I'm working on a system for MLB, its in its final stages

[FellatioReceiver]

These results so far as staggering... its spitting out winners at 71.7% so far, with the average line at -133.

I am still not fully convinced yet... I would like to backtest it further before I start releasing plays. Does anyone know of a good site that shows historical categorized data?

I need more than just the final score of the game, I need to see lines and line movements.

Thanks and I am close to starting the release of these plays soon for those who are interested.

I estimate this thing should hit at around 65%.

[brettd]

make sure you're not back fitting your model mate.... been there done that

[HUY]

These results so far as staggering... its spitting out winners at 71.7% so far, with the average line at -133.

I am still not fully convinced yet... I would like to backtest it further before I start releasing plays. Does anyone know of a good site that shows historical categorized data?

I need more than just the final score of the game, I need to see lines and line movements.

Thanks and I am close to starting the release of these plays soon for those who are interested.

I estimate this thing should hit at around 65%.

If you measure the performance over matches whose result you already know then 71.7% is actually very low... I can tell you the results of matches that have already happened with 100% accuracy right now, without developing a model.

In short, look into what is in SBR called "data mining error" or, in statistical terminology, "overfitting".

[FellatioReceiver]

71.7% games I am predicting them before the game starts. Obviously anyone can pick 100% after the game is over, what are you saying?

[Juret]

He means that you can't test your model on those games that you used for developing the model.

[thom321]

To add even more and probably redundant detail to what Juret (nice avatar btw) et al said, if you developed your model using data up until e.g. 2009 and when testing it on data for 2010 and 2011 season, you hit 71.7% with an avg line of -133 for those two seasons, that might be promising. Whatever your hit rate was using data prior to 2010 in this scenario isn't relevant other than as an indication that doing further out of sample testing might/might not be warranted.

[HUY]

There is no way you are getting 71%, this is definitely overfitting. Try to remedy this otherwise you're in for a very rough landing once you start betting it.

[mathdotcom]

There are probably 950 threads in HTT almost identical to this one, despite J7's sticky at the top

[bulichm]

covers.com

[MonkeyF0cker]

Thanks and I am close to starting the release of these plays soon for those who are interested.

I estimate this thing should hit at around 65%.

LOL.

Good luck with that.

[Inkwell77]

To add even more and probably redundant detail to what Juret (nice avatar btw) et al said, if you developed your model using data up until e.g. 2009 and when testing it on data for 2010 and 2011 season, you hit 71.7% with an avg line of -133 for those two seasons, that might be promising. Whatever your hit rate was using data prior to 2010 in this scenario isn't relevant other than as an indication that doing further out of sample testing might/might not be warranted.

interesting

[GunShard]

When it comes to new systems, always bet with play points on Covers.com instead of risking real money.

Always experiment with fake money.

[a4u2fear]

How many games and years is this over? If its not multiple years and 1000s of games, it's crap

[big0mar]

Mods: pls change thread title to "early stages"

[FellatioReceiver]

This thing is good to go. Ready to launch ThursdayIt is 67% ytd... this thing rocks.

[HoulihansTX]

Sounds like a chase SYSTEM.

Awesome.

[FellatioReceiver]

Not a chase system, I hate those chase systems.

This thing spits out well over 60% winners, Ill show you.

[JMon]

How many games and years is this over? If its not multiple years and 1000s of games, it's crap

Was going to ask this same question, yet IMO minimum 1000 games is not necessary to substantiate.

[FellatioReceiver]

1300 picks

[koz-man]

looking forward

[Dad]

[EVPlus]

To the OP. I wish you luck. However, you will need to excuse the seasoned vets here if none of us hold our breath.

Sure, a minority can profit in this crazy meat grinder business - but no one I know hits 70 percent winners in a season.

[JMon]

^, I would be damn happy with 55%

[FellatioReceiver]

Are you guys serious?

Wow we are about to have some surprised people around here.

55%? 55% wont even net a profit betting MLB

[aznbluff]

Are you guys serious?

Wow we are about to have some surprised people around here.

55%? 55% wont even net a profit betting MLB

You dont need to only bet favorites....

[Frank-King]

Prove everybody wrong

[FellatioReceiver]

This system does not like many dogs. Its not a pet lover.

But it is a winner.

[cmaulsby]

Looking forward to a Friday roll-out.

[matt1216]

dont take this personally but most Good Cappers can hit 60-65% i know some good ones up around 70-75%. just cap the games if its only gonna hit 65% if u can sh me a system thats 75-80% then your on to something.
I need more than just the final score of the game, I need to see lines and line movements

[JMon]

Are you guys serious?

Wow we are about to have some surprised people around here.

55%? 55% wont even net a profit betting MLB

, oh dear

[2daBank]

this guy acts like 55% aint shit..WE ALL GONNA BE RICH!!!!!!!!!!!!!!!!!!!!!!

[FellatioReceiver]

I realize everyone here is their own aspiring God here and a professional. Everyone knows best, thats why this site is full of millionaires I suppose.

In any case, Units won result is the number that matters in the end. Talk means nothing.

[TwoWays]

Yawn. Whatcha doin, just pick some games already. Stop hyping your system and let the results create the buzz.

[FellatioReceiver]

For Starters I was on the NL tonight, posted it in the All Star game thread. +115 was nice monies.

[2daBank]

I realize everyone here is their own aspiring God here and a professional. Everyone knows best, thats why this site is full of millionaires I suppose.

In any case, Units won result is the number that matters in the end. Talk means nothing.

yet that is all you have done...