Baseball Handicapping Theory..."Magnet" Lab

Search

Member
Handicapper
Joined
Mar 7, 2005
Messages
8,849
Tokens
I have my own "system" for handicapping baseball. (It's been flat this year...so far.) But I often come up with ideas that I would like to explore.

The problem is "time". I just don't always have time to explore everything I would like to explore.

Would anyone like to take the ball and run with it so to speak? Perhaps we can have an experiment here if someone is willing to do the work?

I have some ideas to share...anybody have the time to test it out?

Lets call this our "Magnet" Lab. I'll explain further if anyone shows an interest.

Lets keep the whole process public here....it might be an interesting thread for others to follow as well.


It will take a daily....or near daily committment to do some basic research...make the picks...and keep track of the record for us. I would highly recommend using Wagerline for this...it makes it very simple to make the picks...and it will track the record as well.

Any volunteers that can make the daily committment and also have the interest? (estimating about 30 minutes a day...maybe less. It could be done early in the morning or preferably by 5pm CST. each day at the latest. Earlier the better. )


Any volunteers? :dancefool

Multiple volunteers would be fine as well...we could run variants of the system to try to fine tune it.
 

New member
Joined
Nov 3, 2007
Messages
95
Tokens
I would definitly be interested... If you want grab my e-mail and we could exchange ideas farther:103631605
 

Member
Joined
Nov 20, 2004
Messages
3,934
Tokens
MJ,
I wouldnt mind helping as i have all kinds of time.
 

Member
Handicapper
Joined
Mar 7, 2005
Messages
8,849
Tokens
OK...great...we have some people with an interest.

I'll write up an explanation of my idea later tonite...and then we can start talking about it...and maybe bounce some ideas off each other.

:toast:
 

Member
Handicapper
Joined
Mar 7, 2005
Messages
8,849
Tokens
Ok...lets see if I can explain what I have in my head.

Thats not always an easy thing to do. :lol:

Lets start off by going to ESPN. Why ESPN? Because I like the way they organize their stats for baseball.

Use the following link, we want to look at MLB Relative Power Index - 2008: http://sports.espn.go.com/mlb/stats/rpi

All the teams are in order according to the ESPN Relative Power Index. It's kind of a cool formula they use.

The basic formula is 25% team winning percentage, 50% opponents' average winning percentage, and 25% opponents' opponents' average winning percentage.

Maybe we can figure out a way to use the RPI but my idea is to use something else on this page. Something very powerful, a formula that has merit in all the major sports.

It's called the Pythagorean theorem of baseball.

Now don't get worried...it's all done for us...no math involved here...it sounds much more complicated than it really is.

All you really have to understand is this:

This formula was designed to relate a team's runs scored and runs allowed to its won-lost record.

Thats it! A teams runs scored and runs allowed correlates with its won-lost record...in theory.

And it has proven to be true...over time...not all the time...but for most teams...over time...most of the time. I'ts not a lock...but its a pretty solid piece of information to work from. It's a good start for a handicapping experiment.

Well...we hope it is anyway...this is all an experiment.

Now here is where it gets interesting. Look again at this link:
MLB Relative Power Index - 2008: http://sports.espn.go.com/mlb/stats/rpi

Notice on the right hand side of the page where it has an Exp W-L column?

This tells you...according to the Pythagorean formula...what the expected wins and losses should be for each team.

Now look at the actual Win column on the left side of the page.

What we are looking for are teams that are under performing or over performing according to the expected wins/ losses.

In theory...the expected wins/ losses should be equal to actual wins and losses.

When a team gets too far away from the expected win/loss number...in theory...it should get pulled back...like a "Magnet" to what is expected according to the formula.

So all we would be doing is looking for a way to make plays based on teams getting too far away from the expected wins and losses versus what the actual wins and losses are.

Some teams will have more wins than expected...some will have less. The expectation is that teams will have more of a chance to win or lose because of this as they get pulled back to what is expected.

Thats enough for now...its maybe still somewhat confusing...this is just a start.

We still have to figure out how to apply this information.

Lets see if anyone understands what I'm getting at here.

Questions or discussion are encouraged.
 

New member
Joined
Oct 31, 2007
Messages
656
Tokens
I don't know how you feel about chase systems but maybe one would be appropriate here. Say once a team reaches a certain point away from where they are supposed to be you chase them back. For example, a team is losing and reaches a certain point, play on them to win in a chase manner. Make sense? Just a thought.:monsters-
 

Member
Handicapper
Joined
Mar 7, 2005
Messages
8,849
Tokens
I don't know how you feel about chase systems but maybe one would be appropriate here. Say once a team reaches a certain point away from where they are supposed to be you chase them back. For example, a team is losing and reaches a certain point, play on them to win in a chase manner. Make sense? Just a thought.:monsters-

I don't have a problem with a chase system...as long as it doesn't get too crazy. Starting out with 1 unit and ending up chasing 8 units is crazy in my opinion...but I've seen some chase systems that have been effective for several seasons now.

