Middling NFL props?

**oveedee** · 10-10-09, 06:07 AM

Hard to say, it is not a clear pattern here. Sure, some general laws apply, but there is always the surprise factor. I guess that what matters the most is how the player is playing the current season because that's the most relative factor, but many things can change..

**Peeig** · 10-10-09, 06:53 AM

If you find a middle that big, I would jump on it.........I don't know the exact edge, but it is a good bet.

**rm18** · 10-10-09, 08:19 AM

i do not find many of these, sometimes I find where like I bet the same guy to go under in receptions and over in yards and try to hit both or get a win and push

**stats13** · 10-10-09, 08:37 AM

Originally posted by rm18

i do not find many of these, sometimes I find where like I bet the same guy to go under in receptions and over in yards and try to hit both or get a win and push

rm, check your personal email. i need routing info

**Ominous** · 10-10-09, 11:14 AM

I wrote a program to calculate poission value for middling.

Here is my other play
Jones rushing
O50.5 yards -117
U60.5 yards -117

Assumption here is 4.1 yards/rush attempt

This graph computed EV as a function of lambda(Expected number of rush attempts)

http://data.fuskbugg.se/skalman01/g3.png

Lambda should be around the lines posted by the books which makes this bet look like a good one to me, with a EV of about 15%. what you think?

Looking good?

what worries me is that the EV seems to "good to be true"

**rookie** · 10-11-09, 04:30 AM

One of the conditions to be met for poisson distribution is that the number of opportunities for the event occurring need to be very large relative to the likelihood of the event. This is not satisfied in case of pass completions or rushing attempts for main RB(s).

**Ominous** · 10-11-09, 07:57 AM

Originally posted by rookie

One of the conditions to be met for poisson distribution is that the number of opportunities for the event occurring need to be very large relative to the likelihood of the event. This is not satisfied in case of pass completions or rushing attempts for main RB(s).

Hmm yeah, its obviously not ideal but question is if it is "good" enough.

Would you try to use binominal distribution instead?

It has 2 indata parameters as far as I understand, n attempts and p probability of success per attempt. Problem here is that it will be far harder to guess these parameters than in the case of poission where you can assume that <the expected number of occerences> is close to the actual line offered by the book. Any suggestions here?

**durito** · 10-11-09, 09:05 AM

the limits are way too low

**rookie** · 10-11-09, 10:01 AM

Originally posted by Ominous

Problem here is that it will be far harder to guess these parameters than in the case of poission where you can assume that <the expected number of occerences> is close to the actual line offered by the book. Any suggestions here?

Dangerous assumption. Especially when we are talking about those sportsbooks who will not put up lines unless they flash on their Don Best screens. Some of them put up props for a week or two but then discontinue. The limits are so low and juice is high. All indicators of lack of belief in their own numbers. I will not rely on them.

**Thremp** · 10-12-09, 01:47 PM

Originally posted by durito

the limits are way too low

I've never been one to scorn the effort of bending down to pick up $20.

**durito** · 10-12-09, 09:51 PM

Originally posted by Thremp

I've never been one to scorn the effort of bending down to pick up $20.

yea but you should be able to figure out the +ev side in <1 minute. Probably worth the extra effort.

**Frank** · 10-18-09, 06:57 PM

Isn't poisson distribution only valid if things happen in increments of one?

If thats the case, attempts would work but not yardage.

**shantystar** · 10-26-09, 03:51 AM

Originally posted by Ominous

Is it profitable to middle NFL props? When I use poission to examine them I sometimes find rediculous values like 20% value etc.

Lets say I have Sanchez passing
O175,5 yards -117
U199,5 yards -117

Do you like to bet this blindly on poission analysis?
With an avarage yards/passing attempt for this player of 6,75 y/Att the lines will converts to
O26 Attempts
U29,5 Attempts

If the expected number of Attempts is anywhere between 21 and 37 then this bet will show a positive EV according to poission analysis. With expected attempts between in the range 26-30 the bet shows a EV of 17% according to poission both legs taken into account. This seems almost to good to be true.

What do you guys think?

Do I make some logical error with this analysis? Which avarage is best to use to convert yards to attempts?(player this season, player previous season, team avg, NFL avg? not that +-25% on the avg would matter that much)

Are there any bookie specific interpetations of rules that can cause obvious problems here?

Edit:I recalculated using completed passes instead of attempted passes by assuming 60% pass completetion. This lowered the lines to O15,5 and U17,5 but it stills shows an EV of 8-10% in reasonable ranges of expected completed passes, and positive EV in the range of 12-22 expected completed passes

good example to understand pro members here.