Chief Delphi - Terrifying Karthik

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)

- General Forum (http://www.chiefdelphi.com/forums/forumdisplay.php?f=16)

- - Terrifying Karthik (http://www.chiefdelphi.com/forums/showthread.php?t=143460)

Re: Terrifying Karthik

There's a disjoint between low bar capable robots, robots capable of crossing defenses reliably, and robots capable of manipulating balls reliably. As many of us are learning, any one of these is hard.

Karthik is scared teams are going to overestimate how much they don't suck.

And I don't drink Mt Dew, but I need judges for Boston District, I'll take payment in points of contact at iRobot. ;)

Re: Terrifying Karthik

Obviously he's terrified that there's 10% out there who don't understand how crucial it is that every team be able to go under the low bar this year. You're all just looking at this the wrong way.

Re: Terrifying Karthik

Quote:

Originally Posted by Joe G. (Post 1532937)

I'm pretty surprised, and just a little worried, by how many teams are responding "yes" to this poll. I say all this as part of a team that, after much deliberation, decided to go for it, but I'm worried that there are a lot of teams out there who are not fully thinking through the rather dramatic implications of designing for the low bar, and the actual strategic value of it to the average team.

I think that there are a lot of FRC game tasks, or even just general robot characteristics, which teams do not attempt or prioritize every year simply because somebody told them "trying to do everything is bad, simple is good" rather than out of a tangible, well thought out reason that it's going to be difficult for the team to pull off, or because of a valid trade-off which improves performance elsewhere. Shifters, for example. I have a hard time seeing how choosing one bulletproof, battle-proven COTs gearbox over another bulletproof, battle-proven COTs gearbox makes a robot appreciably simpler, or quicker to put together. Last year, the classic case was canburglers. The vast majority of FIRST teams dismissed this task as "too hard," only to have teams that didn't see it this way rapidly retrofit their robots to steal cans during lunch. This year, I expect to see teams fail to meet their potential in this way in regards to scaling. It's an easy task to dismiss, but also an easy thing to add after the fact (look at the WCP MCC, for example). The common thread is, it can be achieved through an "auxiliary" mechanism, something that can just be slapped on top of a robot without affecting the rest of it that much. And it's pretty close to a "binary" task; unlike something like shooting where there will be a huge spectrum of performance with gains to be made by optimization at every level, you either scale or you don't, and there isn't much to be gained by spending a huge amount of time optimizing how quickly you can do it. I would argue that some of the defenses also fall under the category of tasks more teams will avoid based on philosophy than sound engineering analysis.

The low bar is not one of these tasks. It is the opposite of these tasks.

The ability to do the low bar is immensely integral to a robot's design. It affects every single element of it, and disqualifies a number of otherwise viable designs and approaches.

The low bar takes practically every archetypical design from the previous game to which you could effectively say "build team XYZ's robot from that year," Rebound Rumble, and throws them out the window.

The low bar will make your electronics team cry.

The low bar has a direct and dramatic impact on the effectiveness of every single subsystem of your robot. Instead of releasing boulders from four feet up, you're either releasing them from one foot, or adding in systems you didn't need without the low bar to make up the difference. Same with hanging, your reach distance changed dramatically.

The low bar also has its advantages. It's one more defense that you're guaranteed to be able to breach, taking the number of other defenses to design for down from 8 to 6, and possibly eliminating some of the ones which require dedicated mechanisms to achieve. It's also the most direct path to/from the secret passage, probably the fastest defense to cross, and provides you with an optimal cycle time.

I'm worried, however, that a lot of teams are overestimating the degree to which they'll be able to take advantage of this.

By doing the low bar, you have made being an accurate high goal shooter quite a bit harder. You have also made your shots easier to defend if you stick to a low release point. Teams doing the low bar are betting on being able to make up the difference through an increased cycle rate. The number of extra shots a team can expect to miss by building for the low bar is hard to estimate, but likely not trivial, and I would argue that for many teams and the rate at which we've seen that defenses like the rock wall and rough terrain can be crossed, it may be more effective to cycle over these with a taller robot. They are also betting on consistently being effective enough to take priority over their alliance partners in use of the low bar. If as many teams want to use it as people say there will be, there's going to be a traffic jam through the thing.

