Who came up with "home field = +3 point advantage?" (1 Viewer)

[Anarchy99]

Because that's not what the numbers show. Someone must have pulled this one out of a hat. Even some of the experts seem confused by this concept, as Jeff Sagarin uses +2.49 as the home field advantage differential (which is too low) in his weekly team comparison and expected results breakdown.

For argument's sake, let's just use 3 points as the HFA, as it has been that way for as long as I can remember. In doing so, that would essentially mean the home team would be expected to score 1.5 points per game more than a neutral site game and the visiting team to score 1.5 points less than a neutral site to equal a net of 3 points. So if two teams were considered "even," then TEAM A would be a 3 point favorite at their place and TEAM B would be a 3 point favorite at their stadium

However, I ran the numbers for every regular season game over the past 5 seasons (I did not recalibrate for neutral site games due to weather relocations, playing at a stadium that was not truly a home game, or playing in London). Taken as a whole, the league average in net scoring differential home vs. road was 5.31 points. And I hesitate to want to use a "league average," as some teams have played way better at home and way worse on the road. Here's what the numbers showed.

To explain what the numbers represent, the first column is the net point differential for all home games. Over their past 40 home games, GB had a scoring differential of +13.53 points per game. In their last 40 road games, their scoring differential was only +0.80 points per game. So the net difference is 12.73 ppg when they play at home vs. on the road.

To illustrate, ARI has been terrible on the road. So if the Cards played the Packers in GB, the combination of Green Bay's home dominance and Arizona's weak road performance would not be a simple 3 point advantage for the Pack as has been the HFA that has become commonplace since the dawn of time. Instead, based on the on-field performances of both teams, it would be +13.53 for GB at home and +7.78 for ARI playing on the road = 21.31 points.

And unlike the old school mentality, if they played in Arizona, the three point advantage would not just switch from GB to ARI. Arizona would have as a baseline +2.38 points for playing at home and -0.80 for GB playing on the road, so ARI would end up +1.58 playing at home, which obviously is a HUGE difference compared to playing on the frozen tundra at Lambeau.

Home Away NetGBP 13.53 0.80 12.73ARI 2.38 -7.78 10.16BUF 0.98 -8.23 9.21BAL 8.55 -0.35 8.90NOS 9.03 0.45 8.58SFO 8.68 0.18 8.50SEA 9.25 0.80 8.45DET 4.78 -3.05 7.83IND 3.55 -4.13 7.68SDC 6.00 -1.10 7.10NYJ 0.50 -6.45 6.95PIT 7.70 0.80 6.90DEN 7.33 0.65 6.68MIN -0.33 -5.98 5.65CIN 4.50 -0.38 4.88NYG 2.65 -2.23 4.88STL -2.03 -6.68 4.65ATL 4.58 0.28 4.30OAK -4.35 -8.60 4.25NEP 12.90 8.68 4.22KCC 0.75 -3.33 4.08CLE -2.70 -6.10 3.40TBB -3.75 -6.83 3.08TEN -2.45 -5.53 3.08HOU 2.23 -0.63 2.86DAL 2.20 -0.25 2.45PHI 2.30 0.20 2.10CHI 1.05 -1.00 2.05CAR -1.10 -2.68 1.58MIA -0.25 -1.65 1.40JAC -8.40 -9.25 0.85WAS -4.55 -5.13 0.58Obviously there are a lot of things inherently wrong with this breakdown and analysis, first and foremost the fact that the roster and competitiveness of teams change over time and what happened 5 seasons ago is not really relevant today. Using 2014 teams and using data from 2010 as a basis for analysis is not a great comparative tool. However, that still doesn't change the baseline numbers that in most situations a simple HFA of +3 seems archaic and mathematically incorrect.

Again as an illustrative model only, using only raw data and no tweaks for 2014 values, injuries, weather conditions, strength of schedule, etc. . . .

ARI @ CAR would be expected to play out to CAR +6.68 (just played, CAR won by 11)

DET @ DAL would be expected to play out DAL +5.25 (just played, DAL won by 4)

CIN @ IND would be expected to play out IND +3.93 (just played, IND won by 16)

BAL @ PIT would be expected to play out PIT +8.05 (just played, BAL won by 13)

For the weekend upcoming . . .

BAL @ NE = NE +13.25

IND @ DEN = DEN +11.46

DAL @ GB = GB + 13.78

CAR @ SEA = SEA +11.93

 

[ghostguy123]

Because the people who set the lines dont care what the outcome will be, they want equal action on both sides, and that seems to be a pretty accurate rule of thumb for them.

 

[Marauder]

I could be wrong but I've always understood the +3 to be vs. a Neutral site, so the point spread difference between home vs. road is really +6.

 

[ghostguy123]

I could be wrong but I've always understood the +3 to be vs. a Neutral site, so the point spread difference between home vs. road is really +6.

this is true.

