Are we overestimating the difficulty of this game? I refuse to include a poll because it seems like every thread has a poll these days.
There has been much discussion about the low bar and how it will affect the way Stronghold will be played. This discussion led to many polls which yield interesting results. Some of these results indicate that about half the teams (on CD) are trying to do it all. Is doing it all really so hard?
First we need to define “doing it all”. Here is the criteria as I see it (I’d love to hear some discussion on this as well):
Shooting High
Shooting Low
Traversing all the defenses
Climbing
The defenses break down into 2 categories, those that can be done with a drivetrain only, and those that require some kind of manipulation. It would be advantageous for every team to build a drivetrain capable of traversing every defense that depends on a drivetrain only. Every team probably came to the conclusion that the more capable their drivetrain is, the better off they will be. There is a COTS solution for these defenses (Rhino tracks) and there are many examples of what works and what doesn’t (Ri3D).
Shooting high and low can be accomplished with one or two mechanisms and there are many examples out there for teams to learn from. It isn’t a giant leap from building a capable drivetrain, to building a capable drivetrain with a scoring mechanism on it.
Climbing is hard. I’m not going to argue that one bit. Climbing while also doing low bar… that’s very hard. So building a do-it-all robot is very hard by extension. However, that wasn’t the initial question.
Where does a do-it-all team end up when they fail to climb, or they fail to do low bar, or both? They have a robot that can traverse most defenses and can score high and low. If they do those tasks at a high level, they can go very far. I don’t believe it’s terribly difficult to do those tasks at a competitive level given the resources that are available.
So how hard is FIRST Stronghold really?
Edit: I should probably clarify that I’m wondering how hard it is relative to previous games. I believe that FIRST in general is “The hardest fun you’ll ever have.”
(nearly) Every team plans and tests in a vacuum. There’s no defense, stress of a match schedule, etc etc. Once they arrive to a competition and the matches kick off, the dynamic changes. Week two competitions shift even more, and the game is more refined/more predictable by then.
So week one will be VERY difficult for many teams as alliances work to establish winning strategies inside a VERY complex game.
Week two will be the massive high scoring matches, and the shut-out alliances.
The issue that many teams (including mine) experience isn’t so much attempting to do everything and ending up only doing most of it, but attempting to do everything and spreading resources too thin to do anything effectively or on-schedule.
The problem isn’t so much whether a robot can or can’t do a certain objective. It’s less black and white, and the real question is “How effective can a robot which does everything be at a single objective?”. For example, a robot may be able to go under the low bar and shoot, but the shot may be low-based and easily blocked. Another robot may be able to do everything, but they need to take 20 seconds to climb and have a weak drive train.
I know exactly where you’re coming from. I just think that in FIRST Stronghold by attempting to do everything it leads you down the path to being more competitive than it usually does.
Example: In 2013 if you tried to do everything you probably put a lot of time, effort, resources, etc. into a 30 point climber. I’m sure for most teams that would kill their shooter effectiveness. Or in my team’s case, take shooting off the table.
This year, to do it all, the first thing you need is an effective drivetrain, the second thing you need is a way to score high and low. I’d be willing to bet most teams prioritized climbing lower on their lists. So the resources invested in climbing, if they are wasted, are not as significant. You still end up with a robot that can play the game competitively.
The beauty of this game is not in the difficulty of each individual obstacle. The beauty of this game is the GDC creating a game where each team will needs to make difficult decisions about how they engineer their robot.
By teams accepting the lowbar as the primary defense to attack, they will limit how capable the robot can climb, scoop, and score. It will also limit how the robot can attack the other defenses. Since the lowbar is the low-hanging fruit - most teams will attempt this one first.
This low-hanging fruit of the lowbar means that somewhere between 75-90% of teams have designed a GreenHorns style robot - and thus overlooked essential portions of the game. If 75-90% of the teams at a regional can successfully attack the lowbar (or at least try) - how many of these teams will gum up the lowbar area?
The 10-25% of teams that attacked the rest of the defensive walls disregarding the lowbar altogether - they will be much more capable to create a robot that can traverse the Sally Port, CdF, Portcullis, etc. It also sets up their team to create a shooter that is more difficult to defend. These are the robots that will be the most attractive come alliance selection. Does your team choose from a wide variety of GreenHorns, or does your team take a highly-effective breacher that can scoop and score? Or maybe there are enough GreenHorns style robots that can truly do both…
Then you are left choosing from robots that are best adapted for the end game.
