Chief Delphi

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)
-   Rules/Strategy (http://www.chiefdelphi.com/forums/forumdisplay.php?f=6)
-   -   Ultimate Ascent Strategies (http://www.chiefdelphi.com/forums/showthread.php?t=110453)

BrendanB 09-01-2013 13:50

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by jwfoss (Post 1211463)
I would like to pose a question: Ultimate Strategy for What?

- Ultimate Strategy to Qualify #1
- Ultimate Strategy to Get Picked
- Ultimate Strategy to Win a Regional
- Ultimate Strategy to Win Championships
- Ultimate Strategy within your team's abilities

I would strongly argue that these goals are not all the same.

They aren't which is why teams really need to consider what their goals are this season, their resources, the events they attend, etc.

There is no ultimate strategy.

pmangels17 09-01-2013 14:49

Re: Ultimate Ascent Strategies
 
Frankly, the strategy should revolve around your teams resources. Many teams, ours included, have overstretched their means in a season, and need to be aware of the mechanical complexities of their robots and whether or not they can make such complex systems work. I like Robot in 3 Days, because they built something that can score a reliable, decent amount of points. Almost every team has the capability to do something like what they did. So the first thing to consider is, what is the best way for my team to score the most amount of points, not the best way for any team to score the maximum amount of points, depending on machining ability, shop time available (for those of us that work in schools, and don't have 24/7 access to their shops), budget constraints, driving skill, and design/CAD ability. Side note: every team has access to free CAD programs, both from Autodesk and Dassault (Solidworks). OUr team uses a combination of both, since we have students familiar with each.

PayneTrain 09-01-2013 15:41

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by jwfoss (Post 1211463)
I would like to pose a question: Ultimate Strategy for What?

- Ultimate Strategy to Qualify #1
- Ultimate Strategy to Get Picked
- Ultimate Strategy to Win a Regional
- Ultimate Strategy to Win Championships
- Ultimate Strategy within your team's abilities

I would strongly argue that these goals are not all the same.

If you want to be #1, first try to predict the alliance score. Then cut at least a third off of that number because the average alliance score is always overestimated for a number of reasons. You would want to build a system that guarantees a large autonomous score to break as many ties as possible, and because it can provide almost as many points per robot as the highest hang. You also want a consistent hanging mechanism, but to 1-seed a regional it just needs to work at level one or two consistently, not three.

If you want to ensure someone picks you, build a defensive-minded drive train and spend the rest of your time working on a 30-point hang that never fails. Intelligent scouting doesn't mean that as a captain you try to find the two best robots available for your team, but the two that best complement your strategy. Those that know that will probably pick a robot like that.

The ultimate strategy to win a regional is as team specific as it is event specific. Maybe you should aim to build a #1 robot that can pick up a great #2 and ideal #24 complement. Maybe your team should try to build a dark horse robot to be a surprisingly good late pick in selection.

The ultimate strategy to win championship lives or dies by the armory of resources a team has at its disposal, so that may be out of the picture for a lot of teams. I assume that on top of satisfying requirements for the #1 robot at a regional, you should have a way to consistently deliver the colored discs to the pyramid goal and definitely go for a solid 30 point climb.

StAxis 09-01-2013 15:56

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by jwfoss (Post 1211463)
I would like to pose a question: Ultimate Strategy for What?

- Ultimate Strategy to Qualify #1
- Ultimate Strategy to Get Picked
- Ultimate Strategy to Win a Regional
- Ultimate Strategy to Win Championships
- Ultimate Strategy within your team's abilities

I would strongly argue that these goals are not all the same.

To qualify #1: Deadly accurate shooter with 10 point hang and fast floor pickup
To get picked: 30 point climb and dumping (shouldn't be underestimated, but I imagine alliances 6-8 will rely on climbing like they did balancing last year)
To win a regional: Robot as described to rank #1, a manual loading shooter or feeder, and a climb and dumper
-To win championships: 1 robot as described to rank #1, a decent shooter with 30 point climb ability, and a climber with dumping ability
-Depending on ability: For rookie teams, create a robot that can get frisbees loaded into you and then disperse them with an identical slot on the other side, fast load and shoot combo to get picked, then 10 point hang. For mediocre teams it's going to be a decision between climbing and shooting, but pick one and go after it. Finally for the elites of the world, if you can climb 3 levels in less than 20 seconds and still have a good shooter, which I believe at least someone will do, I will just sit there is awe.

karomata 09-01-2013 16:53

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by Sean Raia (Post 1208198)
This.

Also, your teams lifting mechanism in 2010 was very interesting... it was brought up a couple of times at our meeting today.

Now that you mention it, one of our prototyping teams is trying to create a design, in which our robot goes up to the corner of the pyramid, latches on with a mechanism similar to our 2010 robot, then lifts itself up. Then, using an identical mechanism on the other side, grabs the pole and flips itself over again, and it does that 2 or 3 times to reach the top.

