Is Power Up Good at Ranking?


#1

Do you think Power Up’s ranking system was effective?

I’d argue no. I saw plenty of amazing teams unable to rank high due to contraists such as the lack of a buddy climb or just having partners who don’t move in autonomous. The autonomous ranking point doesn’t necessarily distinguish good/bad teams and pressure to get it pulls teams from running scale autos. The triple climb ranking point also requires good partners, even if it’s just to get on your ramps or lift.

Even teams that were amazing at the scale, switch, and vault and had a buddy climb sometimes didn’t rank high. In addition, I saw what I considered to be many carried teams at district and regional level events. Teams would rank high and just get declined by the rest of the top 8. Some teams ranked high from amazing qualification schedules, others could triple climb every time but were ineffective at the scale.

I think Power Up’s ranking system is too based on an entire alliance. Your alliance can win by an outstanding amount and still only get 2 ranking points. What does everyone else think?


#2

While a truly great team needs to have solid auto and teleop gameplay, good teams sometimes only have promising teleop performance. Since the two extra ranking points are tele-op and endgame, ranking wasn’t based on teleop performance as much so I think it did a poor job at ranking teams.


#3

Very few got it…Coopertition was IMHO for FIRST Power Up, built RIGHT into the game, but that coopertition angle was not on the field, it was work in the pits (or at least not on the field the competition was held on, sometimes often on the practice field I saw people collaborating together with all alliance partners who would listen at least, and accept the help when needed, to assure those all important ranking points).

Those that “Got It!” Went on to do really great things, won really big all season long, and some of them set a major record and never lost. Others who put in the (extra work), also, rarely lost.

If you still don’t “get it” after witnessing it, and hearing those admit exactly how they did collaborate with others (hint…search 254 team leader and members posts here on CD for some of the Power Up keys to their success, not that they were the only ones to “Get It Good” as others put in the additional work also, but they actually perfected it), to assure their own success and the success of all their alliance partners from start to finish, all year long. Then you will never “Get It!”

Think a srcond here really hard…Do the math. If 254 won every match this year (and they did), then twice as many plus (as CMP is 4 on an alliance at times), alliance partners also won those same matches. And achieved those same ranking high points.

A 53-0-0 Season, so how many alliance partners cooperated with them to achieve that? Ask them how many they no doubt had to convince diplomatically, emphatically, and very carefully, to accept thier help in the spirit of FIRST coopertition?

Admittedly though, their team experience, depth, record as a team, deep proven knowledge and diplomacy, all certainly had to help. But, that is FIRST, and a FIRST Hall Of Fame team that works/worked hard to achieve the dream.

I harped early, and loud here after reading all the game rules. I was working at AZ North, I saw the work they put in, and expected that perfect season would happen from that point on, because, I witnessed that, beyond a great robot they dreamed up, designed and built…They "Got It BIG, and put in the extra work to make sure others “Got It Too!”

The game was about collaboration (forced or voluntary coopertition to win if you will), as much as it was about picking up cubes that were not really cubes, and putting them on those teeter totters really accurately fast.

If you worked really hard…You could “Get It Too!” But to collaborate very early in FIRST Power Up is as much a key, as picking up almost cubes. ::rtm:: :smiley:

I can be a critic too;:yikes: :ahh:

BTW, not everything was all roses either for them, while practiced before the first event no doubt, the driving in the first few practice matches that I personally witnessed, was not extremely stellar (not bad, just not the usual 254 stellar)…But, it improved really fast, and kept improving every match from then on. And that is the usual game.

And they were ready upon arrival to use that full practice session up, and did they ever…We finally had to say no. It was time for connections only 5 minutes ago guys.


#4

I was worried I misunderstood something.


#5

The purpose of extra ranking points is to seed the teams in the theoretical best order of performance for eliminations. This means that whatever these bonus RP are awarded for need to be high-scoring objectives as well. For example, the first year of bonus RP (2012) didn’t work well at all because the co-op bridge did not translate at all to elimination matches.

