Problem's with 2015…

FIRST General Forum
#45

I don’t think you can beat perfection. But I like this game too.
I was going to post about how many lower seeded alliances were winning competitions after I watched NYC and KC, where both 6th seeds won.

Then I decided to do some maths.
Week 1: Regionals 2.43 average seed won, Districts 1.5
Week 2: Regionals 2.2, Districts 3.5!
Week 3: Regionals 2.3, Districts 1.

Overall so far: Regionals 2.33 Districts: 1.96 All competitions: 2.13

I had a thought in an earlier post about the deeper the field, the greater chance that alliances from the bottom half of the seeding order stood a better chance of winning.
Shallow vs. Deep
So far I have diddly statistically. Hoping more evidence will show up in the next 4 weeks as the game gets better understood and teams gain experience and we add in the District Championships.
I don’t feel as optimistic about Einstein. I don’t think the Maroon Crew from 2007 is likely this year.

Unfortunately, this is more true than in recent games. People I have invited to competitions didn’t last long. If you’re going to invite “outsiders”
Make sure that they show up for the playoffs.

#46

While technically correct, this analysis is flawed. The amount your average moves depends on where it is otherwise.
The 14 points added to your QA from a single match was under the assumption that your average was low otherwise (40). However, the average will move less the higher up in QA you go.
Crunching the numbers, the QA shift is still significant while there is a radically different on the table, but I just wanted to clarify that 14 points is not always 14 points.That isn’t to say the phenomenon doesn’t exist.

In fact, looking at GTRE, if you remove the highest scoring Qualifier (which had the top 2 OPR on the same team), then 1246 would drop from rank 6 to around 36 - assuming they kept their average for their other matches.
However, this is also a function of the low point-scoring of a majority of teams there - “low point-scoring” to me is averages less than coopertition + a few noodles. This leads to many teams being near the same QA, thus you get rank jumping.

#47

My initial opinion of this game hasn’t changed much (if you’re curious) so I’ll take your prompt.

1.) The step bisecting the field that eliminates robot interaction should go, but keep the landfill set up as is, an initial barrier with a small path to the other side that can be removed by stacking or moving totes out of the way.
2.) Establish platforms on either side of the field and punish teams for knocking down opposing alliances stacks. (Alternatively establish one scoring platform that teams will have to jockey for space on, but that might get difficult.)
3.) Reduce the number of totes available behind the drivers station.
4.) Go back to Win/Loss/Tie structure.

These changes would allow teams to play active defense without having the stack knockdown problem of 2003 and at the higher levels could turn the game into a struggle for scoring resources as teams try to gather totes and bins before the other alliance can steal them.

It keeps the interesting engineering challenge but doesn’t diminish the feeling that each match is a direct competition.

#48

I guess the big problem that I have with the game is the way that the rankings work along with coopertition points. Teams that may take an entire match to do coopertition points and never do anything else end up ranked extremely high seeing as co-op factors into your ranking at multiple levels. Then during playoffs, these robots who end up as alliance captains have almost nothing to do during the match and struggle to complete other aspects of the game, because all they needed to do throughout quals to get a reasonably high ranking was co-op. On the other hand, versatile robots that perform a lot of functions often get short-changed on the ranking.

#49

Most of the FRC games are like this. In 2012, if a robot focused only on co-op they would usually seed higher than a robot that focused on winning. Then in finals, they wouldn’t be able to score much outside of a double.

In 2011 robots that had a working mini bot (in competitions that didn’t have that many) would rank high, but lose because the other top teams had faster mini bots and could score tubes.

I could go on, but my point is that as competitions go on, most quals will have co-op points and power house teams will be able to outscore teams that rely on them.

#50

Further to that train of thought, the analysis is also flawed, because of the elimination of the NEGATIVE effect on ranking that playing AGAINST elites has.

In prior years: Your QS was boosted by a factor of probably 1.7 or 1.8 per match allied with an elite, and shrank by about the same for every match you played opposing one.

In 2015, playing against an elite has no impact on your QA (unless you yourself are an elite, and they’re stealing step RC’s you need). Playing allied with an elite boosts your QA. In 2015, there is no such thing as a ‘hard’ schedule (like in prior years if each match you were put up against elites). Just one that doesn’t ally you with as many elites, which you can combat by simply being a little better yourself.

#51

I completely agree that there are no hard matches this year. I think as teams tune in their robots this year, it’ll be a match against themselves rather than against another alliance. Teams will have a set limit of how many points they can earn.

This, I think, is a problem with this year though. Using finals at Michigan States as an example, I believe that it will be extremely predictable. Anyone paying attention to what each single robot can put out will be able to add together that alliance scores and predict who will win. The only variance upon this is teams that grab from the center 4 RCs.

To put that in contrast, 2011 Michigan States finals saw some of the most competitive strategies I have seen. The number 1 alliance was knocked out in the quarter finals by the 8th alliance from strategy alone. No team broke down, no minibots exploded, just pure strategy (sorry to bring this back up 217, 469, and 201).

The lack of defense this year is a double-edged sword. On one hand you aren’t purposely trying to break your opponents so you can get points for it (looking at you, week one 2014), but on the other higher levels of play will lack any diverse strategy.

