Chief Delphi

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)
-   Rules/Strategy (http://www.chiefdelphi.com/forums/forumdisplay.php?f=6)
-   -   Will Your Robot Be High Goaling, Scoring Gears, AND Climbing? (http://www.chiefdelphi.com/forums/showthread.php?t=153477)

MrForbes 12-01-2017 10:39

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
The last time we seeded first at a regional was 2013, and we did it with a robot that was kind of flaky, but did the main scoring thing reasonably well, and the end game only for the minimum number of points, and not very reliably.

I know that you can seed well at many regionals with a robot that leaves out a scoring method, if you are really good at the other two. You'll definitely do better than the large group of robots that are mediocre at all three elements.

But each team will do what they do...as always...

TDav540 12-01-2017 10:46

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by Nemo (Post 1629890)
If you're a team that's good enough to consistently seed as an alliance captain, then sure, it's more convenient for you to draw qualifier round allies and 2nd round picks that are capable of doing one function really well. But those allies aren't necessarily doing themselves any favors (competitively speaking) by capping their abilities at a scoring level that's too low to win a regional except as a lottery winner 2nd round pick. I'm not trying to disparage the value of those 2nd round teams, but I do say that banking on being that team is a low probability proposition. I'd only want to go that route if my team's resources were low enough to rule out any reasonable possibility of building the everything bot and seeding high.

The scoring math is depressing to look at for robots that are missing one of the key elements.

It comes down to team resources. It makes sense for lower resource teams to pick something and stick to it. This isn't a year when I'd advise a middle resource team to skip out on part of the game, and that's due to the nature of the game scoring.

I agree with this. I see a lot of posts about "Well if you put the best fuel robot and the best gear robot and a third robot to play defense together that's going to be the best alliance".

Somehow, I severely doubt that.

I still expect that most third robots at the regional level will not be able to climb, especially the 24th robot in the tournament. Even if you want to include two climbers, no one robot is going to get 12 gears. Especially if the other team is thinking any kind of gear defense. You need multiple contributors in AT LEAST the gear aspect, and in fuel too at the higher levels of play.

efoote868 12-01-2017 10:55

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by Nemo (Post 1629890)
People talk a good game about recommending that other teams leave out a core game function.

If you're a team that's good enough to consistently seed as an alliance captain, then sure, it's more convenient for you to draw qualifier round allies and 2nd round picks that are capable of doing one function really well. But those allies aren't necessarily doing themselves any favors (competitively speaking) by capping their abilities at a scoring level that's too low to win a regional except as a lottery winner 2nd round pick.

I'll hold my own team up as an example. In 2004, we completely ignored the balls, doubling balls, and goals, to focus exclusively on hanging from and playing defense on the bar. We were one of the few robots able to roll across the bar to play defense.

That made us a perfect pick for a good ball handling robot.

The year that we specialized the most, we did our best in elimination, making it to the finals on Einstein.


Now if only I can get the students on my team to realize this... ;)

JesseK 12-01-2017 10:58

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by Nemo (Post 1629890)
People talk a good game about recommending that other teams leave out a core game function.

Apologies if this is a bit terse - we've decided on a singular path this year so it isn't like I recommend something I don't do (at least this year, lol).

Of the 3 core game functions, one gives a set amount of points (and it is a reasonable amount of points for the time spent). The other two only give a LOT of points after a LOT of effort. It does teams little to no good to give half effort to both game functions to get no bonus when compared to an alliance that is great at one or the other and gets one bonus.

The hidden 'bonus' with balls comes by overwhelming the opponents with overflowing balls and then starting a cycle where balls are continuously dumped on the floor near the boiler they're to be scored in. It really only has to happen once or twice a match to be very effective for points. Yet getting & placing gears means wasting the time spent in gathering & scoring the large quantity of balls required to start this cycle.

GreyingJay 12-01-2017 11:17

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by JesseK (Post 1629934)
Of the 3 core game functions, one gives a set amount of points (and it is a reasonable amount of points for the time spent). The other two only give a LOT of points after a LOT of effort. It does teams little to no good to give half effort to both game functions to get no bonus when compared to an alliance that is great at one or the other and gets one bonus.

Getting the RP for gears is a lot of effort, yes, but even if you fail to get the RP you still get 40 points per rotor. I believe one robot working gears could get 3 rotors going with time for the climb at the end. Assuming one rotor in auto (65 points if you include the baseline cross), two more rotors at 40 points each, that's 145 points that you've contributed to the alliance, not including the climb points.

TDav540 12-01-2017 11:38

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by GreyingJay (Post 1629942)
Getting the RP for gears is a lot of effort, yes, but even if you fail to get the RP you still get 40 points per rotor. I believe one robot working gears could get 3 rotors going with time for the climb at the end. Assuming one rotor in auto (65 points if you include the baseline cross), two more rotors at 40 points each, that's 145 points that you've contributed to the alliance, not including the climb points.

