BetLabs systems.. Anyone ever have success with it?

Has anyone ever used Sports Insights Bet Labs??? I understand the Wizard himself despises betting systems, although its seems hard to argue some systems on that site. I have found a few that were thousands of bets, and had profitable winning percentages at a 95% confidence interval. Not to mention profitability and a smooth winning trend upwards for over 10 years. How can we say that is garbage???

Did you actually pay for those systems? It is a pay site.

I am not an expert.

However, how can you trust the data from a website that wants to sell you a system? If the math says the game is comprised of -ev bets, or depends on past results to identify trends in games with independent events, or is subject to past-posting/doctoring results or other data manipulation, (generalities that would or would not apply to every game), then it seems likely it's an elaborate sales job.

If you gave a specific example of what game/what result seems credible, I think it would be easier to discuss it.

I have trialed and paid for the software.

Quote: beachbumbabs

I am not an expert.

However, how can you trust the data from a website that wants to sell you a system? If the math says the game is comprised of -ev bets, or depends on past results to identify trends in games with independent events, or is subject to past-posting/doctoring results or other data manipulation, (generalities that would or would not apply to every game), then it seems likely it's an elaborate sales job.

If you gave a specific example of what game/what result seems credible, I think it would be easier to discuss it.

By no means am I defending the software, but the results are very transparent. Every past game, date it was played, result and odds are available to see.

Quote: Chodempole

By no means am I defending the software, but the results are very transparent. Every past game, date it was played, result and odds are available to see.

So it's sports betting? How do you know they aren't selectively posting wins after betting both sides? Or taking results and retroactively claiming they bet the correct side? It's not that hard to backdate data.

Quote: beachbumbabs

So it's sports betting? How do you know they aren't selectively posting wins after betting both sides? Or taking results and retroactively claiming they bet the correct side? It's not that hard to backdate data.

Valid point. They also post picks before the games. I get what you're saying, but too many people use the software and I have yet to hear anyone suspect something of that nature.

Quote: Chodempole

Valid point. They also post picks before the games. I get what you're saying, but too many people use the software and I have yet to hear anyone suspect something of that nature.

I know someone that’s a consistent winner (15+ years) at college basketball and college football. High 50s% hit rate. I think it’s possible, though it takes a lot of data and analysis. I have never heard of this site or software. I’m skeptical anyone can beat one of the big professional sports (NFL, NBA, MLB or NHL) in the long run, except maybe for second half bets.

If you're paying for sports betting advice, you're already a loser (in betting terms). Almost impossible to win long term at -110, and adding in the monthly cost of someones random picks essentially raises your vig to a level that can't be overcome.

It's a losing proposition, 100% of the time.

We can't be hypocritical though. Sports betting professionals do exist, so how can we say data mining software wouldn't be achieving some sort of handicapping that a human does successfully?

Quote: Chodempole

We can't be hypocritical though. Sports betting professionals do exist, so how can we say data mining software wouldn't be achieving some sort of handicapping that a human does successfully?

I think it's quite obvious you're a shill for said company, and hitting the forum right before football season. But, it's a losing proposition. You can't win long term if you're paying anything substantial for picks.

Quote: SM777

I think it's quite obvious you're a shill for said company, and hitting the forum right before football season. But, it's a losing proposition. You can't win long term if you're paying anything substantial for picks.

I disagree Chodempole is shilling for this particular company. His body of posts over 4 years has pretty consistently been on this topic of sports betting and probability. I think.he's just a customer looking for validation or system credibility. Which I doubt anyone will provide, but that's why he asked the question.

Quote: Chodempole

We can't be hypocritical though. Sports betting professionals do exist, so how can we say data mining software wouldn't be achieving some sort of handicapping that a human does successfully?

Yes, sports betting professionals do exist but very few of them win from betting sports. The ones that are winning are not selling their systems.

The Bet Labs site seems like it is fairly expensive, but potentially useful. While reference to, "Systems," is made and the lowest price is $49/month, while I wouldn't spend it anyway, they do offer a lot of easily accessible historical data all in one place.

Anyway, the, "Systems," are really more like conditions for making a bet. It's not referring to a Martingale System or anything like that. For example, "Road trip, Off-Loss, Favorable Line Movement," is one such, "System," but all it really is would be the results of all such games for which they have data. That one claims to have gone 167-100 for $6,032 over all bets made for an RoI of 23%. That would be $26,226.09 in total bets (Approx) which comes out to $98.225 average amount per bet.

I don't know if they just assume $100 bets for the purpose of the system or if they are using actual data, but almost all of their systems come out to a $100 average bet, roughly.

One of the things you can do is test your own system against the historical odds. I'm not linking to their site, but this is from it:

Quote:

Think there’s an edge for home teams coming off a loss in cold weather? How about that .500 team who’s lost 3 in a row. There’s no chance they lose the first half at home, right? Whatever you can think of, you can build a system to test your theory.