#52

I was rather surprised by our playoff scores being over 100 at Alamo…we had an alliance that was better than the sum of it’s parts. Using the robots together to best advantage can give you a pretty good score this year. #5 alliance was in second place in the quarterfinals, and third in semis.

#53

I completely disagree. We just removed some factors from it. A match where you are playing with two toasters is a hard one not to have drop your ranking.

#54

Different perspectives will lead to different conclusions on the match schedules and ranking systems. For teams with legitimate aspirations of the #1 spot, this ranking system is an improvement. You can’t be completely “sunk” by alliance partners or opponents the way you could in some other games. However, this is not the vantage point of the vast majority of FRC.

For the middle of the pack teams, these rankings are just as random as always. Your average can be considerably buoyed by great partners, opponents willing/able to co-op, and some good luck. 708 vaulted from the middle of the pack to 9th at Chestnut Hill after scoring 165 points with 225 and 1218 in the last qualification match. Your average can plummet when paired with teams incapable of scoring (or worse, teams that knock over stacks) or working opposite of teams that fail to get their yellow totes on the step. There are still “easy” and “hard” matches this year, they just look different than in years past.

#55

It was interesting (but heartbreaking) watching our rankings throughout the qualifications. We started out alright, but a string of mechanical failures and software failures prevented our robot from doing everything we wanted it to. Then, one of the few matches where everything worked, our two partner robots both died!

#56

The change to eliminations really made elimination strategy one of my favorite things about this year’s game. At Alamo, it was a little boggling before quarterfinals, but once we had a strategy down it worked really efficiently. Having a seperate strategy for each match was interesting as well.

“If they take the two center cans in auto, which totes should we clear first so that our alliance member can get the remaining cans on the step?”

Before actually working with our alliance members, I thought we’d be lucky to make it out of quarters, but by playing on each others strengths really well, we gave one hell of a fight. Before our last semifinal, the difference between us and the #2 seed was .33 of a point.

I think this year’s game is great.

#57

Your alliance was awesome, working in near perfect sync. From our perspective it was a little scary!

#58

A while back I made a post about Stacking vs Capping vs ???. I talked about how the resource pile splits were interesting, two feeder stations, two scoring platforms, two landfill pits. I’m starting to wonder if the GDC underestimated teams abilities and thought teams were going to take one resource and tackle it instead of these teams coming off doing everything. I really wish at some point post season the GDC would step forward and say. “Hey this is what we envisioned would happen with the game, this is what we did to try and accomplish that.” Just start an open conversation with the people who just played their game. A lot of game developers when analyzing game health do extensive research and review including interviews. Does the GDC just do this and I’m not aware?

#59

Put the question to Frank Answers Fridays when that starts back up. That one could go a while…

#60

This game has been great skills wise. I’ve seen some teams really excel, and some teams really have to explore new types of boundaries, thinking outside the box. Technically, this is a really cool game for engineering purposes. However, like I’ve said before, this game is really like playing competitive solitaire. For the past few years, between things like GameSense and the competitive and quick gameplay, I thought FIRST was going for an exciting, crowd enticing type of game, relaying it to more of a sport-atmosphere. However, they are completely cutting that aspect of gameplay out this year.

Honestly, saying “I helped build a robot that competes with and against other robots in competition, and we lunch 2-foot-big yoga balls through goals 10 feet of the ground!” is a lot easier to say and makes being a geek much more okay than saying “I helped build a robot that stacks boxes, and then puts a trash can up on top.”

#61

The game design appears to be an attempt to fix the problems with stack attack, which turned into a game of attacking (http://youtu.be/uiZQJLMgXVU), instead of building, stacks. Removing all defense went too far. One of the best aspects of Rebound Rumble and Ultimate Ascent was watching teams maneuver around defense to get to the protected zones.

#62

I truly believe that this game is in alignment with a cycle…

Not just ‘odd year vs. even year’ or ‘round game piece vs. other game piece’ - but more in terms of the life-cycle of the RoboRio and the accompanying components.

We saw two games in that allowed for wide open fields in 2013 and 2014. Why? Because the CRio was in it’s last stages of its life-cycle. Why not allow teams to ‘have at it’ and go after each other without question of the CRio? I know that FRC 4607 had it’s share of CRio reboots on field when hit hard…

However, if the GDC allowed this style of gameplay (open field hits) and team’s RoboRio resetting in the middle of matches - we would all be calling foul. I think that the GDC is safeguarding their investments and will eventually open up the field as the RoboRio matures (and more teams become acclimated to its nuances). This is a perfect game to keep the RoboRio ‘safe’ while the GDC works out other bugs…

So in my estimation - do not hold your breath waiting for another AA or UA style field for at least 2 more years. When that happens, hopefully you have stockpiles of the RoboRio… and when it does - Have at it!

#63

I’m not seeing how this is true. 2014 SVR results Am I missing something?
1st seed was 11-0, 8th seed was 8-3 (6 QS difference)
9th seed was 8-3, 42nd seed was 5-6 (6 QS difference)

Each of these was a 3 match difference.

#64

While I found your observations incredibly astute and thought provoking, I don’t necessarily think the assertions made hold.