By doing the low bar, many teams are completely neglecting the possibility of scaling. These teams are demanding an extra two high goal boulders a match from their low bar cross, minimum.

For teams that have chosen to neglect the high goal, the picture is even more stark. A team would need to run five extra cycles per match to make up the difference from a scale, a task which becomes dramatically easier if you allow your robot to be tall. I would bet that most teams won't even average five a match, let alone five extra cycles due purely to low bar efficiency gains.

Many teams are designing to be "breaching specialists," crossing all 9 defense styles. This gives them an extra five points per match (and no change in RP), when compared to crossing 8. Scaling, or even a single high goal shot, does the same or better.

And that's all neglecting alliance partners. The low bar is weird, in that it can be reasonably expected that both the best and worst teams in FRC will be able to do it. For the best teams, the advantage in cycle time is clear, and it's integral to their strategies. For the teams that struggle to put a kitbot on the field, taking away an effective way to score points that you're given from the start would be a poor idea. For a team in the middle, it's a reasonable assumption that their partners will be able to take care of it, and may be actively hogging it for their own cycles.

I also think Dr. Joe is right. But teams should consider, which will be the more effective robots? The ones which were designed for five weeks to do the low bar, and then hastily had a few tall bits added? Or the ones which were designed from the beginning to take full advantage of their height?

Unless you expect to be able to take full advantage of the low bar's efficiency gains, it may be in many team's best interests to walk away from the extreme design tradeoffs that the low bar forces.

I would nominate this opinion from the original thread. It was before Karthik's response, but I think sums up the various reasons that poll is interesting to people attempting to guess what the Meta game will evolve into.

Re: Terrifying Karthik

The thing to fear with 90% attempting for the low bar, is that a lack of diversity in robot design will lead to a finite amount of strategies and possible alliances that can be made in order to be effective.

Re: Terrifying Karthik

Karthik is afraid of how many under the bar robots will have shooters that will be mutombo'ed.

Re: Terrifying Karthik

Quote:

Originally Posted by Andrew Schreiber (Post 1537223)

Karthik is scared teams are going to overestimate how much they don't suck.

While I'd hesitate to say this is definitely why Karthik in particular was terrified of the results of that poll, I can definitely say this is why I was terrified of them.

On average, robotics teams really aren't particularly good at building robots* (or, perhaps more accurately, robots that can compete effectively). As someone who loves to see teams succeed both on and off the field, I looked at the results of that (and the other, similar polls about functionality) and wondered how many of the teams that said they were trying to score in the high goal, climb the mountain, and go under the low goal would even be capable of one of those things. That 90% of teams said they want to go under the low bar, and that in other polls, 64% of respondents said they would do both low bar and hang, about 80% of respondents claimed they were doing 4/5 defenses, and 70% of respondents have claimed that they are planning on doing either high goal or both high and low goal indicates that there are a whole bunch of teams that are likely biting off FAR more than they can effectively chew. Obviously I hope I'm wrong, but this indicates to me that there are going to be a lot of mediocre or ineffective robots this year (though one could argue that's the case every year, it's just more of a shame when it's such a fantastic game that supports a ton of valid niches for robots).

So, basically it boils down to that shooting, crossing defenses, and climbing are already tough, and the indication that 90% of teams are compounding the difficulty of whatever else they're trying to do with trying to do is, well, terrifying.

*I know this is an extremely strong and possibly offensive statement, but it's true pretty much across the board-- GOFIRST (my college robotics org/team) just competed in two robotics competitions on one weekend where more than half of the participants were either completely ineffective or somewhere between bad and mediocre at the tasks, including us, and we're a group of people that each have 5+ years of robotics experience. Robots are just hard, however you cut it.

Re: Terrifying Karthik

Karthik was thinking about the wiring.

Re: Terrifying Karthik

I think that the one strategic component that sparked Karthik's concerning response to this poll is the "trade-off".

A team's strategy that involves going under the low bar places a significant design and size constraint on the team's robot. The team essentially needs to design and build a robot that is ~15" tall or less in order to meet their objective of going under the low bar. This is a significant challenge, even for some of the best teams in FRC.