So, that's basically just using the data/odds that they have aggregated since 2003 to formulate your own system, but again, it's not a system in the Martingale sense...just a set of conditions under which you're willing to place a bet.

On top of that, whenever a current line matches the set of conditions you have selected, you'll receive an E-Mail and a text telling you so.

Of course, there's a catch, the $49/month doesn't give you the whole thing, here are some of the key differences:

$49/$249

-Eight sportsbooks instead of 50+
-Ten betting systems instead of unlimited.
-No, "Best bet," picks sent to you.
-No Steam Move or Sharp Money Plays

Anyway, I don't know how much value that $249/month plan has over the $49 one, but that's really because the data would be the only thing that would be of any interest to me at all.

So...those are basically the positives, now let's talk:

The Negatives

1.) The flaw with defining highly specific trends on any kind of post-facto basis is that not only can it come down to limited sample sizing, but more importantly, you're just trying to build a system to fit what has already happened even though it might not have any future predictive value.

2.) On the 167-100 one above: The performance data goes back to 2003, which is when they started tracking everything, so that would be over fifteen NFL Seasons or about 60 months.

That would give the person about 4.45 opportunities to bet per month using that particular, "System," so the 23% RoI, if it could be relied upon, would bring back $102.35 in profits per month assuming $100 bets. That sounds awesome until we go back and look at Item #1 again, which is that the whole thing could be bupkus to begin with.

There are also possibly more details that don't appear on there. In fact, I'm sure there are because, the larger the sample sizes get, the more the RoI returns to Earth.

Really, that's what the build your own system component does. It allows you to select from dozens of different criteria and put them all together to see if any particular set of circumstances yields a player-favorable situation.

3.) The very thing that is the, "Hook," of the product, the systems, are limited to ten unless you want to pay an extra $200 per month. Why would this be? I don't know. Why do unlimited systems cost more? Don't know. Does offering unlimited systems even cost the site an extra dollar per month? No, it can't, how?

Conclusion

I could see where the data aspect of it all in one place could be useful to hardcore sports bettors and I also believe that there might actually be some meat to one out of every several betting conditions that they propose. I think that someone would have to bet sports a good bit to take advantage of any of it given the cost and that's assuming that you pick the actual valid positive conditions and that nobody beats you to it and the line changes anyway.

Also, these systems look at how they have performed all-time, but remember, they are going back in time and deciding the conditions under which the data will be counted as opposed to having made the picks at that time based on those postulations.

If I look at all MLB games that took place on a windy Wednesday in the middle of the afternoon that took place following a loss with the Ace pitcher starting that day, and I get to do this with 32 different MLB teams, I imagine that I'll find one team or another who it just so happens has played above or below expectation given that very specific set of conditions.

If I were going to use this software, and I wouldn't, I'd be more interested in the conditions with a sub-10% RoI over more than a thousand samples than in things with an astronomical RoI over only a couple hundred.

Anyone who pays for sports picks is a fool.

Quote: sabre

Anyone who pays for sports picks is a fool.

/endthread

Quote: Mission146

1.) The flaw with defining highly specific trends on any kind of post-facto basis is that not only can it come down to limited sample sizing, but more importantly, you're just trying to build a system to fit what has already happened even though it might not have any future predictive value.

Great write up Mission. I think the point above is the most important when it come to betting trends.

Quote: Mission146

1.) The flaw with defining highly specific trends on any kind of post-facto basis is that not only can it come down to limited sample sizing, but more importantly, you're just trying to build a system to fit what has already happened even though it might not have any future predictive value.

Sounds like you're describing overfitting. That's a problem in all applications of machine learning (ML) which is why the historical data sets are typically split into a training set and a test/validation set. The question is did the Bet Labs folks take the steps required to prevent that kind of problem?

Mission, great post. If you want to screw around with the data, you can do it for free here: http://sportsdatabase.com

Thanks for the compliment and the link! Unfortunately, I have no hope of putting in the search terms the way that thing wants me to.

there is one and IMO only one thing you can count on in betting sports
that is the home/away differential
it stays very close to a constant % throughout time in all the major sports differing mainly in one sport to another not within each sport

almost always any possible edge using this is wiped out by the spread or money line odds
but not every single time. there are reasons, sometimes, it is not given as much importance as it should
i.e. - sometimes very strong favorites will not pay less than about -180 when they should because the books know bettors don't think a small payout can have value and the books would prefer to have the betting balanced if possible. generally, bettors prefer large payouts to small payouts believing that large payouts are the ones with value. and sometimes they are right. but not all of the time.

Any system - fails in long term

You mast analyze every match

analyzing - is key to success in betting