Number of Matches at the Championship

First, I must say that the Championship was absolutely amazing last year. The experience was great, the tournament and just the atmosphere were all really exciting.

There’s one small issue I feel needs to be addressed, and that is the amount of matches teams play. Last year it was 7 per team, and I think it should be a lot more, because more matches means more accurate seeding. Teams would often be seeded inaccurately. I’ll use 217 as an example. They were seeded 17th in Archimedes, and were clearly not the 17th best robot in the division. Now they did end up winning the Championship, so this is not to say that good robots don’t get rewarded. I want to know what ideas people have for increasing the number of matches at the Championship. I know the logistics are crazy with so many teams at the event, but there’s got to be some way to get more matches in.

An obvious solution would be to increase the number of fields. I can’t picture exactly in my head the dimensions of the Georgia Dome, but it seems like two more fields could be squeezed in. Put Da Vinci and Einstein down at one end of the field, and have the six FRC fields condensed a little bit.

Any ideas?

(I know there has been some discussion on related topics here and here but I felt that this needed its own thread.)

I know I’ve posted this somewhere else as well (I just can’t find it), but instead of adding two more divisions, I think you could increase the number of matches per team by putting two fields in each division. One field could be running reset and having the robots get all set up while the other field is running a match. There’s 9 regionals in North America in weeks 3 and 5 so there should be enough fields. With reduced set-up time you should be able to fit in more matches without having to worry about extra volunteers and crossover between the radios.

There’s still the issue of space on the dome floor but if you only add a field without the need for any additional screen, equipment, queuing space, etc. it seems like it would be possible

~Allison

Honestly, I don’t think that you can add that many more matches.

Let’s just ignore the Dome’s requirements and consider this on just the basis of running fields. Each field requires six radio channels, and runs on a six-minute match cycle. (In other words, we have six robot-matches completed every six minutes.)

The IFI radio modem has forty channels. Six of them can be accessed by teams, which leaves the possibility that someone will make a dumb mistake and run off the field on a radio and lock up a competing robot. Clearly, this prospect is unacceptable. Therefore, we have 34 channels available to us. That breaks down to five fields of 3v3, or eight if FIRST reverted to 2v2. (We’ll ignore 2v2; FIRST’s current-generation field barrier is made for 3v3, no matter how Mission Mayhem rigged it up, and should be used by past experience through the 2009 season.) So yes, FIRST could run a fifth FRC field in Atlanta. Now how do you incorporate that into the elimination bracket on Einstein in a manner that is both fair and easily understood by the average person on the street? I don’t think you can.

Allison made an interesting idea with giving each division the GTR treatment, but I believe a similar radio problem would happen here. You’d have to run both fields on the same channels, for the same reasons as above, which would mean that if any robot or OI was turned on on the inactive field, they’d be transmitting on the same channel as the competing robot. I haven’t tried it when it’s just an OI or RC, but I have empirical evidence to prove what happens when you have more than one robot connected and attempting to communicate on the same channel.

Since we can’t really add more fields to the Championship, we’ve got to get in more robot-matches through some other means.

We could decrease the match cycle time, either by a shorter match or a shortened inter-match period. Shorter matches aren’t as interesting–two minutes flies by as it is. Shorter intermatch periods mean both the teams and the field crew have to be absolutely on top of things. That six minutes involves running the match, having the refs score it, field reset doing its thing, any field issues being repaired, teams moving their robots off, the new robots coming on, getting set up, and ready (including a once-over by the IFI rep). If anything goes wrong, matches get delayed and go later than planned. Especially on Saturday, this is unacceptable as well.

This leaves the final option of lengthening the Championship. This involves either starting earlier or finishing later, both of which cost more, both in venue fees and personnel costs. (Don’t ignore the cost of volunteers–food, T-shirts, credentials, and the oh-so-spiffy volunteer pins all cost money. Less than paying them, but it still costs money.) The other problem is that you get burnout here–volunteers get more exhausted, and some have long drives ahead of them that evening. Getting a crew to stay even longer when they have enough trouble staffing at the current level will require some solution I’m not equipped to produce. (I’m marketing, not management.)

It’s a complex problem, one that would be interesting for any management science folks out there. But I really can’t see a solution to it.

