that are at the 60% discount level now for three years, bet straight (so no trading) and manage to make more than they pay in commission. BetFair once indicated that over 95% of their users are losers, so one step further will mean that of that 5% winners many will be traders or if they are still straight winners, then they will still be "working" more for BetFair than for their own benefit.

In regards to all those factors that you mentioned, yes I do still actively keep a database with them, though I must admit that when I started doing some analyses three/four years ago my enthusiasm dropped to almost zero when I discovered that the bookies (or the smart clients that they follow) already know the value of the individual player strength. This does not mean that it is useless information. In fact very useful, but it will only help, if you get team news earlier than the other smart guys in order to act accordingly. I discovered that the betting-in-running is much more lucrative and less stressful than having your tongue glued to the computer screen in order to attack the market at the right moment.

I mentioned this before: The lunar visibility is a very important factor!
 

New member
Joined
Sep 21, 2002
Messages
2,954
Tokens
"BetFair once indicated that over 95% of their users are losers" Ηοw did they indicate that? Some financial report of theirs? I see what you mean, and I agree. I am still wondering how they are maintaining their liquidity, do you think smaller shops are laying off bets there? But I still do not comprehend your initial point that I could be laying high margin bets to squares to then bet them at decent prices elsewhere without setting up a small shop like I said.

"In fact very useful, but it will only help, if you get team news earlier than the other smart guys in order to act accordingly. I discovered that the betting-in-running is much more lucrative and less stressful than having your tongue glued to the computer screen in order to attack the market at the right moment."

I have some, a lot data, of betting in running at betfair I hope I can spot the same market vulnerabilities you have my friend. ; )

Btw, acw can you please email me? I promise I will abuse your mail with my $$$LOCK$$GOYs every single damn hour!
 

Oh boy!
Joined
Mar 21, 2004
Messages
38,362
Tokens
Salain said:
The simplest way to look for a system is to count the wins and losses of only the teams that fit a set of situational "if"s.

Wish you the best of luck with it!

I like this concept. That being of having a system where only certain teams fit, rather than having every team in the league fit into it.

Thanks for your time and effort.
 

New member
Joined
Sep 21, 2002
Messages
2,954
Tokens
thanks my friend, I already have joe's data, but thanks for the heads up anyways, if you with to, please email me at the mail I am mentioning here.
 

acw

Member
Joined
Sep 21, 2004
Messages
1,313
Tokens
JackDee said:
"BetFair once indicated that over 95% of their users are losers" Ηοw did they indicate that? Some financial report of theirs?
No,

It was mentioned in these now dead links:
http://www.news.com.au/common/story_page/0,4057,11229314%5E3462,00.html
http://dailytelegraph.news.com.au/story.jsp?sectionid=1258&storyid=2164368
Maybe someone can still dig it up?

JackDee said:
I see what you mean, and I agree. I am still wondering how they are maintaining their liquidity, do you think smaller shops are laying off bets there?
Not this (my) little shop. I should have done though the last two months, as my squares are beating me big time!!!

JackDee said:
But I still do not comprehend your initial point that I could be laying high margin bets to squares to then bet them at decent prices elsewhere without setting up a small shop like I said.
Let me try to explain a little better. Assume there is one particular football match every week and always team A gives 1/2 goal to team B. Doing line shopping it turns out that the best prices on team A and B (at decent shops) are Even Money both sides. A smart (local) bookie will now post team A and team B at 1.90. Assuming that both sides have 50% chance of winning, this will give the bookie an edge of 5% no matter what his clients bet. Now instead of taking bets you discovered (through your database) that team A and B do not each have 50% chance, but team A 51% and team B 49%. So (together with line shopping) you bet on team A at Even Money. You now have an edge of 2%. The point I am trying to make is that this 2% will never get more than the 5% a dumb bookie has by laying bets to square customers, so though you may spend hours of hard work in trying to bet as good as possible with different books, you will still earn less than having some honourable square "friends" that bet with you. I am simply trying to make clear that you will not (or at least I have not) find that much value in a database that you may hope for.


JackDee said:
"In fact very useful, but it will only help, if you get team news earlier than the other smart guys in order to act accordingly. I discovered that the betting-in-running is much more lucrative and less stressful than having your tongue glued to the computer screen in order to attack the market at the right moment."

I have some, a lot data, of betting in running at betfair I hope I can spot the same market vulnerabilities you have my friend. ; )

Btw, acw can you please email me? I promise I will abuse your mail with my $$$LOCK$$GOYs every single damn hour!
You may private message me on one of the other forums.
 

New member
Joined
Oct 1, 2004
Messages
1,316
Tokens
JackDee: It's always nice to learn or try to learn something new everyday.

"Keep up the good work".:103631605

-----------------------------------------

I personally not only like to see something work, but how it was created, etc. is important. Dig into it, best way.
 

New member
Joined
Feb 21, 2002
Messages
881
Tokens
Jack, I read elsewhere, that only 2% of Betfair accounts were showing a profit. And only 0.8% were making 15k+ (british pounds) a year.A good portion
of those are probably linked to bookmakers.
 

Forum statistics

Threads
1,106,766
Messages
13,438,770
Members
99,337
Latest member
hbs_solutions
The RX is the sports betting industry's leading information portal for bonuses, picks, and sportsbook reviews. Find the best deals offered by a sportsbook in your state and browse our free picks section.FacebookTwitterInstagramContact Usforum@therx.com