MLB Data Analysis

[mesadog]

I don't make my 1st Inning plays at BJ. I make my side plays there.

As I said before to much criticism I don't account for "value". Lot's of people do that, and they lose their rolls... which is fine with me. I use statistical analysis to predict the most likely plays to win, and then I have a money management system to capitalize on my wins and minimize risk on my losses.

To me, a lot of people try and approach the sports betting market like the stock market. That may work for them, but that's not what I do. At the end of the day my system works, however it's not something everyone can do. If you don't know how to model the result of a future event then I understand why you'd think my system has "no chance" of working. My bankroll says differently.

With that said, using a probit model which is what my partner and I primarily use doesn't work for ATS sports quite as well... so for that we use other techniques. What the probit does do in non-ATS sports such as baseball though, is gives us massive insight as to which plays have a greater than 60% chance of winning. We take our risks on those plays and then use our money management strategy to maximize gains.

[bztips]

Given your apparent math skills, I don't see how you don't recognize the folly in blindly betting on something with a 60% or more chance of winning without regard to the odds (which is where the "value" comes in). If the odds are -150, then betting on a 60% winner will give you zero expected value. It doesn't matter how good your money management skills are -- good MM can't change zero or negative EV bets into +EV.

Or maybe I (and other posters) have misinterpreted what you're saying....?

[mesadog]

I don't use them for first inning plays... and I'd rather not say where I make my FIs. The last thing I need is them being taken away.

[mesadog]

We rank plays based on estimated win %. Amongst MANY other methods how can you not see how this results in positive value?

[mebaran]

We rank plays based on estimated win %. Amongst MANY other methods how can you not see how this results in positive value?

There is only value if the true odds of winning (those that are predicted by a model) are greater than the implied odds of winning (odds the line indicates). And even then, there is no guarantee. Most people use a cutoff in their models to deal with inaccuracies. Myself, if there is more than a 5% difference in price from my line to the book's line, I rate it as a play.

[bztips]

We rank plays based on estimated win %. Amongst MANY other methods how can you not see how this results in positive value?

Maybe it's just a confusion of terminology. Most people in this forum use the term "edge" to describe a +EV situation, which can only occur if the estimated win % from your model is greater than the implied win % from the the posted odds.

But if you simply bet all picks where your model estimates that side will win the game 60% of the time or greater -- without considering what the posted odds are on each of those games -- then no, that makes no sense.

[fax6202]

Maybe it's just a confusion of terminology. Most people in this forum use the term "edge" to describe a +EV situation, which can only occur if the estimated win % from your model is greater than the implied win % from the the posted odds. But if you simply bet all picks where your model estimates that side will win the game 60% of the time or greater -- without considering what the posted odds are on each of those games -- then no, that makes no sense.

Interesting thread....BZtips, a question for you. I am relatively new to sports betting, at least serious sports betting. I have done fairly well following cappers who are doing well at a particular part of the season. My question is...if i wanted to handicap my own picks..besides delving into a ton of stats, which i do. How do i get to the point where i can determine +EV bets. Sorry for the rookie question, but i have to start somewhere.

[bztips]

Well, modeling is not for everyone; it takes a fair amount of basic knowledge in statistics, knowing how to gather and organize data, and some Excel or database skills. But if you want to consider that route, Justin7's book is a good place to start -- http://store.sbrforum.com/products/300-CONQUERING-RISK

[mathdotcom]

My evaluation:

1. there is no doubt he is making +ev bets, has a decent to very good database and knows how to use it
2. spread way too thin across too many markets, obviously inexperienced at sports betting, not been around a long time
3. small bankroll but growing quickly

which leads to the recommendation:

You do not need any more people in your fold. You need more money and more experience with offshore so just keep doing what you're doing.

And do not neglect your thesis.

[fax6202]

Well, modeling is not for everyone; it takes a fair amount of basic knowledge in statistics, knowing how to gather and organize data, and some Excel or database skills. But if you want to consider that route, Justin7's book is a good place to start -- http://store.sbrforum.com/products/300-CONQUERING-RISK

thank you

[evo34]

Oh my fukking god. You are not just a beginner; you are a genuine waste of talent. Good luck with your terminal masters degree.