Hypothetically, they could set it up like the 2001 Championship, with additional fields with limited-capacity, portable bleachers in another location. Surely there are additional exhibition halls in the GWCC complex which could accomodate that.

If you keep enough distance (and concrete/steel) between the conflicting transmitters, you might be able to get away with it.

You’d also want to make sure that every team was playing some matches on the Georgia Dome fields, as well as some in the basement.

But there’s another concern here: at a lot of regionals, the transit time between pits and field is miniscule. At Waterloo, for example, it’s less than a minute. At the Championship, it’s often much longer. Then, there’s the time spent in-queue. That means that you’ve got less time to work on the robot between matches, and means that it is rather more difficult to keep everything running smoothly, from a team’s perspective. Some teams may actually welcome the fact that they’re not quite so rushed.

Even if the difficulties could be overcome, I’m not convinced that more matches at the Championship would be a good thing.

Scouting is, and IMHO should be, as important as seeding. I don’t think more matches will make selecting alliance partners, or promoting your team to be selected, any easier.

I disagree with your assessment that seven matches are not enough for accurate seeding at the Championship.

For those teams who perform consistently at the same level, seven matches are plenty to accurately determine their ranking. For teams who don’t perform consistently due to periodic mechanical breakdowns or driver inexperience or lack of testing time, perhaps seven are not enough to allow them to “find their groove”. But at the Championships, the pinnacle of the FIRST competition season, it only makes sense that the teams that can perform consistently well by that point should be rewarded accordingly in the rankings, as opposed to holding more matches as an equalizer. There are five tough levels of elimination rounds to get through in order to win the Championship, and for a team prone to inconsistent performance, that’s a lot of time in which something could go wrong.

That’s not to say that teams who performed inconsistently in the qualifiers can’t perform well in the eliminations, and that’s where scouting becomes tantamount to seeding, as Richard said.

As far as the example of team 217 is concerned, the ThunderChickens deserved a seeding of 17th or thereabouts, as they were having trouble with their shooter and their autonomous mode throughout the qualification rounds. In the eyes of scouting teams, to choose them would be a bit risky as compared to some of the more consistent performers in Archimedes. But luckily for both our teams, after the conclusion of the qualifiers 217’s personnel were successful in convincing 296’s scouts that the bugs had been ironed out and that they were once again shooting 10 for 10 in autonomous.

That makes a lot of sense to me now. Although seven matches may not be optimal, I think you are right that seven does a good job. Robots do need to perform consistently, and my team’s was an example of that. Our robot was not nearly as advanced as a lot of other teams, but we made 10/10 in the lower goal in every single autonomous and ended up with the sixth seed. Regardless of seeding, the best and most consistent teams will prevail in eliminations.

I don’t entirely agree with this. Due to alliance partners, there are factors which teams can’t control. Regardless of how good your robot is, a 1v3 match is next to impossible to win. I know for a fact that our robot consistently malfunctioned to the point where most of our alliance partners were playing 2v3 matches at best. If you have more matches, those sorts of factors even themselves out, and the robots who are the best do end up on top.

If you have fewer matches, such as the 7 at championships, one or two matches with malfunctioning or no-show alliance partners can mean the difference between a top 8 seed and the middle of the pack, regardless of how perfectly your own robot preformed.

This does again reinforce the importance of scouting. There is no reason such a robot should not be picked for eliminations, but given the limited number of matches, they would probably end up lower ranked that they might deserve.

That said, I don’t see any resonable way to increase the number of matches at championships short of making in longer.

in response to the 9 matches at a time, i think you’re forgetting the practice fields. (they don’t use as much stuff, but its still a field, and 5 fields for competition, 4 fields for practice, right?)

I’d like to see einstein’s field be used, instead of just blank for the entire time. Also there could be another field squeezed in… which leaves an interesting idea… Wildcard matches. Teams would be randomly (er… yeah) picked to play against each other, from all divisions. These would be played on the new field and on einstein. That would be pretty fun.

It is used. FLL is run on it.

Scouting, scouting, scouting. In an ideal world, if a team scouts effectively, with a good system, then seven matches should be about enough.

Where else are they going to put more fields? 4 Divisions, another is for VEX, another is for FLL. Georgia dome has room for 6, I think.

