Evidence that scoring margin matters

[InstiGATOR1]

My subject line is the title of this Ken Pomeroy essay:

http://kenpom.com/blog/index.php/web...margin_matters

In it he says:

There are several other interesting pieces of information. For instance, it takes a 15-point home win to get a team to the point where you’d expect them to win the road. Is home-court advantage that strong? No, it’s not. Keep in mind this method is going to use conference games almost exclusively and in most situations conference foes are somewhat equally matched. Cases of a team winning (or losing) by 20 or more points are often cases where the winning (or losing) team has overachieved (or underachieved). They weren’t truly 20 points than their opponent to begin with, and this is manifested in the fact that the return game results, on average, in a small win.

(This notion applies to Florida pretty well right now. Yes, I think they are very good, deserving of strong consideration as the best team in the country if you are into ranking such things that way. But unless they are one of the greatest teams in history, they have also overachieved their true ability by a fair amount to this point.)

The essay shows a very high variance and thus gives UF fans some cause for worry about the game at UGa.

[gatorbogey]

yeah, i think alot of fans - and it's probably only natural - will look at a great game and expect that performance 'all the time'....when it's probably an out-lyer data point. and do the same for a poorly performed game - and think that's the norm for the team.

my own sports experience - w/ golf: i know not to do this (assume the over-achievement is now the norm) and yet, i still fall into that trap w/ gator sports. i never assumed that, after firing a 68, that that score now became my 'norm'...i knew it to be 'playing out of my mind', you know. and likely, i'd have a bad round the next time out just to average things out. and really, after a 68, even a 73 "felt" bad - and when you sense a bit of a loss of momentum, well, the score can then spiral out of control and get even worse....probably happens in team sports too. case in point was the sugar bowl for the gators. a single bad play begets even worse play due to the mental aspects.

[your_perfect_enemy]

I don't expect 25+ point wins from here on out, but from the 4 SEC games I watched this team play we were that much better

[ETGator1]

Did I read it right in that we have a 73.7% chance of winning by 7.6 points based on our 33 point win at home? If so, I like our chances better than UGALY's 26.3% chance of winning as the home team.

I don't expect us to win big in Athens. Just look at last year as to what can happen. We won by 22 at home and then lost by 14 in Athens in a game we should have won close.

This is no time for the team to relax. On average, we are in for a tough 7 to 8 point win if we play our best.

Florida 65
Georgia 58

[VTGator]

I think one crucial factor is whether the scoring margin is due primarily to good offense or good defense. It is much more difficult to sustain a high octane offense every game, especially on the road, but a good, disciplined defensive team should be much more persistent.

For instance, we gave up 76 points to UG last year on the road. 76 points was high but not exceptionally so last year. By the time of the 5th SEC conference game we had given up 68+ points in 7 games. This year we have none. Last year opponent scoring average was 65. This year it is 51.7. I would be surprised if we give up 76 points to any team this year, let alone to a weak SEC team scoring 60 ppg.

[tommyuf21]

UF were struggling at that point in the season when we lost to the Poodles. Young was not playing well and they let their frustration with themselves and each other effect their defense.

I think its safe to say that the team at a much better spot this year and I'll be very suprised if they have a stinker tonight.

[ETGator1]

Wasn't Will Yeguete injured and out for the 2nd game last year?

[tommyuf21]

He was hurt in the Auburn game, just a couple of days before the Poodle travesty.

http://scores.espn.go.com/ncb/recap?gameId=320520057

[trufloridagator]

Great thread, variance discussion itt is good

Sent from my iPhone using GatorCountry

[themistocles]

Every team has on-games and off-games - Big Surprise.

You can't use point differences, other than MASSIVE ones, for anything worthwhile, because they frequently are not an indication of how close a game really was.

The home court advantage is very, very important, but it is far more important for relatively inexperienced teams than for more experienced teams.

Believe it or not, on any given night/day, almost any team can upset a considerably "better" team if things fall their way. Things include, but are not limited to: Referees, LUCK, Biorythmical influences (people have major performance ups and downs based on where they are in their cycle), momentum shifts in a specific game, whether a player or more than one has a hot or cold hand, whether the head coach has a stomach ache or not, etc., etc..

Basketball is about as complex a game as exists from the perspective of changes, motions, reactions, and interactions.

Everything, and I mean EVERYTHING VARIES, and frequently the variation is very, very large.

Nothing new here for an empirical statistician.

[GatorRade]

Originally Posted by themistocles

You can't use point differences, other than MASSIVE ones, for anything worthwhile, because they frequently are not an indication of how close a game really was.

I think I agree with what you are saying, them, but this line intrigues me. If the score doesn't give an indication of how close a game is, what can you use to determine this?

[Jonas]

You'd probably need the time history of the margin throughout the game. Maybe the average of such a thing would give a useful metric.

[swampgas44]

Originally Posted by GatorRade

I think I agree with what you are saying, them, but this line intrigues me. If the score doesn't give an indication of how close a game is, what can you use to determine this?

Our pre-mentrual cycles?