AllieS4246 09-01-2013 17:43

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by Citrus Dad (Post 1211268)
However, an alliance is more likely to have 3 solid climbers than 3 good shooters. 3 climbers will score 120 points before any baskets. If a bot can climb to 20 points, they can get 30 points, so the real option is 10 points for climbing. So an all shooting alliance would have to score 90 points and hold the other alliance scoreless.

Please explain the logic behind 120 points in just climbing. The Max is 30 for each team, and an alliance has 3 teams.:confused: 90 points....?

mwmac 09-01-2013 17:53

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by AllieS4246 (Post 1211695)
Please explain the logic behind 120 points in just climbing. The Max is 30 for each team, and an alliance has 3 teams.:confused: 90 points....?

Perhaps assuming that 6 colored disks are scored by the climbing alliance?

AllieS4246 09-01-2013 18:04

Re: Ultimate Ascent Strategies
 
Assuming the quote I am going %100 climbing only.

Adamc4 09-01-2013 19:17

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by mwmac (Post 1211704)
Perhaps assuming that 6 colored disks are scored by the climbing alliance?

You also can't preload colored discs, so on those grounds the dedicated climber bots would probably need a solid drive system to load those discs, and capable storage for getting the 5 pointers. You might as well just have a dump robot then that doesnt need to shoot, but just dumps 1 pointers, and then climbs. That way it can be a climbing robot, but can gain easy dumping points

Donut 09-01-2013 22:19

Re: Ultimate Ascent Strategies - Scoring possibilities
 
Quote:

Originally Posted by Citrus Dad (Post 1211244)
Given the parameters and observations, I expect the following scoring breakdown in the Einstein round:

...

- Total: 216-235 points

This would be some of the most epic Einstein matches ever if we saw scores like this. I can only hope it happens.

I think that the ability of teams to build successful 30 point climbers is being a bit overestimated. Given how some previous end games have gone (and I think most previous end games are easier than this one) there will be Regional events where maybe one team at the entire competition can successfully climb for 30, and it may not be consistent. I could see it at IRI but I think it will take the stars aligning for a triple 30 point climb alliance to form even at the World Championships. I just don't see a good 30 point climber being left for a 2nd pick anywhere.

I still really like the climbing and dumping bot idea, and could see a robot like that seeding #1 at a regional if it is built to execute consistently. 30 climb points, 20 dump points, 6 autonomous points, and 12-20 low goal points a match depending on how it loads would give it close to 70 points a match. That's pretty much unbeatable in most qualifications barring an unlucky pairing against 1114 and 67 at once. An 18 point autonomous shooter with a 10 point hang still needs to hit 4 volleys of 4 disks with near 100% accuracy to stay even with super dump bot.

sharrington9614 09-01-2013 22:33

Re: Ultimate Ascent Strategies
 
So far 1757 is going with a shooter as well as a grappling hook gun and a winch to hoist the robot to the second tier, thoughts?

CalTran 09-01-2013 22:43

Re: Ultimate Ascent Strategies - Scoring possibilities
 
Quote:

Originally Posted by Donut (Post 1211975)
This would be some of the most epic Einstein matches ever if we saw scores like this. I can only hope it happens.

I think that the ability of teams to build successful 30 point climbers is being a bit overestimated. Given how some previous end games have gone (and I think most previous end games are easier than this one) there will be Regional events where maybe one team at the entire competition can successfully climb for 30, and it may not be consistent. I could see it at IRI but I think it will take the stars aligning for a triple 30 point climb alliance to form even at the World Championships. I just don't see a good 30 point climber being left for a 2nd pick anywhere.

I still really like the climbing and dumping bot idea, and could see a robot like that seeding #1 at a regional if it is built to execute consistently. 30 climb points, 20 dump points, 6 autonomous points, and 12-20 low goal points a match depending on how it loads would give it close to 70 points a match. That's pretty much unbeatable in most qualifications barring an unlucky pairing against 1114 and 67 at once. An 18 point autonomous shooter with a 10 point hang still needs to hit 4 volleys of 4 disks with near 100% accuracy to stay even with super dump bot.

12-20 low points is really pushing it...that's three to FIVE trips, not including your color run, across the field, not including ample time to climb up the pyramid and defense

Justin Montois 09-01-2013 22:50

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by sharrington9614 (Post 1211989)
So far 1757 is going with a shooter as well as a grappling hook gun and a winch to hoist the robot to the second tier, thoughts?

Remember that you need to contact the first tier and get completely off the ground before going to the second rung.

Also, it's important to remember your teams ability, resources, past successes ETC before deciding on a strategy. A robot that can consistently climb to the 2nd level to will be successful at most regionals. Focus on doing one thing really well as opposed to multiple things at a lower ability.

nathan_hui 09-01-2013 23:06

Re: Ultimate Ascent Strategies
 
G22
Expect to have to demonstrate a ROBOT’S ability to constrain itself per B above during
Inspection. PLAYING CONFIGURATION constraints may be implemented with either
hardware or software.

