Team Statistics Explanations

From Dave Sebek - OSSCA Web Team and Statistic Lead

First off, we do not use any of those numbers for any sort of auto-ranking.
In fact they are something that I used on my own personal site (www.highschoolsoccerohio.com) and the OSSCA guys ported over to their site when we combined data(bases).

I'll do my best to explain what they mean.

Pyth - is a teams pythagorean score
E(Pts) - is the expected # of points a team should have (based on their Pythagorean score)
D(Pts) - is the difference between the # of points they actually have and their expected points.

Some things to keep in mind for purposes of this discussion:
Each game is an event worth 3 points.
The total possible points a team can gain is 3 * the number of games played.

To calculate the pythagorean score, you need to know the following:
Goals Scored by your team (GF)
Goals Conceded by your team (GA)

Pythagorean score = (GF * GF) / ((GF*GF) + (GA*GA))

To calculate the E(Pts) or expected points, take the total number of games played * 3 points/game * the Pythagorean score. Round to get a whole number.
To calculate the D(Pts) take the actual points earned - E(Pts)

An example at this point may be helpful, and for that we'll look at my team, the Olentangy Braves to date, we are 4-2-2 and have scored 22 goals and given up 13 so our calculations are as follows POINTS = 14 (where points = 4 wins * 3 points/win + 2 ties * 1 point/tie = 14)
GF = 22
GA = 13
PYTHAGOREAN = 0.7411 based on (22*22)/((22*22) + (13*13)) = 0.7411
E(Pts) = 18 based on 8 games * 3points/game * Pythagorean = 18
D(Pts) = -4 based on Actual Pts - E(Pts) = 14 - 18 = -4

What does it all mean?
It means that based on their results, my team could be expected to "win"
@74% of the points that they have played for. 74% of the points they've played for is @18. We have "won" 14 points for a difference of -4. The fun part is looking at why? In my case its pretty simple, our one win (8-0 over Big Walnut) skews are number pretty drastically.

What are the flaws or things to keep in mind?
1 - Every high school team is not created equal and every high school schedule is not created equal. If I have a good team and play in a good league/play a tough schedule, I'll most likely have a lower pythagorean score than a good team that plays in a weak league or plays a weak schedule.
2 - Incomplete results can definitely effect the pythagorean score (or your combined opponents pythagorean score)
3 - Scoring more goals and giving up less goals results in a higher pythagorean score. Period. Does that mean running up the score would result in a higher pythagorean score? in the long run, yes? in the short run a 1-0 win is the same as a 13-0 win. Both give a PYTH = 1.000. but as you can see from my results, add one 8-0 win over the course of a season and you can skew your pythagorean score.

How do I recommend you use that page?
I look at two things, how a team has done and how their opponents have done. I believe that in addition to your pythagorean score, your opponents combined pythagorean scores should be displayed. Check out that number as well. In our case, our opponents Pythagorean score is 0.7560 and their Pythagorean score for their adjusted record (their combined record against everyone but me) is 0.8345.

So I would conclude that: we've done pretty well and we've done pretty well against good opponents.

Hope this helps. Feel free to respond to webmaster@ossca.org with followup questions and I'll answer as best I can.

Thanks
-- Dave

[close this window]