[Ganchrow]

ok what does this mean? why not just:

.334/.0871= 3.835

Sorry, just force of habit. They're the same, obviously.

8.71% (8.71% + 33.40%) is the probability of drawing team A along with Home.

The reciprocal, 8.71% (8.71% + 33.40%) are the fair decimal odds on this.

Subtracting 1 gives you fair fractional odds ... approximately 3.835.

[acw]

Much of it comes down to the refs being influenced by the crowd.

I read about this before and decided to make use of refs and the crowd in all sorts of way in my equations hoping to find some refs being more biased than others. It turned out to be totally insignificant; they all have the same bias or they are not impacted by them. Just the crowd itself does have some impact. Not that much though.

[pico]

ganch, how do i type math equations like you? you should add some math buttons.

[Justin7]

How about a different, simpler approach. Forget complex formulas.

Assume that the homefield advantage is worth X points. Use straightforward conversions. So if a 10-point favorite wins 80 % of the time on a neutral field, it becomes a 7-point favorite on the road. Conversions (for NFL) = 72% win rate.

Going the other way... a 10-point favorite becomes a 13 point favorite, and its win rate goes up to 85% (if you assume the conversion is accurate).

Perhaps you can do the conversion with a formula, but it seems much easier to do it. I don't know that you lose any accuracy this way either. I've tried a lot of ways to model for HFA, and I have never found anything better than a straight spread adjustment (although the adjustment changes substantially at the end of the year).

[Wheell]

How about a different, simpler approach. Forget complex formulas.

Assume that the homefield advantage is worth X points. Use straightforward conversions. So if a 10-point favorite wins 80 % of the time on a neutral field, it becomes a 7-point favorite on the road. Conversions (for NFL) = 72% win rate.

Going the other way... a 10-point favorite becomes a 13 point favorite, and its win rate goes up to 85% (if you assume the conversion is accurate).

Perhaps you can do the conversion with a formula, but it seems much easier to do it. I don't know that you lose any accuracy this way either. I've tried a lot of ways to model for HFA, and I have never found anything better than a straight spread adjustment (although the adjustment changes substantially at the end of the year).

And Justin scores! So then, what would one put the lines at of a game that was +600 at a neutral court?

[Justin7]

And Justin scores! So then, what would one put the lines at of a game that was +600 at a neutral court?

With NFL? Assuming a normal total (call it 42):
A true line of -600/+600 is about -12.5.

At the dog...-9.5, call it -380 novig
At the fav... -15.5. Call it -850.

A caveat though - conversion charts aren't really good for big road favorites. A 10-pt fav wins at home a somewhat more than on the road (both in NFL, NBA and NCAAF, never studied NCAAB). Most conversion charts don't adjust for that... so I might have to drop the dog price down to -350.

[acw]

A caveat though - conversion charts aren't really good for big road favorites. A 10-pt fav wins at home a somewhat more than on the road

But the 10 points are still correct? If so this would indicate that away favourites when they win, will win by larger margins. My initial thoughts in European football where exactly the opposite, so I did some testing on that. As far as I can say in European football there is totally no bias between a home or away team being the favourite in their winning chances.

[Wheell]

And to think, this was originally a basketball discussion. Still, well put Justin.

[Degenerate]

Thanks for the info Ganch, Wheel & Data

It seems for teams close to 50/50 on a neutral field the HFA is 11%, given the lg ave is 61%

HFA is smaller, the greater the disparity between teams.

If we estimated HFA for each game across a season, then averaged it, wouldn't the HFA always be below the 11% lg average?

[Data]

Thanks for the info Ganch, Wheel & Data

It seems for teams close to 50/50 on a neutral field the HFA is 11%, given the lg ave is 61%

HFA is smaller, the greater the disparity between teams.

If we estimated HFA for each game across a season, then averaged it, wouldn't the HFA always be below the 11% lg average?

As a commonly accepted term, HFA is a league wide constant and in your post it is 61%. It is NOT 11% as you say. If you want to discuss other aspects of this then you should use different wording or give your definition of HFA as you use it. This is the same thing I asked Wheell earlier in this thread. Without this, it is unclear what do you mean.

[Degenerate]

Over the coarse of a season, if teams win at 61% at home, & we'd expect them to win 50% of the time on a neutral field, then the average advantage gained from playing at home is 11%.


so basically its est win % at home - est win % on neutral field

It seems to me like the formulas earlier in the thread would undervalue HFA.
In a league that Home teams historically win 61% of the time, the formulas would have the home team winning more like 57-58% of the time.

[Data]

so basically its est win % at home - est win % on neutral field

That number is meaningless and as such useless.

[Formulawiz]

Guys I find your complex math very interesting but I am not sure how you would apply your math to predict the outcome of future team performances. I am not sure what exactly your trying to prove here. If we look at last nights WAS St Vs ORG ST game how would you apply your math to make a prediction. No pun intended.