Most teams will end up making significant trade-offs and compromises when it comes to their robots functionality in order to accomplish their goal of going under the low bar. Where some of those teams could have had an excellent shooter, climber, or other defense manipulator(s), they might have had to reduce the effectiveness or eliminated the capability in order to go for the low bar.

A robot that can breach the outer works by itself (regardless of what defenses are on the field) would require the capability of crossing maximum of 8 different defenses. Being able to go under the low bar only reduces that requirement to 7 defenses (including the low bar).

I think that Karthik is terrified that most teams that choose to go under the low bar will have made so many design trade-offs that they won't be able to do much of anything else on the field. A team might have been better off with a robot that can shoot and/or climb rather than one that can only go under the low bar.

Teams that are designing for the low bar might actually end up lowering the bar for themselves in the process.

Re: Terrifying Karthik

He could be worried that if an entire alliance is relying on the low bar for fast cycling, a single tortuga blocking the low bar would ruin the alliance's game plan.

Re: Terrifying Karthik

Karthik is Achondroplasiaphobe. He's afraid of midgets. 90% of the robots going under the low bar will drive him insane at events

[/thread]

Re: Terrifying Karthik

90% of robots are going to have a low profile on a field that already has hard lines of sight and bad visibility for drivers.
Don't worry though the refs will see everything better due to better LOS and positioning so when you ask.
"Why are we getting a foul?"
The answer could easily be...
"Cause we are caught in the secret passage and making contact with a robot we can't see due to poor lines of sight."
or
"Cause you are making contact with a robot traversing a defense and you shouldn't be doing that."
What about the spy?
The one spy who will obviously not be invested first and foremost in just watching his or her own team, and even with good awareness communication from station to spybox involves two people who have eyes on the field suddenly deciding they want to look somewhere besides where everything is going on for second hand information from someone who may not even be looking at what you need to know.
90% of robots are planning on using the same point for transportation. Remember those movie scenes where 90% of people are stuck in a disaster and everyone goes "Now would be a very good time to take the quickest way out of town nobody else will be doing that right now". Oh also some people will be taking that one exit because in some instances they cannot physically take any other street due to defensive counter picks, it is the same as before except its one of the movies where everyone on an island is facing a disaster and they all converge on the one bridge to make a run for it.
Its a major bottleneck on scoring that could easily destroy teams score cycle times, which can easily be blockaded. If 90% of teams can only go under the low bar then in a majority of matches you could park a robot on each side of the low bar and your opponent couldn't move you out of the way because no robot is going to be designed to pull a robot away from a defense.
Actually no... My guess is maybe just maybe.
Maybe Karthik just hates limbo.

Re: Terrifying Karthik

I don't know about Karthik, but I'm afraid that teams will be making 15.5" tall robots that can't really get under at all unless they go agonizingly slowly.

We are not designing to just get under the low bar, we are designing to get under the low bar at near to our maximum speed.

Re: Terrifying Karthik

Quote:

Originally Posted by Caleb Sykes (Post 1537325)

Maximum speed is fast. Sonic is fast. Sonic is also blue. What else is blue? Water.

Water game fears confirmed.

Statistically, spiders seem most likely.

Re: Terrifying Karthik

Quote:

Originally Posted by Joe Johnson (Post 1537179)

What do YOU think Karthik is terrified of?

I think Karthik is terrified of teams that made sacrifices to attempt the low bar realizing--at their events--that they can't clear the low bar. At that point, it is very difficult if not flat impossible to recoup the capabilities you sacrificed to attempt the low bar. Just as teams that slapped on passive hook hangers in 2013 realized weight distribution was critical, or tall robots in 2012 realized they couldn't use a dingus to help balancing, the low bar is a serious set of limitations that should be regarded accordingly.

Alternate take: Karthik is terrified of having to emcee this year because even his formidable ups may not be enough for a Portcullis/Drawbridge combo, and ramparts and rough terrains may have a Libby Kamen-esque effect on Jordans.

(I drink Diet Dew, thanks.)