The dome could easily fit 2 more fields. There’s a massive amount of empty space in the middle that could be minimized with a rearrangement of the fields.

Doesn’t really matter though, because the limiting factor is radio channels.

More fields=more radio channels, more space, more game pieces, more vounteers (including the specialised and skilled volunteers such as head refs, FTAs, etc.), less time to repair your robots, etc.
As it is, it takes a minimum of 20 matches to win the championship event (7 qualification, 6 divisional elims, 4 Einstein elims). Adding more spells problems for ANY robot, especially after the wear and tear of regionals beforehand.
Any veteran team will tell you, sucess recquires ALOT more than just a good robot too. Scouting is a key factor, especially in detemining your elimination partners. Marketing, in being able to get the elimination partners you want. A top notch pit crew, in keeping your robot running.
And finally, more matches would mean more time travelling between the dome and the pits, which is long enough as it is. Unlike regionals, where the robot will leave the pits for about 10-15 minuets, and sometimes as few as 6 or 7 if you cut it close, around each match, it can take almost a half hour before your robot returns to the pits from a match during championship (counting the time it leaves both before and after the match). More matches would mean almost no time to repair your bot.

Ok, the scouting is what helps you get the stats. You dont need to have the best record to be considered good.
After seeing the GTR last year and seeing 2 fields going almost simulatiously and having to have only 6 ppl scouting between 2 fields and getting good information, i would have to say that more matches would only cause for more time not needed.

The better teams will always win the competition. i.e. 1114 & 1503 at the 2 regionals we went to. 1114 did have the top spot, but 1503 was not second. You could have a bad record but if you get picked by a team that needs your abilities then it really doesnt matter.

For the record, I was not trying to downplay scouting. If anything, my argument helps justify the need for scouting. What I was really trying to say is that the best robots may not be RANKED at the top. This does not mean they’re not great robots, and it does not mean that they’re less likely to get picked. What it means is that the number of matches is too low to accurately rank all robots.

what if we were to add another day to championships that would give the possibility to have more matches.

Figure out a solution to getting the funds to rent the Dome and GWCC for another day (which can’t be cheap), staff the fields (which isn’t cheap, as we volunteers tend to be awful hungry), pay the extra hotel room nights for the hundreds of teams, and persuade the hundreds of schools and employers to let their students and mentors go for an extra day, and we’ll call it a deal. :slight_smile:

Believe me, I am in favor of having more matches… one or two bad matches can put your team out of the top 20.
Being on drive team, it seems like your always on the move. If there were more matches with the same number of fields, that would be crazy. If there was a way to both have the pits closer to the field, and a way to have more fields, then I think the idea of more matches could be plausible.

However, by the time Nats roll around, there has been plenty of time to research and scout the teams that will be there. Regional winners have been announced, and Divisions have been released.
This is when the most scouting should occur- between regionals and Nats. You have time to research robots, and see what teams are going to be good in your division. You should already be going into Nats knowing what teams you either want to pick, or teams you want to get picked by. However, this doesn’t mean that you should be closed minded. There are always teams that you don’t expect it to make it, but do. And those usually are the best ones. (Just because I’m on 340, and its fresh in my mind, look at 2004, with 1126, 67, and 340 in Newton. Nobody expected that to happen, with having at least one team out of the top 20.)

So, this is just my insight, and thoughts…

More matches would be all fine and dandy but remember if your team is not ready to perform that may say just as much as the ranking system.

If you lose one game in the NCAA tournament your out of national championshipo contention. It also very similar with the BCS in college football.

In FIRST at the Championships marketing, and scouting is all that evens the playing field. If not for these 2 then a lot of the best teams in the nation could potentially never get their chance.

If you expect it to be a championship its going to be tough, and adding more games may really not be the solution.

I think that more matches could be possible. Thursday is used for practice right, well you have already been to 2 or 3 regionals on your machine and some have a practice bot. Since the championships are earlier this year and there is not a big layoff, you could make it to where everyone gets one practice round which is 2 matches of practice which should be enough. you start the qualifying round around 1 on thursday and that will allow each team to get at least 2 or 3 more matches in, plus what they get on friday and saturday. This will allow for at least 10 matches and not even interfere with all of the money and extra day it would take for the venue or the teams to take out of thier pockets. It is a reasonable thought. :slight_smile: