How to determine +EV?

[CrazyTrain]

Hi guys,

New member but have been reading posts for a while. Seems like this is the most intellectual sports gambling forum and am looking forward to some good discussions.

Regarding my question, as I read books and posts and listen to betting podcasts, it looks like determining +EV is extremely important for a successful betting "career". While I do understand the philosophy of taking advantage of a +EV situation, my question is how do you as a gambler determine the % needed to compare to the offered lines?

From what I see, you must have a spreadsheet/model or something that generates definitive numbers rather than just using non-quantifiable research such as reviewing the game data and thinking Team A has a 65% chance to win this game.

My first gambling book was/is Yao's "Weighing the odds..." and while it provided interesting info, to me the most important part missing was how I am to create my own numbers to compare to the line so I can determine +EV. Without that, it's like having a nice car without an engine.

Recently, in an effort to create my own numbers, I decided to take MLB games wherein both teams had a high Over % and the ump had a high Over % and vice versa for the Under. I added all three percentages together, divided by three and took that % and compared it to the line and if I had a difference of 3% or more, I took the wager. I'm slightly ahead in a small sample size, but I figure that's more luck than any brilliant system on my part.

While considering this "system", I wondered if I should extend the variables to the pitchers, parks, home, away, day, night, etc. As you can see, it becomes stat overload and then my eyes gloss over.

Anyway, I've begun to ramble so if any of you smart fellers out there can offer some answers, advice, or whatever i greatly appreciate it.

And please note, by no means am I a math whiz so if you can keep it to a layperson explanation, that would be wonderful.

Thanks and Good Luck to all.

[RickySteve]

Spam the board and get there, that's what I do.

[dj_destroyer]

While considering this "system", I wondered if I should extend the variables to the pitchers, parks, home, away, day, night, etc. As you can see, it becomes stat overload and then my eyes gloss over.

You got everything down except for your eyes glossing over when it starts to get tough... When you think you have enough information, find more. Wade through the bullshit though and stick to the pertinent information. Knowledge is power.

[Flight]

From what I see, you must have a spreadsheet/model or something that generates definitive numbers rather than just using non-quantifiable research such as reviewing the game data and thinking Team A has a 65% chance to win this game.

Quantitative analysis is the only way to win (although the complexity of the analysis can vary greatly based on what you're betting.) Start small - big sports have big bettors and people that work all day to find EV; so you need to shop somewhere else: small sports, or other markets.

Recently, in an effort to create my own numbers, I decided to take MLB games wherein both teams had a high Over % and the ump had a high Over % and vice versa for the Under. I added all three percentages together, divided by three and took that % and compared it to the line and if I had a difference of 3% or more, I took the wager. I'm slightly ahead in a small sample size, but I figure that's more luck than any brilliant system on my part.

Just data mining - you looked at something that occurred in the past and are trying to predict the future with it. Won't work. You need to identify predictive parameters, for example, what contributes to a team's over %. In regards to umpires, you can use that to make adjustments, but I would caution against it (due to variance and lack of predictive strength). At least put a coefficient on that term to only allow maybe 3% adjustment to your estimate.

While considering this "system", I wondered if I should extend the variables to the pitchers, parks, home, away, day, night, etc. As you can see, it becomes stat overload and then my eyes gloss over.

More variables is better - I use everything you mention (except day/night) and then many more (everything in a box score is useful). If you're eyes are glossing over, start with something simpler and figure out a way to get simple numbers easily. For example, you should be able to use Bill James's pythag and have an estimate on two teams within seconds. Work from there and begin adding improvements.