I also think with the RoboRio’s tendencies towards brownouts, teams with 2-3 CIM motor boxes will have to limit their other capabilities. Tank drives, tread drives, etc that wish to push and bully for defense will find that they have difficulties. Last season, most teams could run around with light drive trains and not worry about a defender. So this may place in the minds of some of the younger teams that you can build a stalwart drive train and at the same time operate a high-functioning manipulator.
My goodness, the GDC did a great job. But back to GingerPower’s op:
This is the importance of a highly active HUB. If you have a handful of teams in your HUB willing to truly cooperate, then you can have a distinct advantage.
I don’t know about what the teams that are attempting to “do it all” are up to…we are attempting to do what we think are the most important things:
Cross enough defenses to be able to breach, as an alliance
Pick up boulders relatively quickly
Shoot high.
Or shoot low, if high isn’t working so hot (defense, lousy aiming, inconsistent shooter)
We are looking at those magic ranking points. I think we should be able to get the breaching one pretty often. The capture one will be relatively rare (at the lower-mid level regionals we play), but we intend to do our best to attempt it.
As far as defenses, I keep coming back to the fact that there are 3 robots on an alliance, and it just looks like the defenses are designed to be a cooperative thing among an alliance. Stronger robot pushing weaker robot across harder defenses, one robot opening door/bridge for the other two to cross, etc. I’m not too concerned about the defenses thing. But maybe that’s because we were able to cross 5 of them today relatively easily in testing, after getting our robot mostly together.
edit: I just noticed that I left climbing out of my post. That’s because we left climbing out of our strategy.
I disagree with the premise that if a team designs for the low bar, they are not also designing for the other defenses. I’ve worked with or seen 10 different teams’ robots this build season and not one has told me “we just designed for low bar”. Every team that is designing for low bar is also going to design for other defenses, or they’re limiting their effectiveness to an extreme degree.
The overall point that I’m trying to make is that I believe the floor of the competition will be higher this year than it has been in the past. Partly because the game pushes you to make competitive decisions, and partly because there are more resources for teams then ever before.
However, if you figure in that there are a very large amount of Rookie Teams trying to get a robot built in 6 weeks - most will go for the lowbar. Then you add in the large number of second year teams that did not have to interact with another alliance last season (that also have to adjust for defensive play as well as building bumpers) and had only 1-2 objectives to deal with in Recycle Rush - I fear there will be an exorbitant amount of teams that are as a default ‘rookie teams’ when it comes to interactive gameplay as we found in games such as UA, AA, or Rebound Rumble where there is so much going on.
And then you pile on the Rank Points to the scoring - this game is going to come down to specialization and who chose correctly, and then capitalized on that decision.
The way I see it is Stronghold’s main difficulties are in “Timing and Vision”. I am fairly confident teams will come up with bots to do some/most/all scoring plays (assuming they have enough vision) the hardest part is doing the what I call 16 scoring plays( goal and/or cross a defense twice) as an alliance in 130 seconds with a 20 second endgame.
Last year timing was an issue too… the best bots efficiently stacked 6 with a can and noodle and had ultra efficient cycle times.
As for vision… last year was half court this year full court with obstructions.
This year there are no relatively quick "multiple score"cycle possibilities to a close scoring platform, for instance there is no relatively quick 42 point plays to be had in under 30 seconds. This year its individual scoring plays the entire game many “scoring plays” of which take the same about of time as a 42 point play did last year.
That is where the difficulty is IMO this year.
Bottom line in Stronghold there is no one obvious killer strategy like last year due to lack of multiple high-score play actions. Teams will struggle match to mach based on their alliance partners and only the best will be good enough to be effective every match.
This game is definitely harder than Aerial Assist, because there was a wide open field in AA and every point (except the auto mobility bonus) was from doing something with the ball. The game was about one thing.
It is also harder than Recycle Rush, because you have to worry about defense again.
It is slightly harder than ultimate ascent, because UA had no obstructions on the field apart from the pyramid, and the goals are smaller, though the climb is easier because you can do it in one cycle.