I can see why the GDC chose these two options for bonus RP. The Auto Quest is simple enough that it guarantees that most every robot will move in auto, and Face the Boss is difficult enough for the top teams to have to design specifically for it.

The main issue with these two options is that they’re heavily dependent upon your partners. Even if you make sure every robot can move forward, errors happen and you could get stuck with a dead bot in auto. Compare that to the 40kPa auto from last year where top teams can push themselves to solo this RP. I don’t think Face the Boss had this issue, mostly because of the Levitate power up.

Another key issue with these is that they don’t really scale into elims that well. It’s easy to say that afterwards, but that’s the case.

The main determining factor for finals was who won the scale in auto. That’s because of the snowballing nature of the scale/switch, but also because teams can’t really score enough points from other means to make up for losing the scale. If Face the Boss awarded bonus points in elims, then maybe there would be a chance.

How would I adjust the bonus RP in Power Up to make it reflect elim performance more?

  1. I would remake the Auto Quest to award RP for whichever team gets more points in auto. That would push more teams to make scale autos from the start, leading to more advanced scale autos. Rookies would also be pushed to make a “move forward” auto so they don’t miss out on a potential 5 points. I would not make this give bonus points in elims though, because that would only worsen the snowballing.

  2. Keep Face the Boss as is.

  3. Add another bonus RP task called “Power Up Pros”, where you would need to use all the Power Ups to get the RP. This would be the easier option, but would make the titular Power Ups be used in every match. I would make this award points in elims so now there’s a decision to either use cubes to keep the scale or use cubes for this bonus.


#6

I think the fact that even 254 went undefeated but was still unable to achieve a perfect ranking score at 2/3rds of their events says something. No matter how hard you try to control your destiny this year part of it still comes down to luck.


#7

I’m not even sure I fully understand the question here… Every single game is perfect at ranking how teams managed to rank given the ranking system of that game.

Is Power Up good at ranking the individual on-field performance of robots? That doesn’t matter, because that’s not the game being played.

As mentioned above, if you take the time to make your future partners better at achieving the objectives you need to achieve in order to rank high, you will likely rank higher than if you do not do that. Teams with just-okay robots that did that probably ranked higher than teams with better robots that didn’t. Is that “good” by this metric, or is that “bad” by this metric?


#8

I like to think of “good at ranking” to mean “qualification ranking correlates well with playoff performance.” One of the issues I have with FIRST game design is when the game introduces bonus objectives for RP, and these objectives either disappear in playoffs (in 2018 and 2012 the task for bonus RP didn’t translate to points in playoff) or the point conversion becomes imbalanced during playoffs (in 2017 one team could achieve 40 kPa by themselves and rank high, but the 100 point 4 rotor bonus was a “must achieve” in playoffs).

Edit: I didn’t read the whole thread. Looks like Loose Screw already made my main points.


#9

I’d argue Coopertition has been nonexistent on the field since 2016. It’s all been in the pits.


#10

Good. Keep it there.


#11

The ranking system was definitely capable of being a major hurdle to making competitive alliances. Case in point: NC champs. There were 5 top tier scalebots in the state (1533, 2059, 2642, 4561, 5190) and by the end of qualifications, only 2 were in the top 5 (1533 [ranked 1] and 5190 [ranked 3]). The rest of the top 5 was occupied by shooterbots (6502 [ranked 2] and 3737 [ranked 4]) and a switch/exchange specialist (2655 [ranked 5]). Because 5190 was the GOAT, 1533 picked them. After that, the remaining top tier scalebots got gobbled up by the rest of the aforementioned teams, making it so that no alliance could even come close to matching the raw scale firepower of the first alliance and they won the event with little resistance. No shame to the rest of the teams in the top 5, they were all very good teams, just a shame the finals of NC champs couldn’t have been a massive scale shootout.


