I'm serious about building a succesful model to predict baseball games. I have collected all the historical data that I need from baseball reference dot com. I have found a run estimator on the internet. I'm using Jim Furtado's Extrapolated Runs Formula. I'm going through each line-up card and calculating singles, doubles, triples, etc... The problem is this is very tedious and time consuming. It takes an hour each game to do it by hand. Is there a way to automate this process? Can I have the computer calculate everything for me? The calculations are simple. Addition, subtraction, multiplication and division. Should I use team stats to speed up the process? I hesitate to do this, because I will lose some accuracy. I don't have sixteen hours a day to devote to baseball handicapping. Any suggestions?
-
-
Excel is your friend for stuff like this.Comment -
Open Office has a good spreadsheet alternative for free download.
You need to learn a programming language. I use C# because it has a great API and library. For beginners, Python seems to be the consensus on the forum. Python also has a rich library for doing things like interacting with web data and parsing XML.
Based on your description, you would be crazy not to automate this. I've automated everything I can from gathering web data to analyzing teams and players, and it still takes me about 10 minutes per game/prop to go analyze.
One last comment: don't bet using the results of your model. (At least not for a long time, and dont' use MLB. Find a smaller league to bet on)Comment -
Open office spreadsheet is basically excel - if I didn't have excel thanks to work I'd still be using it.Comment -
Usiing the XR formula:
Structure your spreadsheet by column in this exact order of variables (1Bs in first column, 2Bs in second column, 3B in third column...)Code:XR – Extrapolated Runs = (.50 × 1B) + (.72 × 2B) + (1.04 × 3B) + (1.44 × HR) + (.34 × (HP+TBB−IBB)) + (.25 × IBB)+ (.18 × SB) + (−.32 × CS) + (−.090 × (AB − H − K)) + (−.098 x K)+ (−.37 × GIDP) + (.37 x SF) + (.04 × SH)
add a total sum column at the end of the row of XR calculations
Then write a macro, I'll get you started:
Comment -
I didn't realize Works had access to VBA macros.Comment -
you can also use google doc as you can create an excel spreadsheet on there. Other then excel like document, does anyone have an actual software they use to help them predict games?Comment -
If you're running on a mac like me, there's a handy new program that comes standard with Snow Leopard called "Automator"... its similar to a macro recorder in excel but a LOT more user friendly and it lets you automate tasks across programs on your computer. Sounds like this could be helpful IF you're on a mac... I'm currently doing a search for something similar that you could run on a PC but so far, no luck... I'll be sure to keep you updated though, sounds like a very cool project!
Comment -
well once you have written the visual basic you just have to change the properties and naming conventions to the framework of a visual basic form, i suggest getting visual studio, or just write it as an offline web page, using tables
if you understand program it doesn't take long before you can manifest some sort of automated game predictor, i have one for each of the major sports (includiing WNBA)Comment -
I am not at all meaning to rain on your parade, but unless you come up with a really interesting, novel sabermetrics insight, you won't succeed in beating the line this way. ALL public sabermetic knowledge is incorporated into the line. Where the line varies from it, there's a reason.
Common sabermetric score estimations are the new vanilla in MLB sportsbetting.
But you can still have lots of fun trying. Enjoy.Comment -
Ive been making money all year with sabr stuff. I dont think its all priced in quite yet.Comment -
This statement is too vague to be judged.
MLB doesn't run all "year", so you probably mean "season."
No one wins "all" season, everyone has losing days, so you probably don't mean "you've been making money all year", you probably mean that "you're ahead for the season."
"With sabr stuff" can mean just about anything. "Sabr stuff" is a huge range. And further, just because you've been winning doesn't mean it's because of "sabr stuff." If you go to a race track and bet all the horses in red silks and make a profit, you didn't win because you bet all the horses in red silks. Such "winning" isn't causal, it's coincidental.
If a hundred guys try to beat MLB for a season with "sabr stuff," a few will end the year ahead. If a hundred guys try to beat MLB for a season without "sabr stuff," a few will end the year ahead.
And anyway, "winning," if for only a little bit, or for only a few hundred bets, can mean just about nothing. Up a dollar after a thousand bets? Winner. Up a thousand after one bet? Winner. For the year.
And LOL, this response reads a little bitchy, so I apologize. But my point is right: your statement, as it stands, isn't evidence of anything.Comment -
I'm not doing this to have fun. I'm running this formula to have an edge. Have you tried using the Extrapolated Runs formula? How do you know it won't give me an edge? Past experiences or are you guessing? I know it is available on the internet for the public to find. Does this automatically make the formula a waste of time? Is it better to guesstimate what a line should be?Comment -
Don't get frustrated at the messenger. I'm green and you even sound bright green to me.
If you are as serious as you sound, what Pokerjoe is saying will be self evident to you pretty soon.
And that doesn't mean you are doing the wrong thing here. Most of us have to put in this work ourselves to really understand/believe it's actually going to take 'something' more than just following the formulas..Comment -
Just don't bet the model before you test it.
It would be useful to read Justin7's book. There is a chapter focused on MLB run estimation.Comment
