Chief Delphi

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)
-   General Forum (http://www.chiefdelphi.com/forums/forumdisplay.php?f=16)
-   -   The math of the cans... (http://www.chiefdelphi.com/forums/showthread.php?t=135198)

Jared 26-02-2015 17:48

Re: The math of the cans...
 
Quote:

Originally Posted by Lil' Lavery (Post 1450341)
Interesting idea that could benefit the game, but it would also have created a ref quagmire. With the speeds that teams would attempt to engage and disengage with the containers (not to mention the results of containers and gravity afterwards), it would be awfully dicey in determining if a team was only touching one container at a time.

I still don't know how the referees will be able to judge the status of 6 robots, 6 totes, and all the containers in the 200 milliseconds between autonomous and teleop.

asid61 26-02-2015 17:51

Re: The math of the cans...
 
Quote:

Originally Posted by Lil' Lavery (Post 1450327)
Nobody stopped 469 in autonomous on Einstein.
469 didn't win Einstein.

Even with an overwhelming strategy, you still have to execute. Getting the cans alone is not a chokehold strategy.

I take this quote as one example of all of the similar quotes in this thread; yours just happens to be short.

I disagree with this viewpoint because this game is totally different from 2010. There is a limited number of cans and totes, to the point where if all the cans are acquired at the highest levels it is likely that no other can grabbers can compete.
I did my math with something that solely grabs cans. Add a drivetrain and a can topper and it becomes unstoppable.
I will elaborate on this later.

Daniel_LaFleur 26-02-2015 18:08

Re: The math of the cans...
 
Quote:

Originally Posted by asid61 (Post 1450348)
I take this quote as one example of all of the similar quotes in this thread; yours just happens to be short.

I disagree with this viewpoint because this game is totally different from 2010. There is a limited number of cans and totes, to the point where if all the cans are acquired at the highest levels it is likely that no other can grabbers can compete.
I did my math with something that solely grabs cans. Add a drivetrain and a can topper and it becomes unstoppable.
I will elaborate on this later.

While this is correct, it is only applicable at the highest levels of play. Until then the value of the RCs on the step is very low ... especially if they go unused.

Lil' Lavery 26-02-2015 18:40

Re: The math of the cans...
 
Quote:

Originally Posted by asid61 (Post 1450348)
I take this quote as one example of all of the similar quotes in this thread; yours just happens to be short.

I disagree with this viewpoint because this game is totally different from 2010. There is a limited number of cans and totes, to the point where if all the cans are acquired at the highest levels it is likely that no other can grabbers can compete.
I did my math with something that solely grabs cans. Add a drivetrain and a can topper and it becomes unstoppable.
I will elaborate on this later.

You're missing the point. I wasn't disputing the importance of the cans from a game theory perspective. I'm simply pointing out you still have to execute after obtaining the cans. Unlike controlling the goals in 2002, acquiring the cans is not a chokehold on its own. The cans AND scoring points are a chokehold. If you don't execute in the scoring department, you don't win.

Obviously the teams at the highest level of play should be able to execute in almost every match. However, almost every match is not the same as every match. Mistakes still happen, sometimes crippling ones. Look no further than the Einstein finals in 2010 for proof of that.

The alliances that win the battle for the center step will undoubtedly have an incredibly high winning percentage. However, that winning percentage will not be 100%

asid61 26-02-2015 18:40

Re: The math of the cans...
 
Quote:

Originally Posted by Daniel_LaFleur (Post 1450355)
While this is correct, it is only applicable at the highest levels of play. Until then the value of the RCs on the step is very low ... especially if they go unused.

"Highest levels of play" is variable.
For example, this year we are going to Utah, SVR, and Championships (probably for the lottery, hopefully if we win). Utah may not be as challenging as SVR, although I don't know what powerhouse teams are attending. SVR ahs 971, 254, 1678, sometimes some of the texas teams. It's very competitive IME. And of course Champs will be interesting.
So depending on the competition, it changes the play. At SVR, for example, I would definitely play a tote-and-flopper, maybe with a wimpy elevator for can topping. I don't know if our team could support it at Utah, and so I would ahve to let one go.
At Champs it's a pretty good chance at victory IMO.
The purpose of a bot like this is to win, not to seed high. Sure, it's a gamble, but a traditional bot has to undergo far more fabrication and testing, as it has to be a multipurpose bot. But a tote-flopper with can topping abilities has to only be loosely tested because it has to lift only 8lbs at once. Add a capable drivetrain and the only thing that has to be tested is the grabber itself- and you have weeks to perfect that. So the odds of winning, even for a middling team otherwise (resources, mentors, etc.) are actually quite good.
Driver practice is much less because they simply have to avoid stacks and top off cans at a rate of 6 or 7 per match. One repetitive action over and over again.

asid61 26-02-2015 18:43

Re: The math of the cans...
 
Quote:

Originally Posted by Lil' Lavery (Post 1450361)
You're missing the point. I wasn't disputing the importance of the cans from a game theory perspective. I'm simply pointing out you still have to execute after obtaining the cans. Unlike controlling the goals in 2002, acquiring the cans is not a chokehold on its own. The cans AND scoring points are a chokehold. If you don't execute in the scoring department, you don't win.

