Idea to Balance Championship Divisions

[EricH]

I propose that they just go back to what they used previously.

Let the divisions be letters A, B, C, D, E, and F.
Let the teams that register be 1, 2, 3, ..., n, sorting by numerical order.

1 goes in A, 2 goes in B, 3 goes in C...6 goes in F, 7 goes in A, 8 goes in B...n goes in the next-in-sequence of A-F.

Assign A-F to random division names.

The problem (or not) with this method is that anybody can generate a division list with a few minutes and some programming knowledge, and knowing which teams are signed up.

[Citrus Dad]

Originally Posted by AGPapa

I don't think "most people" would say that, and the OP definitely didn't. He said "improve the Championship experience for everyone", not just the few teams that make Einstein. Many teams want to play with the best in the world, not watch them compete in another division.

I agree. The problem with division stacking is not who wins the title, but rather who gets the opportunity to play for the title. The best example is this year comparing Curie to Newton. There were teams in Newton left off playoff alliances that would have played on Saturday in Curie if divisions were more balanced. And the Curie division winning alliance clearly wasn't up to the level of the other division alliances on Einstein because it didn't have a deeper pool of teams to draw from.

I too like the segmenting approach to assigning. It's really only a variation on what FIRST is already doing.

[AdamHeard]

Originally Posted by EricH

I propose that they just go back to what they used previously.

Let the divisions be letters A, B, C, D, E, and F.
Let the teams that register be 1, 2, 3, ..., n, sorting by numerical order.

1 goes in A, 2 goes in B, 3 goes in C...6 goes in F, 7 goes in A, 8 goes in B...n goes in the next-in-sequence of A-F.

Assign A-F to random division names.

The problem (or not) with this method is that anybody can generate a division list with a few minutes and some programming knowledge, and knowing which teams are signed up.

You can fix that by grabbing the next 6 teams, randomizing them, then assigning them A-F.

[Cory]

Originally Posted by Taylor

I vaguely remember, perhaps a decade ago, teams were accused of gunning for particular awards and registering for particular regionals because they believed it would get them in a "favorable" CMP division. Does anybody else recall this?
The reason I bring this up is setting the precedent for teams being accused of colluding before the season to end up in the same CMP division, if they knew and could outsmart the placement algorithm.

This is the first I've heard of it. To the best of my knowledge at no point in the last 17 years has what regional you qualified at or what qualifying award you won had anything to do with your division assignment.

[IKE]

Originally Posted by Cory

This is the first I've heard of it. To the best of my knowledge at no point in the last 17 years has what regional you qualified at or what qualifying award you won had anything to do with your division assignment.

The only correlation I am aware of is that by using the registration timing mechanism, and then assigning them to each division, and the deadlines after winning an event, you would likely get a scenario where the teams that won an early event together would not be in the same division.

[AGPapa]

Originally Posted by Chris is me

Anyways, we have all of this information - anyone invested in this idea could do some programming and spit out 8 hypothetical divisions this year sorted by this algorithm, and then analyze these divisions using different metrics (OPR, etc) to see how unbalanced they are. I'd love to see this sort of analysis.

I generated some division lists based on what was outlined in the OP. The pools were slightly different though (mostly because I got the qualification types from a different project I've been working on).

The pools are:

Winner AC/1
Wildcard AC/1
District Points
CA/EI
RAS
Prequalified
Waitlist
Winner 2/B & Wildcard 2/B

Note that I used first qualification method, rather than the "highest", and that DCMP winners are in "District Points" pool because FIRST doesn't give DCMP winners automatic qualification.
I've attached an OPR chart for the different divisions, along with the actual distribution last year. The balanced distribution looks slightly better, but there is still a division that's clearly above everyone else.

[M217]

Just wanted to say I find it a little strange that the OP placed DCMP Winner Second Pick/Backup in Category II while he put Regional Winner Second Pick/Backup in Category V. I understand that obviously District Champs will have significantly deeper fields than Regionals, but is it really this big of a difference? Seems kinda extreme...

[Hitchhiker 42]

Originally Posted by M217

Just wnt it to say I find it a little strange that the OP placed DCMP Winner Second Pick/Backup in Category II while he put Regional Winner Second Pick/Backup in Category V. I understand that obviously District Champs will have significantly deeper fields than Regionals, but is it really this big of a difference? Seems kinda extreme...

Originally Posted by AGPapa

...and that DCMP winners are in "District Points" pool because FIRST doesn't give DCMP winners automatic qualification.

Problem fixed.

[Brian Maher]

Originally Posted by M217

Just wanted to say I find it a little strange that the OP placed DCMP Winner Second Pick/Backup in Category II while he put Regional Winner Second Pick/Backup in Category V. I understand that obviously District Champs will have significantly deeper fields than Regionals, but is it really this big of a difference? Seems kinda extreme...

There are three reasons behind this decision:

1) The fact that number of district points earned by these robots almost always puts them above the cutoff to qualify as a District Points team

2) Average OPR from this year's Championship shows that the DCMP Winner 2nd Pick / Backups are considerably stronger than those of regionals:

• DCMP Winner Pick 2 / Backup: 36.6
• Regional Winner Pick 2 / Backup: 28.3

Note: 2nd picks and backups were lumped together because I felt the sample sizes (1 for DCMP winners and 1 for Regional winners) for backups were too small to evaluate individually.

3) Having attended or watched footage from a number of regionals and District Championships, anecdotally, I have seen considerably higher performance from the winning second picks at DCMPs than at regionals. This makes sense when the reasoning for why this may be is considered.
The barrier for entry to a DCMP is way higher than that of a regional: strong performance at District events (or District Chairman's) versus nothing at all.
In addition, the winning alliance at a regional represents the 30-65 teams attending that regional. The winning alliance at a District Championship represents the 50-400 teams in that district.

Originally Posted by AGPapa

because FIRST doesn't give DCMP winners automatic qualification.

While this is generally a non-issue due to the impact of the 90 playoff round performance points most DCMP winners receive, that is technically false.

Originally Posted by Section 7.4.4, FRC Admin Manual 2016

These slots will be made up of qualifying award winners from the District Championship (Chairman’s Award, Engineering Award, and Rookie All Star winners), all teams on the Winning Alliance from the District Championship (including any Back-Up teams participating), any qualifying award winners among district teams who traveled to Regionals, plus the teams on the final District ranking list, as deep in the ranking list as the District needs to go to fill their allocation.

[AGPapa]

Originally Posted by BMSOTM

While this is generally a non-issue due to the impact of the 90 playoff round performance points most DCMP winners receive, that is technically false.

Hmm, I was looking at this document which contains a table that does not state that DCMP winners automatically qualify. Although looking at it now, it also seems to leave out Regional Winners and Wildcards! I'm going to assume that this table is in error and that the Admin Manual is correct.