#12

To me “Good at Ranking” means that the outcome of individual matches was more a result of the differences between the alliance’s capabilities, strategy, and execution than luck (that is, factors beyond or largely beyond the control of the alliances). By this understanding, I would consider POWER Up to be good at ranking. Once the switches were synchronized by Team Update 03, the only real random factor introduced after the match started were field equipment failures and inconsistency in refereeing, both of which were kept to a minimum.


#13

QFT.

I enjoy games with bonus ranking objectives because they do two things:

  1. They help get the best teams to the top of the rankings.
  2. They give an alternative task during the match to shoot for when you’re outgunned

I think Power Up’s ranking system did the first one fairly well, but not as well as, say, 2016’s did.

I didn’t like that there was no side objective. Perhaps instead of the auto quest we could have had something concerning the # of cubes placed in some location?


#14

In my opinion, it was fairly good at ranking teams. I think those who think that Power Up didn’t rank teams fairly are going to cite the teams with great scale scoring that didn’t rank well due to buddy climbing. However, the decision to focus on scale over the climb ranking point was a conscious one, and thus they must have been prepared to sacrifice ranking in exchange for more firepower. The biggest issue I personally had was the auto-run ranking point. In early districts, even teams with consistent switch autos could not always get the ranking point due to partners not having line cross autos, and in later competitions the auto-run was virtually a free ranking point, which I’m not a fan of. In my opinion, the auto-based ranking point should have been to own both the scale and the switch coming out of auto.


#15

I’d argue that it’d be cool to see more Coopertition on the field, but not necessarily built into the game like auto runs and buddy climbs. I’d like to see more games like 2013 (and 2017 in a sense) where it’s impossible for one robot to do everything themselves.

2013 gave the teams almost too many game pieces, which led to some cool designs that work well together. You could build a full-court shooter, but you’d have to sacrifice accuracy to shoot all of the frisbees you can. Pair that with a floor pickup robot and it doesn’t matter too much if you miss; your partner can pick up what you miss. No single robot could score every frisbee on the field. 2017 also had this idea with top teams focusing on fuel while their partners scored gears.

I would love to see another FIRST game where the GDC give teams more game pieces than you’d be expected to score.


#16

You’re describing cooperation. Coopeetition on the field involves both alliances, like balancing the center bridge in 2012 or completing a center stack in 2015.


#17

Particularly Switch specialist robots didn’t do so well in the rankings. If they happened to team up with robots who were not good at the Scale or didn’t do the Scale at all they usually would end up losing. Despite the skill of the Switch bots. So it can be quite skewed in the favor of Scale bots despite Switch bots being just as important on the field.


#18

FIRST is about much more than creating a great or fair game–it’s about getting students excited STEM. Who would ever believe that students would be working tens of hours a week on an engineering project and then spend 2 days cheering in the stands watching those projects?

And a core part of FIRST’s program is getting students and mentors to work together outside of just their own team. That’s where gracious professionalism and coopertition come in. Three of the six core values in Frank’s blogcenter on this approach.

This year’s bonus ranking points were about emphasizing building those relationships. Teams were encouraged to think of how to work outside of their small group. Success was about more than just building the coolest scaler (although this year’s best team also had the coolest scaler).

I have only one criticism of this year’s RPs–unlike the last 2 years, they didn’t convert into extra points in the elimination rounds. That would have quelled some of the criticism here that the “best” robots weren’t ranking high enough.


#19

I think Power Up is good at ensuring that the best of the best robots (product of ability times consistency) at an event usually wind up at or near the top.

I don’t think it sorts the teams below that particularly well. Luckily, only the ranking of the top 8 to ~12 teams matters, so I think that yes, Power Up did a pretty good job at ranking.


#20

I agree with all of your post except this last bit. The bonus points in 2016 and 2017 completely skewed the scoring and made playoff gameplay very different than quals, not to mention confusing for spectators. I was very happy that FIRST finally decided this year to just let the $@#$@#$@#$@# game be played as is.