Frank Answers Fridays: Expanded Championship Qualification

[cgmv123]

http://www3.usfirst.org/roboticsprog...idays-07252014

Today’s good question comes from Adam Salem, from FRC Team 3419, The Rohawks, out of New York City, New York, USA:

Question:

Hi Frank,

My name is Adam Salem and I am a mentor on team 3419. I read your note about increasing the size of nationals and think it's a terrific idea. I have a suggestion about how to select some of the additional teams. One of the awards for engineering the robot (either the Excellence in Engineering Award, the Innovation in Control Award, the Industrial Design Award, or some combination of those) should come with an invitation to nationals, in the same way the Engineering Inspiration award does. I think this would help to solve two major problems I see transpiring with FRC:

• The popularity of the "Robot in 3 Days" videos has led a lot of teams to build very similar looking robots (how many loaders did we see this year which were almost identical to each other?). These videos have taken a lot of the desire to innovate out of the students. I think this generation of students - who have had things like Google at their fingertips for most of their lives - are used to immediate answers and don't see much value in researching a problem themselves when another solution is promptly available. By having an Innovation in Engineering award that gets you to the finals, it would encourage students to think more out-of-the-box rather than taking the easy way out and copying what they see on the videos.
  
• Every year my team faces at least one design decision wherein one option would build a robot more likely to win the game, and the other option would present a more interesting engineering challenge. The students always push for the former, as they are focused on winning. I always push for the later, as I am focused on the educational and fun nature of the build process. I know other teams struggle with the same thing. By allowing teams to get to nationals via an innovative robot, I think you'd see a lot more unique and interesting robots get built. In the extra 200 robots at nationals, you would see a way more diverse set of designs. One of my favorite parts of every tournament is walking around with my students and talking about the interesting designs other teams have and seeing how they solved the same challenges we experienced in different ways. By encouraging a wider variety of robots, it would make this experience more enjoyable and educational for everyone.

Thanks,

Adam Salem

Mentor, Team 3419

SVP Of Technology, AllianceBernstein

Answer:

Thanks for the question, Adam.

We currently have an eleven-person task force working on the question of FIRST Championship FRC eligibility for 2015 and beyond. Four of the task force members are volunteers, and the balance are FIRST staffers, both from HQ and the field. We plan to be able to announce the eligibility criteria to teams by the end of August.

The system used in 2014, under the 400 team FRC limit, was ‘almost broken’. With only five waitlist slots awarded, we were very close to being overbooked for the FIRST Championship. While it would seem that the significant increase in slots we’re looking at for the 2015 Championship - exact number still TBD – will solve this problem to the point where we would actually have too few teams attending Championship on merit, this isn’t necessarily the case, at least longer term. Districts get a percentage allocation to the FIRST Championship based on their representation in FRC. So, if a given District has 10 percent of all FRC teams, they get 10 percent of the slots (excluding pre-qualified slots) at Championship. This means that as we increase the number of teams we can host at Championship (and add Districts!), the number from Districts increases as well.

Under the existing system, we typically qualify 6 or so teams for Championship from each Regional. In practice, it doesn’t average exactly 6, because we occasionally have multiple award winners – who use only one slot - or 4 team winning alliances if back-up robots are called. If I understand you correctly, you are proposing an additional award that would raise that number to about 7. Based on current estimates, and some educated guesses, while we could accommodate this approach for 2015, in 2016 we would be close to being overbooked again, even assuming we have 600 teams at Championship. I appreciate your interest in encouraging design diversity, but I think it unlikely the task force would be taking this specific approach. We are looking for a solution that will serve us comfortably for several years.

Thanks again for your question.

Frank

[PayneTrain]

inb4 Gregor comments on "nationals"

[AGPapa]

Originally Posted by Frank

So, if a given District has 10 percent of all FRC teams, they get 10 percent of the slots (excluding pre-qualified slots) at Championship.

This is rather interesting. Last year Michigan had about 10% of all FRC teams. If this continues next year then they will have about 60 slots at Championship. Only 4 or 5 of the teams at the Michigan State Championship will not qualify for St. Louis. At this point the state championship is essentially meaningless in deciding who advances. In fact, many teams can completely skip MSC and still qualify for World's. This seems to go against the point of the state championship.

How can Michigan solve this problem? Increase the size of MSC by moving to a new venue? Adding divisions like World's?

Jim Zondag addressed this issue a few years ago in his whitepaper.