 

[Dinsy Ejotuz]

I could be wrong but I've always understood the +3 to be vs. a Neutral site, so the point spread difference between home vs. road is really +6.

Which is pretty close to his 5.31 differential in the OP.

 

[Anarchy99]

I could be wrong but I've always understood the +3 to be vs. a Neutral site, so the point spread difference between home vs. road is really +6.

Which is pretty close to his 5.31 differential in the OP.

3 + 3 = 6.

5.31 + 5.31 = 10.62.

And factoring in that some teams are higher or lower and COMBINATIONS are all over the map, it is not as simple as 3 points.

 

[shadyridr]

common people bet on the game so they set the line for common people not mathemeticians.

 

[ghostguy123]

common people bet on the game so they set the line for common people not mathemeticians.

this again

they dont care if some people win every game they bet. They care if they get equal wagering on both sides.

 

[Marauder]

I could be wrong but I've always understood the +3 to be vs. a Neutral site, so the point spread difference between home vs. road is really +6.

Which is pretty close to his 5.31 differential in the OP.

3 + 3 = 6.

5.31 + 5.31 = 10.62.

That's not what you are saying in your OP:

"However, I ran the numbers for every regular season game over the past 5 seasons (I did not recalibrate for neutral site games due to weather relocations, playing at a stadium that was not truly a home game, or playing in London). Taken as a whole, the league average in net scoring differential home vs. road was 5.31 points."

Anyway, I seriously doubt that the Vegas linemakers are actually still using a flat +3.

 

[ZWK]

I could be wrong but I've always understood the +3 to be vs. a Neutral site, so the point spread difference between home vs. road is really +6.

Which is pretty close to his 5.31 differential in the OP.

3 + 3 = 6.

5.31 + 5.31 = 10.62.

And factoring in that some teams are higher or lower and COMBINATIONS are all over the map, it is not as simple as 3 points.

If CIN is playing at home against an evenly matched team, the conventional rule would put them as 3 point favorites (+3).

If CIN is playing on the road against an evenly matched team, the conventional rule would put them at -3.

Net, for the 2 games against evenly matched teams, would be +6.

So that's +3 at home, -3 on the road, +6 net.

In your table, the average team is +2.67 at home, -2.64 on the road, +5.31 net.

That's slightly less of a home field advantage than the conventional +3, -3, +6 (and slightly more than Sagarin's +2.49, -2.49, +4.98).

 

[Anarchy99]

I could be wrong but I've always understood the +3 to be vs. a Neutral site, so the point spread difference between home vs. road is really +6.

Which is pretty close to his 5.31 differential in the OP.

3 + 3 = 6.

5.31 + 5.31 = 10.62.

That's not what you are saying in your OP:

"However, I ran the numbers for every regular season game over the past 5 seasons (I did not recalibrate for neutral site games due to weather relocations, playing at a stadium that was not truly a home game, or playing in London). Taken as a whole, the league average in net scoring differential home vs. road was 5.31 points."

Anyway, I seriously doubt that the Vegas linemakers are actually still using a flat +3.

Yes, net average difference per team is 5.31 points. Two teams in a game - 10.62 point differential. Just like 3 points swings either way, so does the 5.31 points.

 

[Dinsy Ejotuz]

David, what am I missing?

Team A is a +3 favorite at Home, and a -3 dog Away (traditionally, assuming the two teams are evenly matched). That's a 6-point difference.

And you're saying that the average difference vs Home and Away is 5.31 points. Why are you doubling it?

 

[ghostguy123]

Anyway, I seriously doubt that the Vegas linemakers are actually still using a flat +3.

No, they are not. And never did. Each line they set is unique, but the +/-3 is just a generalization

 

[squistion]

Because that's not what the numbers show. Someone must have pulled this one out of a hat. Even some of the experts seem confused by this concept, as Jeff Sagarin uses +2.49 as the home field advantage differential (which is too low) in his weekly team comparison and expected results breakdown.

For argument's sake, let's just use 3 points as the HFA, as it has been that way for as long as I can remember.

It has been that way as long as I can remember, too, and using the fallible Occam's Razor, I think it was just the assumption that home field advantage was worth roughly a field goal (in the days when low scoring matchups were the norm rather than the exception).

 

[IHEARTFF]

3 is the generic number and is close to the avg HFA. Teams with better advantages such as GB get a larger HFA than say, Jax when making lines.

 

[Dinsy Ejotuz]

I could be wrong but I've always understood the +3 to be vs. a Neutral site, so the point spread difference between home vs. road is really +6.

Which is pretty close to his 5.31 differential in the OP.

3 + 3 = 6. 5.31 + 5.31 = 10.62.

That's not what you are saying in your OP:"However, I ran the numbers for every regular season game over the past 5 seasons (I did not recalibrate for neutral site games due to weather relocations, playing at a stadium that was not truly a home game, or playing in London). Taken as a whole, the league average in net scoring differential home vs. road was 5.31 points."

