Is my model flawed?

TomG · 10-15-12, 11:18 AM

What is the probability of you going 2-5 in NFL this week given your model is not flawed?

If you can't answer this then yes your model is flawed.

hutennis · 10-15-12, 01:23 PM

Great point!

I hope OP will try to answer just as an exercise, if for nothing else.

big0mar · 10-15-12, 04:51 PM

Originally posted by TomG

What is the probability of you going 2-5 in NFL this week given your model is not flawed?

If you can't answer this then yes your model is flawed.

17%.

My ultimate point here is that I beat the closing number - sometimes by a good margin - on almost all the games I lost. Theoretically, can a model be broken even if its regularly beating closing numbers?

buby74 · 10-15-12, 05:42 PM

Big0mar 17% is very high thats 1 in 6 do you check if a die is loaded if it comes up 6 the first time you throw it? you cannot tell anything from 7 games come back when you are 20-50 despite beating theclosing line and then we will have something to sink our teeth into

sayhey69 · 10-15-12, 06:31 PM

check line 10 min before closing. taiwanese steam changes everything

big0mar · 10-15-12, 08:51 PM

I had Denver +3 tonight!

clarkd32 · 10-15-12, 09:31 PM

hard to tell if i can't see your model. send it to me and i'll get back to you...

chunk · 10-15-12, 09:38 PM

Originally posted by big0mar

I had Denver +3 tonight!

Denver +3 was the correct play

hutennis · 10-15-12, 10:22 PM

Originally posted by clarkd32

hard to tell if i can't see your model. send it to me and i'll get back to you...

Just curious, how do you plan to evaluate his model without knowing in a great details specifics of the model he competes against?
What if data and methods he is using are already part and parcel of that competing model?

Or maybe it is not a concern at all?
If so, why?

HUY · 10-18-12, 02:33 PM

Beating the closing line is overrated. There was a study that said that most of the time line movements are due to new information, which means that beating them or not beating them means jack shit.

The study went on to say that when the line movement was not due to new information, it had no predictive power. I will try to find the study, but I don't have it handy at the moment.

VLR100 · 10-18-12, 04:01 PM

Learn to do hypothesis testing. Anyway with such a small sample, drawing conclusions is meaningless.

hutennis · 10-18-12, 04:30 PM

Originally posted by HUY

Beating the closing line is overrated. There was a study that said that most of the time line movements are due to new information, which means that beating them or not beating them means jack shit.

The study went on to say that when the line movement was not due to new information, it had no predictive power. I will try to find the study, but I don't have it handy at the moment.

Study is about beating the line that moves, not a CL, and I agree, considering the line movement to be some kind of opportunity is most likely a mistake, since line moves on new info most of the time and thus every new line essentially becomes a new opener and there is no advantages in taking new line vs previous line.
But the CL is basically the end of the road for line development. CL incorporates in itself all available information. There is nothing more to consider pre game thus making CL most efficient.

Beating most efficient fair line just has to be beneficial.

chunk · 10-18-12, 04:35 PM

Originally posted by HUY

Beating the closing line is overrated. There was a study that said that most of the time line movements are due to new information, which means that beating them or not beating them means jack shit.

The study went on to say that when the line movement was not due to new information, it had no predictive power. I will try to find the study, but I don't have it handy at the moment.

Totally agree. This ongoing phobia @ beating the closer is B.S. I don't doubt that sayhey is correct about Taiwanese steam coming late either...with maybe a little afghani and Egyptian steam in the mix.