[jakeandba]

Have a technical question regarding betting.

Lets say I handicap baseball games and found out when teams without their last names on the back of their uniforms play other teams without their last name on their uniforms go over the game total 55% of the time.
when betting the over:
assuming the over/under will always be -110 then i would have a positive EV correct?

55%x100=55
45%x110=49.5

rounding up then would be a +5. Can you confirm this is correct.

A followup question is this, lets say i find 2 more baseball handicapping systems that hit 55% or higher(for example betting the over for the first 5 innings when it is raining and bet the home team total over during day games). The odds offered for these 2 additional systems are at -110. All three are a positive EV correct? would you bet all three or just one ?

[Data]

Can you confirm this is correct.

No, I cannot. Any angle based on past ATS performance is useless.

[jakeandba]

perhaps an amendment is needed to the statement. In 2000 a hypothesis is tested regarding over/unders. The research shows that when two teams play each other with last names not printed on the uniform, the over hits 55 or higher.
The next yr the same hypothesis is tested to validate the results of 2000. Validation testing is done for 9 more years. Each year the results are the same, results are at 55% or better.

with the above information, is the EV a positive number?

[Indecent]

perhaps an amendment is needed to the statement. In 2000 a hypothesis is tested regarding over/unders. The research shows that when two teams play each other with last names not printed on the uniform, the over hits 55 or higher.
The next yr the same hypothesis is tested to validate the results of 2000. Validation testing is done for 9 more years. Each year the results are the same, results are at 55% or better.

with the above information, is the EV a positive number?

No. See previous answer.

[jakeandba]

Then i am confused, what is the purpose of creating a hypoethesis and then validation. what information is modeling based on?

[Data]

Then i am confused, what is the purpose of creating a hypoethesis and then validation. what information is modeling based on?

The model takes the line into account but it must not be based on it.

[Blax0r]

Your EV calculation is correct, but I think these guys are trying to tell you that your methodology isn't right (ie, despite the historical lines patterns, it still may not be positive EV to make that bet now).

[CrimsonQueen]

This... is the craziest thing I've ever read. Teams with no names printed on the backs of their jerseys make them score more runs??? Why, because the lack of fiber on their backs allows them to run faster?!

Pirates have been trending towards non-existence from hundreds of years ago, and the Earth has been getting warmer during that same time. Obviously then, pirates make the Earth cooler.

[jakeandba]

ok thanks for the comments. What I have read(outside sources) so far could be summarized:
An observation is made regarding an event. Hypothesis is made and then validated. This is done by back testing. Including in validation is the introduction of odds to see if this observation was profitable. Now I understand all this work is no guarantee this observation is profitable for future games yet to be played. Also is to remember is basic mistakes such as sample size and relationship error. Sure there are a few others i am forgetting. With that being said models and or angles have to be based on basis of past performance. If not then models might as well be based on air. After all with any scientific approach to observation, the results have to be validated. Meaning anyone and their dog should be able to follow the methodology and confirm or disprove said observation. If this observation has been confirmed over and over, with no exceptions being found then it is considered a law.

It would be interesting to determine if a law had been found regarding wagering...anyway back to the point, models/observations are based on an idea, this idea is tested against the past. From this point, the ultimate test is for future games yet to be played. If playing/following this model/angel would yield a positive EV then this has to be considered an edge. This can only be "proven" by tracking bets after these future games are played.

Thanks again for all comments

[jakeandba]

This... is the craziest thing I've ever read. Teams with no names printed on the backs of their jerseys make them score more runs??? Why, because the lack of fiber on their backs allows them to run faster?!

Pirates have been trending towards non-existence from hundreds of years ago, and the Earth has been getting warmer during that same time. Obviously then, pirates make the Earth cooler.

Perhaps i should have highlighted the first sentence that said...say I handicap baseball and found said observation...that was an example no where was it mentioned that this is actually the case....next time i guess i need to say teams that players wear caps play teams that dont wear caps when out in the field.

the whole point of the paragraph was about the EV, the observation was added as background and not meant to be taken seriously

[Blax0r]

Your EV calculation is correct, but I think these guys are trying to tell you that your methodology isn't right (ie, despite the historical lines patterns, it still may not be positive EV to make that bet now).

My view still hasn't changed; in addition, the answer to your latter questions is yes. If you do find an identical situation (in terms of odds and true win probability), the bet should still remain +EV.

[CFA]

The model takes the line into account but it must not be based on it.

Somewhere out there, Bill the data miner is laughing at you.

[Shonner]

Have a technical question regarding betting.

Lets say I handicap baseball games and found out when teams without their last names on the back of their uniforms play other teams without their last name on their uniforms go over the game total 55% of the time.
when betting the over:
assuming the over/under will always be -110 then i would have a positive EV correct?

55%x100=55
45%x110=49.5

rounding up then would be a +5. Can you confirm this is correct.

A followup question is this, lets say i find 2 more baseball handicapping systems that hit 55% or higher(for example betting the over for the first 5 innings when it is raining and bet the home team total over during day games). The odds offered for these 2 additional systems are at -110. All three are a positive EV correct? would you bet all three or just one ?

I think what the guys are trying to say is that if you search hard enough for games that hit "x" percent, you can find whatever you are looking for. Data mining will do that for you. It takes much more than "x" teams cover "x" % of time to create a successful model.

It's very hard to create a successful model, if you listen to the sharp posters around here they say less than 1% of the models out there are done correctly and actually produce +EV results in the long run.

Take a look at the thread about beating the closing line if you really want to see how to get +EV, if you believe in efficient market that is. The thread gets wild but there are a lot of sharp posters giving very solid feedback there.

[Shonner]

Have a technical question regarding betting.

Lets say I handicap baseball games and found out when teams without their last names on the back of their uniforms play other teams without their last name on their uniforms go over the game total 55% of the time.
when betting the over:
assuming the over/under will always be -110 then i would have a positive EV correct?

55%x100=55
45%x110=49.5

rounding up then would be a +5. Can you confirm this is correct.

A followup question is this, lets say i find 2 more baseball handicapping systems that hit 55% or higher(for example betting the over for the first 5 innings when it is raining and bet the home team total over during day games). The odds offered for these 2 additional systems are at -110. All three are a positive EV correct? would you bet all three or just one ?

In theory, if your model was correct and produced consistently in the long run, then yeah, you would have +EV.

[jakeandba]

thanks all for the time to comment, it has been helpful

[mathdotcom]

Sure it is EV if you have a large enough sample, are not relying on extremely dated data, etc.

Of course if you are serious about this example and not just using it as a metaphor, you should be cautious as there is no reason to expect nameless jerseys to score more runs..

While as Data says any angle based on ATS performance is going to fail, this is only in practice. In theory there is no reason why it must fail. If bookies never took into account home advantage then you'd notice home teams seem to cover more often than 52% and it would be +EV to bet them. In practice all these angles are known, and despite what most squares think, everyone knows (for example) that wind conditions affect Wrigley field totals.

[Data]

While as Data says any angle based on ATS performance is going to fail

I did not say this but rather implied that any angle is likely to fail in the future. Not neccessarily it will. I said "useless" as it may fail at any moment. It is useless due to uncertainty even while the angle is still profitable (!). See more in this thread: http://www.sportsbookreview.com/forum/handicappe...kely-lose.html.