Anyway, I seriously doubt that the Vegas linemakers are actually still using a flat +3.

Yes, net average difference per team is 5.31 points. Two teams in a game - 10.62 point differential. Just like 3 points swings either way, so does the 5.31 points.

If you're going to do it that way, you have to double the six points too. Home team is +6 vs being away, and away team is -6 vs being at home.

 

[Stinger Ray]

I believe it comes down to this....if both teams are considered even by Vegas the away team gets +3 in the betting.

 

[Leroy Hoard]

I could be wrong but I've always understood the +3 to be vs. a Neutral site, so the point spread difference between home vs. road is really +6.

Which is pretty close to his 5.31 differential in the OP.

3 + 3 = 6. 5.31 + 5.31 = 10.62.

That's not what you are saying in your OP:"However, I ran the numbers for every regular season game over the past 5 seasons (I did not recalibrate for neutral site games due to weather relocations, playing at a stadium that was not truly a home game, or playing in London). Taken as a whole, the league average in net scoring differential home vs. road was 5.31 points."

Anyway, I seriously doubt that the Vegas linemakers are actually still using a flat +3.

Yes, net average difference per team is 5.31 points. Two teams in a game - 10.62 point differential. Just like 3 points swings either way, so does the 5.31 points.

If you're going to do it that way, you have to double the six points too. Home team is +6 vs being away, and away team is -6 vs being at home.

You are pretty close with this. I have seen what is closer to -5 (-2.5 each way) but 6 is very much in the area.

 

[Warrior]

Two thing:

All teams do NOT get the +3 treatment. It depends on the team and the stadium, etc.

Your sample size is way too small. 5 seasons? Try 15.

+3 is just a rule of thumb, not something that can be or is followed in every circumstance.

 

[Anarchy99]

Your sample size is way too small. 5 seasons? Try 15.

That's part of the problem. The game has changed so much that more recent seasons are more indicative of how things are given the offensive explosion in recent years, the two hand touch rules for defense, and the turnover in rosters due to free agency.

With more games now 41-28 compared to 17-14, intuitively the range in scoring and HFA should be much different in 2014 vs. 1994 vs. 1974.

 

[GregR]

Let's say only 1 NFL game has been played in all of history to create our model from. GB at home beat ARI by 5. Our model ends up with:

GB is a +5 at home.

ARI is a -5 away.

Our model's best prediction for a rematch in GB should be the same 5 point difference that happened before. But the way you are adding both team's differential, you would predict a 10 point margin.

That's what people like wdcrob are talking about that you're doubling the differential. The points for each team's home differential are also reflected in another team's away differential. If you just add them together straight up, you're effectively doubling them.

 

[trader jake]

This thread is yet another example of why Vegas takes home so much filthy lucre each year.

 

[DropKick]

Agree the Vegas odds are strictly designed to get half the betting on each side.

Generally speaking, if two evenly matched teams played at a neutral site, there would be no spread. If they played in one team's home stadium, that team would be favored by 3. There is a six point swing in the spread when you compare these two scenarios (home vs away).

To illustrate the six point swing, if team A scores 17 points, how many points must B score to at least tie ("push")?

14 if B is the visitor (14 + 3 = 17)

20 if B is at home (20 - 3 = 17)

Per your numbers "the league average in net scoring differential home vs. road was 5.31 points". I don't think you double the 5.31.... You're not talking average points scored at home vs. the road - you're already looking at a "net" metric which includes the home teams plus performance and the visitors negative one.

So 4.98 or 5.31 or 6.0 seems pretty close to me given the purpose of a line... Maybe the key point is that - depending on the teams actually playing - there is a mathematical opportunity to exploit a poor line.

 

[DropKick]

So I understand your metrics...

In a grossly simplified two game season between the two teams in my Mom's fantasy league

Home Team A defeats B 20-14 team A wins by 6 (and vice versa)

Home Team B defeats A 20-14 team A loses by 6 (and vice versa)

People might conclude the home team should win by 6...

The league average in net scoring differential home vs. road is 12? Does this already include a "double count"?

 

[CalBear]

common people bet on the game so they set the line for common people not mathemeticians.

this again

they dont care if some people win every game they bet. They care if they get equal wagering on both sides.

That's just not true. Vegas will take whatever advantage it can get, and it's more than happy to exploit inefficiencies in the way people bet. To suggest otherwise is to suggest that either 1) Vegas is unable to identify inefficiencies in the way people bet, or 2) Vegas can identify inefficiencies, but decides not to exploit them.

 

[Nero]

If two teams are considered even, then it should produce an overtime game which would likely end in a FG. I think that is where the 3 points to the home teams is derived from. As stated above, the Vegas betting line has nothing to do with the hypothetical three points to the home team. Its about betting action.