As long as its fun...and nobody gets killed playing it.

Any ideas on how you would want to chase? I'm thinking we still need another qualifier. For example...just chase home series when a team has been under performing...and just chase away series when a team has over performed?

It might give you the best chance for the reversal of fortune?

Just a thought back to you.

:monsters-
 

New member
Joined
Oct 31, 2007
Messages
656
Tokens
That sounds better. My idea was oversimplified but it was just to get the brainstorming going. It may help to incorporate RPI and SOS.
For example, take a dog in 1st game of series who is -5 ExpW-L with higher RPI than opponent then chase it? I don't even know if this situation would even occur, again, just brainstorming.
 

Member
Handicapper
Joined
Mar 7, 2005
Messages
8,849
Tokens
That sounds better. My idea was oversimplified but it was just to get the brainstorming going. It may help to incorporate RPI and SOS.
For example, take a dog in 1st game of series who is -5 ExpW-L with higher RPI than opponent then chase it? I don't even know if this situation would even occur, again, just brainstorming.

I'm not sure either...it might be one of those things we find out after playing around with it for a while.

I wish I could have a daily graph to look at ...like a stock chart...to see how they act.

But thats way too much work for my taste.

I wouldn't be afraid to start out with something simple...if it doesn't work...we can always add layers later.
 

Member
Handicapper
Joined
Mar 7, 2005
Messages
8,849
Tokens
It seems that the expected W/L is right on target with actual W/L.

Magnet Picks Watch List

LA Angels +5
Atlanta -5
Clev -4
Tampa +3
LA Dodger -3
Det -3<!-- / message -->
 

Member
Handicapper
Joined
Mar 7, 2005
Messages
8,849
Tokens
Hahahaha.

Chasing Variance... I'm sure this will end well.

Good Luck.

Don't poop in the punch bowl Captain.

I don't think you really understand my intent here.

It's not so much about thinking we have found a holy grail...its about what you learn trying...or meeting others with the same interest.

I certainly don't expect this to turn into magic pixie dust...but sometimes by trying...you do find valuable tools that add to your handicapping.

You don't learn anything sitting on the sidelines laughing...thats what tailers do.

If you want to make your own picks...you have to be curious. :103631605
 

Fah-New-Gee
Joined
Nov 27, 2007
Messages
3,023
Tokens
Ok...lets see if I can explain what I have in my head.

Thats not always an easy thing to do. :lol:

Lets start off by going to ESPN. Why ESPN? Because I like the way they organize their stats for baseball.

Use the following link, we want to look at MLB Relative Power Index - 2008: http://sports.espn.go.com/mlb/stats/rpi

All the teams are in order according to the ESPN Relative Power Index. It's kind of a cool formula they use.

The basic formula is 25% team winning percentage, 50% opponents' average winning percentage, and 25% opponents' opponents' average winning percentage.

Maybe we can figure out a way to use the RPI but my idea is to use something else on this page. Something very powerful, a formula that has merit in all the major sports.

It's called the Pythagorean theorem of baseball.

Now don't get worried...it's all done for us...no math involved here...it sounds much more complicated than it really is.

All you really have to understand is this:

This formula was designed to relate a team's runs scored and runs allowed to its won-lost record.

Thats it! A teams runs scored and runs allowed correlates with its won-lost record...in theory.

And it has proven to be true...over time...not all the time...but for most teams...over time...most of the time. I'ts not a lock...but its a pretty solid piece of information to work from. It's a good start for a handicapping experiment.

Well...we hope it is anyway...this is all an experiment.

Now here is where it gets interesting. Look again at this link:
MLB Relative Power Index - 2008: http://sports.espn.go.com/mlb/stats/rpi

Notice on the right hand side of the page where it has an Exp W-L column?

This tells you...according to the Pythagorean formula...what the expected wins and losses should be for each team.

Now look at the actual Win column on the left side of the page.

What we are looking for are teams that are under performing or over performing according to the expected wins/ losses.

In theory...the expected wins/ losses should be equal to actual wins and losses.

When a team gets too far away from the expected win/loss number...in theory...it should get pulled back...like a "Magnet" to what is expected according to the formula.

So all we would be doing is looking for a way to make plays based on teams getting too far away from the expected wins and losses versus what the actual wins and losses are.

