[Alex Elliott]

So for part 2 of this series, I figured it would be good to spend some time answering a rather burning question from Nationals:

How much better is your team if you have pros?

First off, we need to define how to measure the impact of a given roster. To do this, I borrowed a page from the pro soccer stats folks and decided to use point differential for each game.

In order to get data for the strength of each roster, I looked at each roster from this year's Nationals and counted up the number of each APPA ranking level for that roster (e.g. Pro, D1, D2 etc).

I also tracked the size of each roster just to see if that had any effect as well.

Now that I had the data, for each roster I created a model that compared:
-Difference in roster size
e.g. if you have 10 players and they have 9 then you have a +1 and they would have -1 for when they played you.

-Difference in players at each ranking
e.g. If you have 1 Pro and they have 2 then you get a -1 and they get a +1 and so one.

So what do the numbers say?

Well here is how it breaks down:

Size  0.83
Pro   1.51
D1    0.44
D2   -0.19
D3    0.67
D4   -0.39
D5   -2.25
Beg  -1.27

NOTE: only about 70% of the variation is attributed to the above.

In English:
-For each additional player you have over your opponent you can expect to score about .8 more points per match.
-Each additional pro player you have over your opponent you can expect to score about 1.5 more points per match
-Each additional D5 player you have over your opponent equals about -2.25 points

So, does having Pro players help you win more points per match: the answer would seem to be yes.

However, there are two key things:
1. Nebraska's pro advantage would seem to be offset by the fact that on average they played teams with 6-7 more players than they did.

In other words, while they gained ~4.5 points due to having pros(3 pros * 1.5 points per pro) they also would have lost almost 5 points (-6 players * .83) due to the differences in the roster sizes.

So how did they win so many games?

2. If you take a look at the numbers above again, you'll notice each additional D5 player you have leads to -2.25 points!

In absolute terms, that's roughly 150% the effect that a pro player has.

If Nebraska is playing lots of teams made up mostly of D4/D5/Beg players, that would account for their huge points margins.

To summarize, having pros does help. That being said, having a team of D5 rated players hurts you more than having pros helps.

To put it another way, Nebraska was exceptionally lucky to have multiple pro players go to their school aka they won the player lottery. If your school has no D1/D2/D3 players then you could consider yourself rather unlucky aka you lost the player lottery.

Up next, something similar but will include some more school related data e.g. school size.

[njballer44]

My Coach Is A Genius!!!!!!

[UMDpaintball]

[tchalien]

interesting statistics... It also follows logic, the more higher ranked players you have, the better you should do. There is still the "unknown" factor of players that are "underranked" example, I have great players at LU that have not played anything but college ball. They have the same skill if not higher than some of my other players that have played PSP for many years.

[Alex Elliott]

Originally Posted by tchalien

interesting statistics... It also follows logic, the more higher ranked players you have, the better you should do. There is still the "unknown" factor of players that are "underranked" example, I have great players at LU that have not played anything but college ball. They have the same skill if not higher than some of my other players that have played PSP for many years.

Very true.

Another point, some of the schools do substantially better or worse than the model would predict. Below is the list of schools and the average number of points per match that they score higher or lower than the model would predict:

Team	Avg
-1.66	Northeastern University Huskies
-1.59	Purdue University Boilermakers
-0.85	University of Maryland Terrapins
-0.63	Florida Atlantic University Owls
-0.23	University of Florida Gators
-0.20	University of Nebraska - Omaha
-0.08	University of Connecticut Huskies
-0.03	University of Tennessee Volunteers
 0.03	Ohio University Bobcats 
 0.07	West Point Black Knights
 0.53	Rutgers University Scarlet Knights
 0.68	Cal State University-Long Beach 49ers
 1.65	Kennesaw State University Owls
 2.21	Liberty University Flames

There are a couple of potential reasons for this:
a. Under ranked players as pointed out above.
b. Coaching e.g. a good coach can make up for inexperience and an inexperienced coach can negatively impact a good team with bad point concessions
c. Fitness
d. Weather effects e.g. Northern teams playing in hot weather etc
e. etc etc

[tchalien]

Originally Posted by Alex Elliott

Very true.

Another point, some of the schools do substantially better or worse than the model would predict. Below is the list of schools and the average number of points per match that they score higher or lower than the model would predict:

Team	Avg
-1.66	Northeastern University Huskies
-1.59	Purdue University Boilermakers
-0.85	University of Maryland Terrapins
-0.63	Florida Atlantic University Owls
-0.23	University of Florida Gators
-0.20	University of Nebraska - Omaha
-0.08	University of Connecticut Huskies
-0.03	University of Tennessee Volunteers
 0.03	Ohio University Bobcats 
 0.07	West Point Black Knights
 0.53	Rutgers University Scarlet Knights
 0.68	Cal State University-Long Beach 49ers
 1.65	Kennesaw State University Owls
 2.21	Liberty University Flames

There are a couple of potential reasons for this:
a. Under ranked players as pointed out above.
b. Coaching e.g. a good coach can make up for inexperience and an inexperienced coach can negatively impact a good team with bad point concessions
c. Fitness
d. Weather effects e.g. Northern teams playing in hot weather etc
e. etc etc

I like this stat...

[jakes78]

While I don't enjoy the pro/no pros in collegiate paintball debate, I'm very impressed by your pretty sound statistical approach. Your next project with overall student populations should be real interesting. Thank you for the good work and read.

[SARGE37]

Originally Posted by tchalien

interesting statistics... It also follows logic, the more higher ranked players you have, the better you should do. There is still the "unknown" factor of players that are "underranked" example, I have great players at LU that have not played anything but college ball. They have the same skill if not higher than some of my other players that have played PSP for many years.