I don't make my 1st Inning plays at BJ. I make my side plays there. As I said before to much criticism I don't account for "value". Lot's of people do that, and they lose their rolls... which is fine with me. I use statistical analysis to predict the most likely plays to win, and then I have a money management system to capitalize on my wins and minimize risk on my losses. To me, a lot of people try and approach the sports betting market like the stock market. That may work for them, but that's not what I do. At the end of the day my system works, however it's not something everyone can do. If you don't know how to model the result of a future event then I understand why you'd think my system has "no chance" of working. My bankroll says differently. With that said, using a probit model which is what my partner and I primarily use doesn't work for ATS sports quite as well... so for that we use other techniques. What the probit does do in non-ATS sports such as baseball though, is gives us massive insight as to which plays have a greater than 60% chance of winning. We take our risks on those plays and then use our money management strategy to maximize gains.

[uva3021]

Can't imagine an NBA "model" analytically reduced to measures of simplicity actually working. The league is an emergent system, and the relationship between the league and the linesmakers is analogous to a parasite and its host. Also having trouble imagining a successful MLB "model" that renders the actual moneyline of little consequence.

If you are bringing in 100 units per sport what exactly is the nature of your inquiry. You should be increasing the size of your bets rather than messing with a more efficient method of cataloging data. Obviously the latter is immaterial to the ultimate means which have already been realized.

[donjuan]

So you can calculate a z-score but are unable to calculate EV? Impressive.

[RickySteve]

Hilarious thread. Would read again.

[mathdotcom]

Just because the guy can't talk the talk doesn't mean he isn't making money and won't continue to make money. He has said some strange things about "60% plays" and not treating betting like the stock market, but there are plenty of guys here who can talk EV and Kelly but are obviously squares or $5 losing bettors (Wrecktangle has got to be the #1 example). Many of you look up to Justin7 who had to hire a PhD mathematician to run a probit model for him -- something mesadog is handling all by himself. I think many of you guys are jumping on him for using the word "system", which for him is not a system as most use the term = betting based on a silly non-causal trend. His method is seemingly based on solid econometrics, but he may have a strange rule of thumb for when to pull the trigger. That's why I suggested he needed more experience and not more data.

There are a lot of sharps who don't calculate their EV but bet based on whether a line is off by more than x cents. They could calculate their EV if you hassled them but they choose not to. They just know it is +EV by some amount they are comfortable with and bet something they are comfortable with (so they don't need to know the EV to bet according to Kelly), or they don't need to know EV because they can't get down as much as they'd like to anyways or they don't want to attract attention. And when you're in a daily MLB grind for 6 months it can save time.

[donjuan]

Yes, I'm sure someone who doesn't care about what the price is on a bet in a ML sport is just raking it in.

[Flight]

Mesadog,

If your model estimates a road team has a 40% chance of winning, and BetJamaica is offering the moneyline at +200 road / -240 home, what would you do?

[bztips]

Mathy, you make some good points. But there's just something very odd about the guy's description of what he's doing. The sharps you're talking about who just "know" when a line is off are not the kind of people who estimate probit models to figure out which is the correct side. If you have enough background to develop and estimate a probit model, then computing your edge is completely trivial -- you can literally do it in your head.

I do agree with you however that he probably just needs more experience.

[mathdotcom]

Again, I think you are being distracted by his poor terminology/explanation. Maybe he never bets -150 or higher, so he makes a play anytime his model predicts a 0.6 probability of winning. Bizarre, but +EV.

[donjuan]

You're just making things up at this point.

[Dark Horse]

Mathdotcom lives in the real world. The sect must not deal with smaller sample sizes. If they do, they can't possibly be as precise as they love to assure everyone. And if they don't, they're leaving money on the table. Which one is it?

[bztips]

Mathdotcom lives in the real world. The sect must not deal with smaller sample sizes. If they do, they can't possibly be as precise as they love to assure everyone. And if they don't, they're leaving money on the table. Which one is it?

Nice -- has nothing to do with what this thread is about.

[donjuan]

Can we get an official list of sect members so when Dark Horse makes his next straw man we know who he's referring to?

[Dark Horse]

Nice -- has nothing to do with what this thread is about.

Isn't much of this thread about the guy's inability to calculate his precise edge? Underlying this popular criticism by the sect, it would seem to me, is the assumption that they are able to calculate their edge with tremendous precision. Aren't smaller sample sizes a reality in sports betting? You do the math.