Exactly. While scoring fuel is important, if you ignore and rely on your partners for climbing and gears, you will not be as successful in winning matches as might to be expected.

And winning is still the best way to get RP, last I checked.

It has to be a balance. You don't have to be the best gear robot in the world, but you need to be able to rely on yourself to do upwards of 3 per match.

You don't have to be the best fuel robot in the world, but if you can get 35-40kPA by yourself, you're sitting pretty good.

You don't have to have a climber, but man, it's sure faster to have a climber than to do another 3-4 gears.

JesseK 12-01-2017 11:52

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by GreyingJay (Post 1629942)
Getting the RP for gears is a lot of effort, yes, but even if you fail to get the RP you still get 40 points per rotor. I believe one robot working gears could get 3 rotors going with time for the climb at the end. Assuming one rotor in auto (65 points if you include the baseline cross), two more rotors at 40 points each, that's 145 points that you've contributed to the alliance, not including the climb points.

We're betting that equally-similar gears vs ball robots will have the same RP over the course of Quals. If a Ball bot loses 50% of its matches vs a gear bot but always gets its Ball RP, then it will be the same RP as a Gear bot that wins 100% of its matches vs that ball bot. So the ranking differentiator will be points scored - which I believe will lean towards the Gear bot. Thus, it seems like Gear bots will out-rank Ball bots (edit) - on average.

So to us, getting the RP for the gears isn't nearly as important as a win margin for Quals. The way we rank above other gear bots is "simply" by being the best at gears for an event, thus winning our matches with good scores.

Thayer McCollum 12-01-2017 12:08

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
I think teams pursuing versatility is important. If your robot only scores gears and you're on an alliance with two other robots that only score gears, your alliance is in trouble. Having the ability in this situation to score balls (even minimally) is better than only being able to score gears. But I think it's also important that a team is somewhat specialized.

If a team is really good at scoring gears but can also score balls, is on an alliance with a team that's really good at scoring balls but can also score gears, the alliance (imo) is much better of than if both robots were completely specialized or not specialized at all.

For example: Our team intends to have a robot that is really good at delivering gears but can also score balls in the low goal if necessary.

I think pursuit of all forms of scoring might mean that people do things only marginally well, and so it's important for a team to monitor what they are trying most to achieve and do it the best. But I also think an alliance with three robots that can all score three ways is strategically a lot better than an alliance with specialized robots.

Chris is me 12-01-2017 12:18

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by Thayer McCollum (Post 1629969)
I think teams pursuing versatility is important. If your robot only scores gears and you're on an alliance with two other robots that only score gears, your alliance is in trouble.

I don't follow how this is trouble, it means you get at least one, probably three ranking points and up to 200 points toward the final score before anyone hangs. This would be a great quals alliance.

That ranking point won't be achievable with just two robots in quals at most events. Think of how many cycles teams did in 2013, when scoring was much faster.

Lil' Lavery 12-01-2017 12:22

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Is being a marginal fit onto each and every possible alliance really better than being an ideal fit onto a subset of alliances?

mathking 12-01-2017 12:25

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
I am generally the person on my team arguing that being really good at some of the tasks is better than being decent at all of them. This year I am not so sure. At least from a building the robot perspective. This year a mediocre gear scorer that can score their initial gear in autonomous or early in teleop and is capable of getting 2-3 more in a match if needed can be very valuable if they can climb. A team that can reliably climb in 20 seconds is not at a huge disadvantage over a team that can climb in 10 seconds. So it might not be worth investing a ton of effort in a better (faster, cooler) climbing mechanism if that takes away from moving another capability from poor to mediocre. Obviously scoring fuel is the most difficult task because it takes so much fuel to get many points. On the other hand it could have a really high marginal value given that there is a hard upper limit on score from climbing and gears. I have been noodling around a lot in the past few days with game theoretic models and it is not at all clear this year (unlike most years) that a robot has a huge advantage being really at good at one thing, mediocre at one and unable to one versus a robot that is mediocre at all three. Particularly if the climbing mediocrity is that it takes longer versus not being reliable.

Optimal strategy is going to change drastically depending on your alliance. Last year there were plenty of robots that could get a ranking point on their own by clearing the outer works. This year it will be the rare robot that can score 40 fuel or 12 gears on its own. This year my plea, to my team, is to go with pretty good mechanisms we can get done quickly in order to give us time to practice with them. I completely agree with the posters who believe that we tend to give short shrift to optimizing pretty good devices versus making really amazing devices.