Some teams will have more wins than expected...some will have less. The expectation is that teams will have more of a chance to win or lose because of this as they get pulled back to what is expected.

Thats enough for now...its maybe still somewhat confusing...this is just a start.

We still have to figure out how to apply this information.

Lets see if anyone understands what I'm getting at here.

Questions or discussion are encouraged.

Here's a problem that i can see with it.

The "ranking" is a self-fulfulling prophecy. I'm no math expert, but i realize that as you compile more data you are in effect changing the stats as you go along. In other words, the worse a team gets, the worse it's record gets, thus the lower the #. The last thing I would want to do is martingale chase a loosing team. While using the data as a snapshot, or picture in time, may be worth doing, I don't know that the team will magnet back towards the centerline. Look at a bell curve. There is a middle that happens most often but the ends of the curve are there and the possibility of something happen.

IMO the best handicappers rely 100% on math and nothing on feelings so I really like the road you're traveling on. I just don't think the magnet is the road we are looking for. I would argue that whatever system is going to be used will need to be based on the starting pitcher, with value being added based on expected pitch count, bullpen, value of the hitters and then a number that reduces value based on opposing hitters.
 

Member
Handicapper
Joined
Mar 7, 2005
Messages
8,849
Tokens
Here's a problem that i can see with it.

The "ranking" is a self-fulfulling prophecy. I'm no math expert, but i realize that as you compile more data you are in effect changing the stats as you go along. In other words, the worse a team gets, the worse it's record gets, thus the lower the #. The last thing I would want to do is martingale chase a loosing team. While using the data as a snapshot, or picture in time, may be worth doing, I don't know that the team will magnet back towards the centerline. Look at a bell curve. There is a middle that happens most often but the ends of the curve are there and the possibility of something happen.

IMO the best handicappers rely 100% on math and nothing on feelings so I really like the road you're traveling on. I just don't think the magnet is the road we are looking for. I would argue that whatever system is going to be used will need to be based on the starting pitcher, with value being added based on expected pitch count, bullpen, value of the hitters and then a number that reduces value based on opposing hitters.

I appreciate the comment...but I'm not really interested in arguing about the validity of the formula. I believe that has already been proven.

Does that mean we can use it to beat the line...probably not...but I'm convinced we may learn something by trying.

You can't find gold unless you pan for it.

I'm not sure what you are getting at with the self fulfilling prophesy part.

It can be said that if a team is far away from their expected wins or losses...it represents "luck"...I think its reasonable to wager against that continuing for a long amount of time.

Here is some more on the theory...perhaps this will shed more light.

=============

Pythagorean expectation is a formula invented by Bill James to estimate how many games a baseball team "should" have won based on the number of runs they scored and allowed. Comparing a team's actual and Pythagorean winning percentage can be used to evaluate how lucky that team was (by examining the variation between the two winning percentages). The term is derived from the formula's resemblance to the Pythagorean theorem.
The basic formula is:
<DL><DD>
43c198866d56f1677288399f3908934e.png
</DD></DL>where Win% is the winning percentage generated by the formula. The expected number of wins would be the expected winning percentage multiplied by the number of games played.
 

Fah-New-Gee
Joined
Nov 27, 2007
Messages
3,023
Tokens
The self fulfilling part is this.

Game 1 - team scores 10 runs.
Game 2 - team scores 1 run.
Game 3 - team scores 1 run.
Game 4 - team scores 1 run.
Game 5 - team scores 1 run.

If you have a really bad team, the "trend" of scoring 1 run then continues.

If you look at the stats after 1 game, you think this is a good team avg 10 runs.
If you look at the stats after 2 games, you think still good team - average 5.5 runs.
If you look after 5 games, now their "average" is runs per game. If you started playing them at game 2 you would have thought they were a good team - in reality they are not.

Here's another way to look at it.

In my business (a collection agency), we compete against other agencies for a liquidation rate - basically how much we collect, and we're judged versus the "average" liquidation rate of the portfolio.

Month 1 - we collect 2%
Month 1 - competition 8%
Month 1 - average 5%


Month 2 - we collect 4%
Month 2 - competition 8%
Month 2 - average 6%


Month 3 - we collect 7%
Month 3 - competition 8%
Month 3 - average 7.5%

While we've collected a lot more in month 2, by collecting more, we raise the bar of the average performance of the portfolio. We think, heck we're a lot closer to the average of 5% - but in reality the average liquidation of the porftolio is 6% since we raise the bar. We then end up being less than the average. Month 3 rolls on and we beat the "average" of Month 1 and Month 2 - we think we've done great - yet since the bar is raise so much we are again "below the average" and we lose the portfolio. The only way to be above the average is to beat the competition.