Originally Posted by Jim Zondag

Q30: “What will you do when the State Championship is no longer large enough to house all of the good teams in the State?”
A30: This will be a good problem to have . We are still many years away from reaching any real limits in this regard. That said, if our population of teams continues to grow at its current rate, the threshold to make it to the State Championship will continue to rise. When we launched the District System in 2009, 1/2 of our team population advanced to the State Championship. After 4 years, this percentage is down to 1/3rd. By 2016, we project this percentage to fall to below 1/4th. Entry will keep getting harder, and the competition will continue to increase. There are over 700 public high schools in our state plus hundreds more private schools. If we grow large enough to reach Dean Kamen’s goal of a team in every school, then perhaps someday we will need to create a second Regional Championship. We will cross this bridge when we come to it.

Well, I think we've arrived at the bridge, now what?

[Jon Stratis]

Yes, having so many teams make it from a district that the championship is worthless is a problem... but perhaps more of a problem is for the areas that aren't districts.

Minnesota, for example, is probably going to have about 200 teams next year. As a state, we have 4 regionals... which means 24 slots at champs. Sure, a few teams travel to out of state regionals, but even more teams from out of state come to our regionals and win slots. What it ends up meaning is that, as a state, the number of teams we send to champs is essentially a constant, and not comparable to the number districts send, as a proportion to FIRST population. We don't grow as champs grows - if we want more MN teams to make it to champs, we have to hold more events. The same isn't true for the districts, which creates, in my opinion, a severe imbalance across FIRST.

FIRST needs to come up with a system whereby they can support both the district and regional model while allowing proportional representation from every distinct area. Picture something like drawing up district lines across all of FIRST for Champs participation, but the method of entry for each individual region could be different - The district model could use the point system with a district championship, while the regional model could pull X teams from each regional, based on the number of teams needed for that area (with each area designed to have a minimum of 6 teams attending). For the regional, you could use a point system for the event, or base it on awards, or whatever.

[Lil' Lavery]

Why does FIRST have to ensure proportional representation from all areas?

[Nate Laverdure]

Originally Posted by Lil' Lavery

Why does FIRST have to ensure proportional representation from all areas?

Of course it doesn't. But it's a way of saying that, to a first-order approximation, all teams have a roughly equal opportunity to advance to CMP. This is part of what makes a compelling competition.

[dodar]

Originally Posted by Nate Laverdure

Of course it doesn't. But it's a way of saying that, to a first-order approximation, all teams have a roughly equal opportunity to advance to CMP. This is part of what makes a compelling competition.

Almost no where in FIRST is an equal distribution of CMP slots to number of teams. Florida for instance, there are 12 total possible slots for CMP and yet Florida hold roughly 4% of FIRST; this would be 17-18 slots of CMP. Florida actually sent something like 10-12 teams to CMP this year too.

[MARS_James]

Originally Posted by Lil' Lavery

Why does FIRST have to ensure proportional representation from all areas?

It really doesn't but with a potential increase of 200 teams I would like it to. If FIRST told Florida that we had to use the District point system to determine who were the highest ranked teams in our state not attending (according to my math to be represented beyond our Regional guarantee spots, we would get 2 more slots.), 744 and 1251 would have been able to attend, personally speaking I feel they were more than championship caliber bots with 744 giving us a tough finals game in South Florida and our twin 1251 was just as tough as we were and arguably the better catcher.

But personally I would rather there be some new determining factor that is fair to all teams other than who can click a mouse the fastest.

Originally Posted by dodar

Almost no where in FIRST is an equal distribution of CMP slots to number of teams. Florida for instance, there are 12 total possible slots for CMP and yet Florida hold roughly 4% of FIRST; this would be 17-18 slots of CMP. Florida actually sent something like 10-12 teams to CMP this year too.

We are actually more like 2.5% of all teams so one could argue previously we were over represented

[Justin Montois]

I'm not sure why a FIRST-wide points system based on the standard district model points system wouldn't work....

Top 600 teams in the world go to champs.

[MARS_James]

Originally Posted by Justin Montois

I'm not sure why a FIRST-wide points system based on the standard district model points system wouldn't work....

Top 600 teams in the world go to champs.

Because of Chairman's, Engineering Inspiration, and Rookie All Star, plus it would mean waiting until the last possible week for all 600 teams to make their plans to go to Champs.

EDIT: Though if that idea would have been applied last year Florida would have been represented by 180, 179, 744*, 1592, 1251*, 108, 79, and 233*, meaning 1902, 3932, 4013, 5145 and 5196 wouldn't have made it (All who won the awards above except 3932 who was the last pick of the South Florida Winning Alliance)

*-did not qualify

[DonRotolo]

Hm. Something is vaguely unsatisfying with the comment about 'a district with 10% of the teams gets 10% of the slots'. That's the case now (more or less), but I'm just not following the logic to expand that to 600 slots, guaranteeing a near-overbooking situation every year. And the absurdity of sending 60 teams from MI.*

I think it's me.