[ForgetWallStreet]

Isn't much of this thread about the guy's inability to calculate his precise edge? Underlying this popular criticism by the sect, it would seem to me, is the assumption that they are able to calculate their edge with tremendous precision. Aren't smaller sample sizes a reality in sports betting? You do the math.

He's not even CONSIDERING payout odds when determining which games to bet, but yes, let's keep giving him the benefit of the doubt because he claims he has models that have won units in the past.

[donjuan]

Isn't much of this thread about the guy's inability to calculate his precise edge? Underlying this popular criticism by the sect, it would seem to me, is the assumption that they are able to calculate their edge with tremendous precision. Aren't smaller sample sizes a reality in sports betting? You do the math.

LOL that is what you got from this thread?

[bztips]

Isn't much of this thread about the guy's inability to calculate his precise edge? Underlying this popular criticism by the sect, it would seem to me, is the assumption that they are able to calculate their edge with tremendous precision. Aren't smaller sample sizes a reality in sports betting? You do the math.

You are clueless. Like I've already mentioned, given that he's already estimated a probability model, he has only to take one more trivial step to get his estimated edge. And he's already stated that he's done all kinds of backtesting so that (at least in his mind) sample size isn't an issue. So guess what, he's actually part of the "sect" that you're trying to defend him against.

[Dark Horse]

LOL that is what you got from this thread?

That is what I take from you, and your misplaced arrogance. Tell us, oh great one, how you establish a precise edge for smaller sample sizes. Enlighten us.

[Dark Horse]

You are clueless. Like I've already mentioned, given that he's already estimated a probability model, he has only to take one more trivial step to get his estimated edge. And he's already stated that he's done all kinds of backtesting so that (at least in his mind) sample size isn't an issue. So guess what, he's actually part of the "sect" that you're trying to defend him against.

Uhoh. I sidestepped into uncomfortable territory. Therefore it must be unrelated.

My question is valid, in the real world of sports betting. A favorite sect mantra is to gloat that people can't calculate their edge precisely. So let's hear it from these geniuses. They either sidestep everything related to smaller sample sizes, or they make themselves look a lot better than they really are.

[Inspirited]

What's going on here! People just want to know if he knows how to convert lines into probabilities (and also probabilities into lines) and then if his > 60% plays all still make long term sense in terms of simple EV calculations given the lines he's betting at. Precision or no precision I think he should know how to do this?

[ForgetWallStreet]

What's going on here! People just want to know if he knows how to convert lines into probabilities (and also probabilities into lines) and then if his > 60% plays all still make long term sense in terms of simple EV calculations given the lines he's betting at. Precision or no precision I think he should know how to do this?

Welcome to the sect.

[Dark Horse]

What's going on here! People just want to know if he knows how to convert lines into probabilities (and also probabilities into lines) and then if his > 60% plays all still make long term sense in terms of simple EV calculations given the lines he's betting at. Precision or no precision I think he should know how to do this?

No doubt.

You haven't been here long enough to understand the background. The sect members, basically dj and rs with a few followers, have attempted to police this place for years with misplaced arrogance, chasing people away left, right, and center. They were, not surprisingly, making fun of the new guy as well.

So I asked them a simple question here. A question related to the reality of sports betting. And while I don't expect to see them at the races any time soon, I would still love to hear their take.

[ForgetWallStreet]

You haven't been here long enough to understand the background. The sect members, basically dj and rs with a few followers, have attempted to police this place for years with misplaced arrogance, chasing people away left, right, and center. They were, not surprisingly, making fun of the new guy as well.

Except there's nothing misplaced about it. If SBR instituted a policy where one had to prove sports betting earnings of $XXX,XXX before posting in the think tank, guess who wouldn't be posting here out of you, DJ, and RS? I'll give you a clue, it's not DJ or RS.

[donjuan]

That is what I take from you, and your misplaced arrogance. Tell us, oh great one, how you establish a precise edge for smaller sample sizes. Enlighten us.

Well if anyone else needed any further proof you are a complete and utter moron, here you go.

[donjuan]

Except there's nothing misplaced about it. If SBR instituted a policy where one had to prove sports betting earnings of $XXX,XXX before posting in the think tank, guess who wouldn't be posting here out of you, DJ, and RS? I'll give you a clue, it's not DJ or RS.

I'm definitely nowhere near RS in terms of lifetime or even annual earnings.