Same thing with looking at these #'s. Again, this is not from a math major.

And as for chasing the gold, I think I will ALWAYS be looking for a system to beat the man. It's out there. We just have to find it. The last few years I had it in bases. Not much of a win but a win none the less. This year the formula hasn't worked. I don't feel bad because I am in very good company with the results of Tomorrow's Newspaper and a bunch of other cappers on the forum that just aren't producing they way they used to.

The holy grail has changed and morphed. So not only do we have to find the formula we have to change it as the book changes.
 

Member
Handicapper
Joined
Mar 7, 2005
Messages
8,849
Tokens
The self fulfilling part is this.

Game 1 - team scores 10 runs.
Game 2 - team scores 1 run.
Game 3 - team scores 1 run.
Game 4 - team scores 1 run.
Game 5 - team scores 1 run.

If you have a really bad team, the "trend" of scoring 1 run then continues.

If you look at the stats after 1 game, you think this is a good team avg 10 runs.
If you look at the stats after 2 games, you think still good team - average 5.5 runs.
If you look after 5 games, now their "average" is runs per game. If you started playing them at game 2 you would have thought they were a good team - in reality they are not.

Here's another way to look at it.

In my business (a collection agency), we compete against other agencies for a liquidation rate - basically how much we collect, and we're judged versus the "average" liquidation rate of the portfolio.

Month 1 - we collect 2%
Month 1 - competition 8%
Month 1 - average 5%


Month 2 - we collect 4%
Month 2 - competition 8%
Month 2 - average 6%


Month 3 - we collect 7%
Month 3 - competition 8%
Month 3 - average 7.5%

While we've collected a lot more in month 2, by collecting more, we raise the bar of the average performance of the portfolio. We think, heck we're a lot closer to the average of 5% - but in reality the average liquidation of the porftolio is 6% since we raise the bar. We then end up being less than the average. Month 3 rolls on and we beat the "average" of Month 1 and Month 2 - we think we've done great - yet since the bar is raise so much we are again "below the average" and we lose the portfolio. The only way to be above the average is to beat the competition.

Same thing with looking at these #'s. Again, this is not from a math major.

And as for chasing the gold, I think I will ALWAYS be looking for a system to beat the man. It's out there. We just have to find it. The last few years I had it in bases. Not much of a win but a win none the less. This year the formula hasn't worked. I don't feel bad because I am in very good company with the results of Tomorrow's Newspaper and a bunch of other cappers on the forum that just aren't producing they way they used to.

The holy grail has changed and morphed. So not only do we have to find the formula we have to change it as the book changes.

It seems to me the formula accounts for what you are talking about.

If a team is in a long trend of underperforming or one run games as you say...then the expected wins will also be lower according to the formula based on runs scored...pulling it toward what is expected...or the baseline. This would be a situation where the team would be doing what is expected...and not be on our radar.

If their is a variance in that ...for whatever reason...we will attempt to exploit it. The theory would be they won or lost more than "expected" based on "luck"...and the fundamentals of the formula would tend to pull it back to the baseline.

It's already been well documented that this formula does in fact predict records based on runs scored and given up...over time.

Again...if there is a variance in that...its reasonable to expect bad or good luck will not continue for a prolonged period of time.

But there will be exceptions and outliers...of course.
 

"Straight Cash Homie"
Joined
Nov 12, 2006
Messages
1,125
Tokens
Great Stuff. MJ, not just picks, but actually theories. One thing that will need to be studied is how far from the predicted record teams actually move away from the expected record before returning. I think if we look at previous years which I did a little of earlier we can see how far some teams travel from the norm. If most teams never go over or under say 5 games then once that number is reached then the your expected magnet effect should pull the team back to a norm.
 

New member
Joined
Oct 30, 2007
Messages
24
Tokens
Magnet Picks Watch List

LA Angels +5
Atlanta -5
Clev -4
Tampa +3
LA Dodger -3
Det -3<!-- / message -->

When a team gets too far away from the expected win/loss number...in theory...it should get pulled back...like a "Magnet" to what is expected according to the formula


I guess the question is, how far away is too far? I'm feeling like 5 games isn't that far away. BTW, i LOVE your other thread. Thanks for that. And I'm not against you at all on this, just putting my two cents in.
 

Forum statistics

Threads
1,106,773
Messages
13,438,917
Members
99,339
Latest member
billcunninghamhomeloans
The RX is the sports betting industry's leading information portal for bonuses, picks, and sportsbook reviews. Find the best deals offered by a sportsbook in your state and browse our free picks section.FacebookTwitterInstagramContact Usforum@therx.com