Very Simple Projected Runs scored Formula

[Heinz57]

I am aware this formula for projecting runs scored is not going to be very accurate and that betting games with it would likely cost you money, but as a quick and dirty tool how bad is it?

Note: Avg refers to Median not mean...

[(Team A's Runs/Team A Games Played) x [((Team B's Starting Pitcher's ERA x Avg # of IP)/ 9 ) + (Team B's Relief Pitcher's ERA x (9- Team B's Starters avg IP) / 9)]

Take the Square Root of the whole thing and this is your projected runs for Team A. Repeat for Team B.

Some questions I'm pondering are:

What are the biggest problems with this formula?

Is forty games into the season anywhere near a large enough sample size for this anyway? I would think so seeing as it's over 30 but am not sure.

Finally, it does factor in a teams errors that result in runs from the defensive perspective, but that seemed difficult to capture and express. Everything I've included is readily available and simple to gather.

Thanks in advance for the feedback, Heinz57

[MHoydilla]

I have never seen a formula like that in a book or talked of. I also have absolutly no idea if it works, I would be intrested in reading a post with some projection pre-game time though.

[The13atman]

Do this and you'll be very very close to the actual lines (usually within half a run), so it doesn't help too much. I've experimented with this formula or one very similar to it and it like I said, it's so close to the actual line barring some exceptional circumstances (such as a "good" pitcher who has a bad ERA to start the season such as Barry Zito or Randy Johnson, where the line will be lower than the formula's projection) that it's not of much use. There would be A LOT of variance with this strategy I think, especially if you're willing to bet on very small edges according by the formula, such as betting the under when the formula comes to 8.7 and the line is 9. Good luck though.

[Heinz57]

Yeah, I don't think it's strong enough by any means to bet soley on, but I like when my instincts and general feel for a game are confirmed by something mathmatical. I want to start betting more based on the numbers as oppossed to my opinion and am just trying out some different things.

[mrbaseball]

[ QUOTE ]
Do this and you'll be very very close to the actual lines

[/ QUOTE ]

Yeah.

What I did was take the home teams median total at home, the road teams median total on the road, the home teams median runs scored at home added to the road teams median runs scored on the road added to the home teams median runs given up at home and the road teams median runs given up on the road. Then average all of that crap together. What you get is 99.9% identical to the actual line unless there is a pitching outlier.

So what I do is just look at number of occurances. How many times has the home team gone over or under the predicted number at home and how many times has the road team gone over or under the predicted number on the road. If the number of occurances, both for home and away teams, drifts significantly enough away from 50% and there is no pitching surprise I have a bet.

[The13atman]

[ QUOTE ]
So what I do is just look at number of occurances. How many times has the home team gone over or under the predicted number at home and how many times has the road team gone over or under the predicted number on the road. If the number of occurances, both for home and away teams, drifts significantly enough away from 50% and there is no pitching surprise I have a bet.

[/ QUOTE ]

Yep, this is much more reliable.

[DougOzzzz]

I have a spreadsheet which calculates the distribution of total runs/game given input of both teams RPG (MEAN expected value). In a way it's similar to this formula - though adjustments for each team's pitching etc. are not included in the calculations - you have to come up with an accurate projected MEAN runs/game for it to be of any use. But for converting mean into median, it is extremely accurate based on historical evidence.

[Danimal]

I find this post very interesting. I like the basis you have, and I'm sure adding things along the way, will make this very helpful.

[Heinz57]

Thank you, I like these types of discussions much better than the typical I'm on these games what do you think of my picks (that no rational was provided for)...I appreciate everyone's input and will continue to tweak and test for sure.