All of that said, try to concentrate on doing what you think will lead to the best experience for your team. Pay attention to the advice you get from experienced mentors, by all means, but don't let that just make the decisions for you. Last year I know several teams that abandoned low goal capability because the "good teams" told them to. I was personally struck by how many matches I watched last year where alliances did not get the ranking point for clearing the defenses because they couldn't clear the low goal and failed at one of the others. I ran a bunch of game theoretic simulations and you had to be really good at high goal shooting to make it worth giving up low goal capability. Furthermore, the lower profile robots seemed to hold up better to the wear and tear last year and didn't fall over as much.

Why_A_Username? 12-01-2017 12:39

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
We on 2910, determined that we will prioritize the tasks as such:

1. Gears
2. Climbing
3. Fuel

Currently we are subscribing to the thought that a Jack-of-all-Trades is a master of none. We would rather have a really good climber and gear system, with the time it takes for driving practice, than have an OK version of all three manipulators, and limited driver practice.

We are still considering handling the fuel in some capacity however, and are thinking of gathering it if it happens to be in our way, and dumping it in our zone for any shooters on our Alliances.

The thought process here is a combination of the points, and considering the time it takes to perfect and tune a shooter.

Hope this helps!

Daniel_LaFleur 12-01-2017 12:48

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by JesseK (Post 1629960)
We're betting that equally-similar gears vs ball robots will have the same RP over the course of Quals. If a Ball bot loses 50% of its matches vs a gear bot but always gets its Ball RP, then it will be the same RP as a Gear bot that wins 100% of its matches vs that ball bot. So the ranking differentiator will be points scored - which I believe will lean towards the Gear bot. Thus, it seems like Gear bots will out-rank Ball bots (edit) - on average.

So to us, getting the RP for the gears isn't nearly as important as a win margin for Quals. The way we rank above other gear bots is "simply" by being the best at gears for an event, thus winning our matches with good scores.

While the numbers you state are correct, I believe that the random pairings will influence this.

It is my belief that 70% of the seeding matched will only have 2 propellers running (at least 2 gears, but less than 6). The reason I believe this is because teams will not realize the difficulty in retrieving and placing gears once the field is littered with fuel, gears, and defense). since 1 propeller is guaranteed (due to the reserve gear) any team that can score 40 kPa (with a little help from their alliance partners) will match that 80 points of the gearbots and have the ranking point. They will thus rank higher and be in a picking position.

I believe that gearbots are more setup to win in the elimination rounds (tougher to defend against 2-3 gear ferrying bots), but need to get there first (which will be hard for them).

Nemo 12-01-2017 13:00

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by Chris is me (Post 1629895)
The advice given, to not be a jack of all trades, isn't the top teams trying to talk down to middle teams to get better second picks. It's not only a good idea because you're a better 16th selection that way; you'll be a more effective robot overall and quite possibly a captain or first pick. The advice is gets middle ground teams to win events. Depending on your team, the event, the task, and how good you're at it, you can definitely seed first without doing all of the game tasks this year.

I agree that it depends on the team, event, and task. My analysis is influenced by the fact that we're going to regional events where we expect the top handful of teams to be very strong.

Being an alliance captain and having a decent chance to win the event aren't quite the same thing. You can get stuck in the bottom half of the elimination round alliances with little realistic shot at winning if you pick a strategy with a hard points ceiling.

Chris is me 12-01-2017 13:55

Re: Will Your Robot Be High Goaling, Scoring Gears, AND Climbing?
 
Quote:

Originally Posted by Nemo (Post 1630008)
I agree that it depends on the team, event, and task. My analysis is influenced by the fact that we're going to regional events where we expect the top handful of teams to be very strong.

Being an alliance captain and having a decent chance to win the event aren't quite the same thing. You can get stuck in the bottom half of the elimination round alliances with little realistic shot at winning if you pick a strategy with a hard points ceiling.

I guess the point of the advice is, the best teams are more likely to pick the best bot at the task they aren't great at, in the first round. They're not going to pick a jack of all trades bot, unless that bot is good at everything or especially great at the one or two things they really value.

For a lot of teams they feel the choice is "do everything well" or "do some things very well", when really the choice might be "do everything at a mediocre level" or "do some thing well". The latter is, very consistently, more successful no matter how good you are at the Some Thing.

In my seasons in FRC, I've never had a season that ended with us hitting our points ceiling, having been the most effective at any task we specialized in, but unable to contribute further due to designing out a key feature. Either we never got great at that one task we specialized in anyway, or we tried to do too much and ended up as a jack-of-all trades who gets picked 4th-9th in the draft and losing before the finals.

Part of this advice is influenced by being in the district system, where being a middle of the pack semifinalist will get my team to DCMP, and then we'll be a top second round pick if we don't add additional functionality and we really do get very good at the key aspects of the game. I dunno.


All times are GMT -5. The time now is 10:49.

Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi