[Justin7]

Intro to stat-based approach.

[j0hnnyv]

great video. thanks.

[RogueJuror]

Good video Justin.

[whatisit]

For the people who use stats/math based approaches to gambling, how high of a math degree do you have? By this I mean, do you use a lot of upper level/advanced statistics or can a lot of it be learned on your own or by taking a few lower level stats courses in college?

[Justin7]

If you passed a class on calc, you could figure out everything else. The harder thing is thinking creatively - how do I think this game works, and how do I use math to represent that.

While statistics is a useful field, there are two that are more useful (for my approaches): basic probability, and discrete and combinatorial mathematics.

[Justin7]

Justin,

Good vid. One question. How do you account for relief pitching? Yes we know what pitchers are more likely to come out for relief based on previous games, but it seems many teams bring out an individual pitcher based more on the circumstances of the game, as opposed to always going to a main guy. Any insight to this?

Take all the pitchers available for relief. Add up all the stats - innings pitched, hits, ... That is your generic "bullpen" stat. Estimate how long the starter will pitch - if he's projected to play 210 innings in 30 starts, I'd assume he goes 7, and bullpen goes the remainder.

[pico]

justin is the man. only reason i am still here at sbr.

[MonkeyF0cker]

You only have too many if one of your variables is not predictive.

[pico]

hoocus poocus with math. still solid video

[MonkeyF0cker]

I prefer simulation models to EV models actually...

[Justin7]

Okay good stuff. How do you know when you are starting to add too many variables? I mean we could throw Umps in here, Weather maybe, park dimensions, other stats about the bullpen ect. It seems this would help better define any model, yet how do you know when you have too many in the mix or adding things that may decrease your models success rate?

As said below, as long as it's predictive...

Park factor - use from Bill James handbook. These affect totals a lot, but not the ml so much. Weather - I ignore everywhere but Wrigley field.

The more you can add correctly, the better. But just using the player stats will give you a slightly winning model if you do it correctly. It doesn't have to be perfect - just better than the darts thrown up on overnights.

[Justin7]

I prefer simulation models to EV models actually...

I've tried that. I had serious offers for my old sim so they could fade it.

[MonkeyF0cker]

LOL. MLB, Justin?

[Justin7]

LOL. MLB, Justin?

Yes. Very serious offers.

[brooks85]

this is great cause I really dont know much about baseball

[MonkeyF0cker]

They certainly aren't easy, but if you get them right, they can be rather profitable.

[Justin7]

They certainly aren't easy, but if you get them right, they can be rather profitable.

The definition of depressing: After losing a ton of $$$, you read a Ph.D. thesis explaining exactly why your model sucks.

[j0hnnyv]

not only is this video solid. it looks like the lines went your way on most of your picks for today justin. they look good, gl to u.

[Justin7]

not only is this video solid. it looks like the lines went your way on most of your picks for today justin. they look good, gl to u.

The average line move for picks in my threads has been 8 cents in my favor.

Do that math

[MonkeyF0cker]

The definition of depressing: After losing a ton of $$$, you read a Ph.D. thesis explaining exactly why your model sucks.

Brutal. It wasn't a Markov chain Monte Carlo sim was it?

[Justin7]

Brutal. It wasn't a Markov chain Monte Carlo sim was it?

Monte Carlo. Play the game 10m times, and set prices that way.

[tweek]

The definition of depressing: After losing a ton of $$$, you read a Ph.D. thesis explaining exactly why your model sucks.

Justin-- cool video! I'd be interested in reading this thesis... do you have a reference?

[Justin7]

Justin-- cool video! I'd be interested in reading this thesis... do you have a reference?

Not any more, sorry... But you can find all sorts of baseball articles if you search for them.

[Flyers09]

justin7- could you or do you have a program made to where you just enter runs scored etc. and the program calc. the percentage?

[Justin7]

justin7- could you or do you have a program made to where you just enter runs scored etc. and the program calc. the percentage?

It's trivial to do a spreadsheet in excel that does this.

[Flyers09]

thats a good idea justin7, i really like your video, im kind of new to the mlb world, and really dont know how to cap these games until now, im going to try this theory out, but i need to watch your video a few more times, im a little slow in math...

[Flyers09]

also could this theory work if i got back to last seasons stats, to get my 100 game marker?

[Justin7]

also could this theory work if i got back to last seasons stats, to get my 100 game marker?

If you have lineups, historical lines and results, yes. You might overshoot though - you won't have any mistakes in your lineups, which will increase your hit rate.

[Flyers09]

so your bascially saying i cant use the stats from this season so far?

[Justin7]

so your bascially saying i cant use the stats from this season so far?

Sure you could, but I ignore them.

[Flyers09]

do you bet mlb? if so how effective is this theory? percentage wise?

[durito]

do you bet mlb? if so how effective is this theory? percentage wise?

It wins 89%. You should go deposit your entire paycheck to an online book right now.