I do understand the logic of equal representation though. Either you run districts, which get a proportional representation (and rewards you for growing FRC), or you run regionals, which gets you (about) 6 slots per regional (and rewards you for being able to fund more regionals).

I don't envy the task force.


*Not that they're not welcome, but as AGPapa pointed out, the MSC becomes pointless for everyone but 4-5 teams.

[Jon Stratis]

FIRST doesn't have to have proportional representation at champs... The problem comes in when some areas are proportional while others aren't. Teams from Michigan don't have to compete for spots with teams from Hawaii, Wisconsin, Iowa, Ohio, etc... But Minnesota teams do. I'm completely willing to back a non-proportional system, so long as it provides equal opportunity for everyone, not two radically different sets of rules that potentially benefit one group but not another. Giving proportional representation to one group but not another simply doesn't present a level playing field.

[brennonbrimhall]

Originally Posted by Justin Montois

I'm not sure why a FIRST-wide points system based on the standard district model points system wouldn't work....

Top 600 teams in the world go to champs.

I think once districts completely displace regional events, then that only makes sense -- especially with the promise of inter-district play.

I'm all for having the most competitive Championship experience, but the impact of a Championship on a less-than exemplary team also needs to be noted -- I've had the distinct pleasure of seeing faces light up with inspiration at the level of play at the Championship. However, some of those teams wouldn't have gotten there if they hadn't been that third robot to the #1 alliance. So I guess what I'm saying is that the top teams should go, but that also needs to be weighed against the merits of letting weaker teams experience Champs.

[Shrub]

Originally Posted by Jon Stratis

FIRST doesn't have to have proportional representation at champs... The problem comes in when some areas are proportional while others aren't. Teams from Michigan don't have to compete for spots with teams from Hawaii, Wisconsin, Iowa, Ohio, etc... But Minnesota teams do. I'm completely willing to back a non-proportional system, so long as it provides equal opportunity for everyone, not two radically different sets of rules that potentially benefit one group but not another. Giving proportional representation to one group but not another simply doesn't present a level playing field.

According to this awesome whitepaper I found, Minnesota is the most under-represented state for FRC teams at CMP (a 14 team difference). (link)

There's also a paper on representation of regions at CMP (link) that goes back quite a bit. (It probably doesn't have much to do with the topic at hand but I think its a good read as well).

Although I do find it interesting, I really have no idea what can be done about it, if anything (especially the under-representation some regions face).

[Boe]

Originally Posted by Jon Stratis

Yes, having so many teams make it from a district that the championship is worthless is a problem... but perhaps more of a problem is for the areas that aren't districts.

Minnesota, for example, is probably going to have about 200 teams next year. As a state, we have 4 regionals... which means 24 slots at champs. Sure, a few teams travel to out of state regionals, but even more teams from out of state come to our regionals and win slots. What it ends up meaning is that, as a state, the number of teams we send to champs is essentially a constant, and not comparable to the number districts send, as a proportion to FIRST population. We don't grow as champs grows - if we want more MN teams to make it to champs, we have to hold more events. The same isn't true for the districts, which creates, in my opinion, a severe imbalance across FIRST.

FIRST needs to come up with a system whereby they can support both the district and regional model while allowing proportional representation from every distinct area. Picture something like drawing up district lines across all of FIRST for Champs participation, but the method of entry for each individual region could be different - The district model could use the point system with a district championship, while the regional model could pull X teams from each regional, based on the number of teams needed for that area (with each area designed to have a minimum of 6 teams attending). For the regional, you could use a point system for the event, or base it on awards, or whatever.

Just to add some numbers to this:
MN Teams in 2014: 186
FRC Teams in 2014: 2707

Percentage of FRC teams from MN: 6.87%
Number of slots if represented at champs by percentage: 27.5
Current in state slots available: 24
Percentage of MN teams at champs if all slots won by MN teams: 6%

2014:
24 slots available, 6 slots won by non-MN teams
2 MN teams double qualified (one of them providing a wildcard slot to another MN team)
Slots won by MN teams: 18
Slots used by MN teams: 17
Percentage of champs teams who are from MN: 4.25%

The percentages currently aren't that far off, but as MN grows it will become farther and farther away from equal representation if more slots aren't available to MN teams. Of course there are two other regionals frequented by MN teams that are available but for the most part MN brings in more out of state teams then we send to other states.

I may go through at some point and run the numbers for 2013 as well if I get the chance.