Obviously the teams at the highest level of play should be able to execute in almost every match. However, almost every match is not the same as every match. Mistakes still happen, sometimes crippling ones. Look no further than the Einstein finals in 2010 for proof of that.

You don't seed high with this; you depend on being picked. A point analysis and minimal scoring capabilities is enough to allow a can grabber to win. That was the point of starting this topic, to discuss how overpowered a fast 4-can is points-wise.

Randomosity isn't something you can really control, and your bot's main mechanism failing entirely in elims would be rare, depending on your mechanism. Not much you can do about that, regardless of whether it's a can grabber or stacker.

Lil' Lavery 26-02-2015 19:01

Re: The math of the cans...
 
Nothing in my post was talking about seeding.

I'm pointing out the 100% undeniable fact that the cans are not a chokehold strategy in themselves. You still have to score points to establish chokehold. That's a fundamental difference from 2002, where controlling the goals meant you won the game, regardless of how many balls were scored. While the odds are stacked dramatically against the team that loses the can race, this game is not over after the can race.

The Einstein finals in 2010 are an example of a similar scenario. The odds were stacked against any team failing to stop 469 from establishing position, but the game wasn't automatically lost, either. 469's alliance still had to execute for the remainder of the match. Same holds true in Recycle Rush.

faust1706 26-02-2015 19:12

Re: The math of the cans...
 
Quote:

Originally Posted by Lil' Lavery (Post 1450370)
The Einstein finals in 2010 are an example of a similar scenario. The odds were stacked against any team failing to stop 469 from establishing position, but the game wasn't automatically lost, either. 469's alliance still had to execute for the remainder of the match. Same holds true in Recycle Rush.

It's pretty easy to execute a strategy when the game prevents defense from being played. It comes down to driver practice at that point. There is zero potential for the other team to "outplay" the team that gets the middle bins, only the hope that they mess up.

asid61 26-02-2015 19:26

Re: The math of the cans...
 
Quote:

Originally Posted by Lil' Lavery (Post 1450370)
Nothing in my post was talking about seeding.

I'm pointing out the 100% undeniable fact that the cans are not a chokehold strategy in themselves. You still have to score points to establish chokehold. That's a fundamental difference from 2002, where controlling the goals meant you won the game, regardless of how many balls were scored. While the odds are stacked dramatically against the team that loses the can race, this game is not over after the can race.

The Einstein finals in 2010 are an example of a similar scenario. The odds were stacked against any team failing to stop 469 from establishing position, but the game wasn't automatically lost, either. 469's alliance still had to execute for the remainder of the match. Same holds true in Recycle Rush.

So we're considering the entire alliance then, my bad. I thought you were referring to the one can grabber.
That makes sense. You do still have to score the cans across your alliance. However, you only need good stackers to win. Sure, they're bound to be less common than bad stackers, but I feel like due to the nature of the game finding good stackers won't be that hard.

Kevin Leonard 26-02-2015 19:43

Re: The math of the cans...
 
Assuming a perfect auto, if your alliance gets all 4 containers, to guarantee victory in finals, your alliance needs 7 fully capped stacks of 5 or 6 fully capped stacks of 6.
While that isn't a small number of stacks, its certainly doable by championship alliances. Although not necessarily consistently.

Interestingly enough it requires more totes to execute a chokehold with stacks of 6 than stacks of 5.

GeeTwo 26-02-2015 22:26

Re: The math of the cans...
 
Quote:

Originally Posted by Kevin Leonard (Post 1450385)
.. to guarantee victory in finals, your alliance needs 7 fully capped stacks of 5 or 6 fully capped stacks of 6.
..
Interestingly enough it requires more totes to execute a chokehold with stacks of 6 than stacks of 5.

Not really, as 5 stacks of 6 and one stack of 5, all capped, score the same as 7 stacks of 5, assuming "normal height" stacking. Both 35 totes, both 210 points. And assuming, of course, that you did get all four RCs onto your side.

faust1706 28-02-2015 19:42

Re: The math of the cans...
 
148 put up 4 stacks of 6 by themselves, would have been 5 if their last stack didn't take a tumble going to the ramp.

Add the totes from the landfill to that and you're looking at at least 7 stacks (more if teams can effectively use upsidedown totes.

The robots grabbing from the landfill better have some sort of on board vision for the end of the game....

asid61 28-02-2015 21:34

Re: The math of the cans...
 
Quote:

Originally Posted by faust1706 (Post 1451226)
148 put up 4 stacks of 6 by themselves, would have been 5 if their last stack didn't take a tumble going to the ramp.

Add the totes from the landfill to that and you're looking at at least 7 stacks (more if teams can effectively use upsidedown totes.

The robots grabbing from the landfill better have some sort of on board vision for the end of the game....

On the other hand, looking at Dallas regional right now, teams cap out at just over 200 points, of that maybe 5 canned stacks. Championships will be interesting, as this is only week 1.


All times are GMT -5. The time now is 14:10.

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