It is about as hard as Rebound Rumble. Obstructions (somewhat more mandatory than RR), small goals (somewhat easier than RR), ball pickup all but mandatory in both, some protection against defense when shooting in both, fairly difficult endgame in both.
That’s as far back as I go with first hand experience, so I won’t go into further details, but this does seem to be harder than half to three-quarters of the previous FRC games. It seems a bit harder because the previous three appear to have been easier than the median.
My experience goes back a little further. I think Stronghold is the hardest FRC game we’ve seen since 2004. Its autonomous challenges are significantly harder than any I can recall. As Dr. Joe says, there will be many matches that begin with one or more tortugas – so teleop will often become an opportunity for drive teams to improvise under pressure. To play the hand they’ve been dealt, so to speak.
Scouting will be the hardest we’ve ever seen, because of the factors above and because of the number of possible variations of defenses.
In 2010, not many robots climbed consistently.
In 2010, not many robots could move between zones quickly.
In 2012, not many robots could move across center field quickly.
In 2012, not many robots could quickly intake a ball.
In 2012, not many robots had a shooting accuracy of greater than 50%, and that was with the opportunity for improvement on 2nd/3rd shots.
I predict that, in 2016:
Not many robots will scale consistently.
Not many robots will cross defenses quickly.
Not many robots will quickly intake boulders.
Not many robots will have a shooting accuracy of greater than 50%.
Doing everything at the highest level will be extremely difficult this year, probably not as difficult as 2013 (where no team did everything effectively), but certainly more difficult than 2014 or 2015.
A large percentage of the teams are aiming to be low bar capable robots. There is little doubt that being low bar-capable places severe restrictions on the design of the robot and how effective it will be on the competition field. There is always a group of highly effective robots, a group of moderately effective robots and a group of robots with low effectiveness. I suspect that the design limitations of being low bar-capable will make the majority of the middle group much less effective than they have been in past years. In other words, there is likely to be a greater gap between the top group and the other groups and the differences between robots in the two less effective groups will be smaller. This will make alliance selection more difficult than in other years since it is likely that a very high percentage of the pool of robots that can be selected will “suck” (as Andrew suggested in Joe Johnson’s thread about what scares Karthik) with few positive attributes to differentiate them. In essence, it may turn alliance selection into a roll of the dice. I am hoping that ours will not be one of the ones that proves Andrew right.
IMO I feel that as years go by, the learning curve for rookies is greatly skewed toward becoming shorter. This is a result of several factors, primarily the plethora of HOW-TO videos and the wonderful concept of gracious professionalism. Back when I started we were lucky to see half of the teams just fielding a robot that could traverse a flat field on the first day of competition. Everyone crossed their fingers hoping that all robots on the field moved at the start of the match. We would spend weeks determining the best method to encapsulate drill motors to make a drive train two speed. Kit bots and CD have made that mostly a thing of the past. Applying that rate of growth to super structure development and we can see that there is a noticeable difference there too. The mere fact that teams are now shown a video (Ri3d) which provides them with at least a model to emulate or follow increases the number of rookie teams with high caliber robots . Yes, not most, but at least many robots should easily be able to intake a boulder (and, subsequently low goal shoot with high frequency)
I agree with Gus’ and Richard’s assessments ^, especially about the difficulty with scouting this year. I feel that teams can now be very good at something, many things, or all things in this game --depending on their focus during robot design (and of course, their ability to carry out that design)
Caleb, I only conditionally agree with your 50% shooting if active defensive is employed (or too much time is taken for the shot).
One area I’m expecting this to be a hard game is the physical toll on the robots. Traversing the B & D defenses are tough on the bots. Add in the normal defensive contact and I expect the mechanics will be kept busy.
Another hard aspect to this game is the variability. That adds so many challenges to the autonomous mode.
There seems to be a number of individual areas a team can master. Doing them all well is really tough IMO.
I’m interested to see how robots from teams that go to 4-5 events hold up. I believe 469 has had seasons with over 100 matches the past few years. That’s quite a bit of matches and with the stresses put on robots via the defenses I’m interested in how well robots stay together.
Drive teams will have more to do in queue this season compared to earlier ones. Time to make repairs between matches will be very limited, especially for district teams.
