To elaborate on our evaluation of the Gonzalez results from earlier in the week, this article will focus on the Perrone System. The System has a quality win component that uses the opponent’s Win Percentage to dictate how much Perrone Points a team receives.
I’m throwing a lot of numbers at you, but that’s just the exact thing that happens. Think of it this way: if one defeats an undefeated team, the most points are awarded. If you defeat a team with 0 wins, you get a tinier amount of points. Flip that around for the loser: losing against an undefeated team won’t be so bad. Giving a team their first win will hurt the losing team.
- GVSU W% = 0.500 and 2.625 points
- MSU W% = 1.000 and 5.000 points
GVSU def MSU 2-1 OT
GVSU gains 4.000 to 6.625
MSU loses 1.500, but gains 0.950 for losing in overtime. MSU loses 0.550 to 4.450
The key here is MSU’s undefeated posture at this point in the Season. GVSU gets the most points for defeating an undefeated team, that is a team with a 1.000 W%. Further, MSU’s loss to a team with their 0.500 W% leads to a loss of only 1.500 points.
In the Perrone System, overtimes give the losing team a slight rebate. Here, MSU gains .950 on their initial loss of 1.500 points because they took GVSU to overtime. So in the end MSU only loses 0.550 points.
MSU def GVSU 2-0
MSU gains 3.250 to 7.700
GVSU loses 1.500 to 5.1250
Here, the winning team has played against a team with a different W% than our first result. The W% is 0.600, falling within the 3.250 point award. Perrone awards range from 2.5-4.0. Losses range from 1.25-2, and this all depends on the initial W% of either team going into the match.
Takeaways:
Rolling System: Both the Perrone and the Gonzalez system are rolling systems (at least when they are viewed in the spreadsheet). As matches progress, points are awarded or taken based on the W% of the team before the match. At this point in development, the point a match is played has a minor impact, i.e. W% can fluctuate during a multi-match tournament. This consideration is represented in the Perrone Award range of 2.5-4.0.
Inherent Consequences of Starting from Game 1: Further, losses at the start of the season (or counting period) may be more impactful than they need to be. That could be nulled over time. The Gonzalez system goes back to the start of accurate records, 2010-2011, and the system adjusts over time to reflect the current strength of a team. (A team can change their stars in as little as 8 games) The same could be applied here, possibly even combining the benefits of a quality win component into the Gonzalez rating exchange concept. This would involve a Rating Exchange modifier that fluctuates based on the W% of the two teams. Most likely that part limited to what… the last 20 games? or just the current Season?
Overtime: At current, the Perrone System doesn’t give or take anything special for the winning team in an OT. The losing team gets a rebate on the points they lost, but nothing affects the winning team. Here, GVSU gets a lot of points for going into overtime and defeating an undefeated team. Some tests with an OT against two undefeated teams may be required.
Hopes: I saw the results of this day as a fairly decent example to show how the ranking systems work using real world results. I even managed to do some further debugging. Learning never stops. We’ve put a lot of thought and effort into these systems, and the learning never stops.
If anyone has any questions, let them be known! Thanks for reading.
One more thing… I’ve been naming these systems after the players/officers that introduced them.