I'm not sure the gun would be able to meet that requirement. Also, I'm pretty sure the gun is illegal under FRC rules. Maybe R08, R09.

Donut 10-01-2013 00:31

Re: Ultimate Ascent Strategies - Scoring possibilities
 
Quote:

Originally Posted by CalTran (Post 1211998)
12-20 low points is really pushing it...that's three to FIVE trips, not including your color run, across the field, not including ample time to climb up the pyramid and defense

I was hedging more towards 3, which would give a total score of 68. If you figure a climb takes 30-45 seconds then you have 1:15-1:30 for scoring those 3 batches of disks. The advantage super dump bot has is that scoring in the low goal should be 95%+ accurate and wouldn't take much time to line up for.

If a team could somehow fit a floor pickup mechanism in there to avoid runs to the feeding station then I think the design could push 80 points.

All of this is assuming pretty low defense. Defense will definitely bring the number of scoring runs down significantly (maybe 4-8 points), but defense will bring down high goal shooters' scores as well. It may even be disproportionate; if you can't keep super dump bot essentially pinned the entire match then it has 56 points that are not defend-able (6 points autonomous, 30 point climb, 20 point dump on top of pyramid), while the accurate high shooter has only 28 points that are not defend-able (18 points autonomous, 10 point climb).

ChurgoII 10-01-2013 16:44

Re: Ultimate Ascent Strategies - Scoring possibilities
 
Quote:

Originally Posted by Donut (Post 1212092)
if you can't keep super dump bot essentially pinned the entire match then it has 56 points that are not defend-able (6 points autonomous, 30 point climb, 20 point dump on top of pyramid), while the accurate high shooter has only 28 points that are not defend-able (18 points autonomous, 10 point climb).

However, you don't take into account the points they score DURING the teleop. Not to mention, I can garuntee there will be teams who figure out how to shoot accuratly AND get the 30pts for climbing. One of those bots could easily trump a dumb bot.

The dumb bot may be easier to build, but the work for an all-around is worth it.

Donut 10-01-2013 18:22

Re: Ultimate Ascent Strategies - Scoring possibilities
 
Quote:

Originally Posted by ChurgoII (Post 1212481)
However, you don't take into account the points they score DURING the teleop. Not to mention, I can garuntee there will be teams who figure out how to shoot accuratly AND get the 30pts for climbing. One of those bots could easily trump a dumb bot.

The dumb bot may be easier to build, but the work for an all-around is worth it.

The last part of that post is comparing points that can be scored while defense is illegal, e.g. autonomous and when in contact with the pyramid. Overall point potential I think is very similar between a 30 point climb dump bot and a 10 point climb shooter.

You made me realize an important effect of the rules I might have missed though, and maybe other teams did too. Is it illegal to defend against a shooter (make contact) while it is in contact with the pyramid but is clearly not attempting to climb? I will have to check the manual more closely to see. If so shooting bots could use the pyramid like the key was used last year.

I definitely agree that a do it all bot would trump either design I am comparing. But I think a do it all bot that works well is outside the capability of 95% or more of FRC teams, mine included.

Gregor 10-01-2013 21:28

Re: Ultimate Ascent Strategies - Scoring possibilities
 
Quote:

Originally Posted by Donut (Post 1212531)
You made me realize an important effect of the rules I might have missed though, and maybe other teams did too. Is it illegal to defend against a shooter (make contact) while it is in contact with the pyramid but is clearly not attempting to climb? I will have to check the manual more closely to see. If so shooting bots could use the pyramid like the key was used last year.

G30
Regardless of who initiates the contact, a ROBOT may not contact an opponent ROBOT

contacting its PYRAMID or
touching the carpet in its LOADING ZONE.


Violation: FOUL. If purposeful or consequential, TECHNICAL FOUL. If an opponent's CLIMB is affected, each affected opponent ROBOT will be awarded points for a successful Level 3 CLIMB.

SheepGuy 11-01-2013 00:27

Re: Ultimate Ascent Strategies
 
It seems that everyone believes the max you can score from the tower is 50. But I think if you make a robot that sort of has a glove for a feeder could possibly score some more 5 point frisbee's. If I am not mistaken you can throw the frisbee's at the end of the match. So, why couldn't the human player (with much practice) toss the frisbee at the glove and let the robot dump the frisbee in the tower goal. <- This being already suspended.
(I am not sure if this is legal as I haven't throughly looked at the rules.)

Donut 11-01-2013 00:37

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by SheepGuy (Post 1212814)
It seems that everyone believes the max you can score from the tower is 50. But I think if you make a robot that sort of has a glove for a feeder could possibly score some more 5 point frisbee's. If I am not mistaken you can throw the frisbee's at the end of the match. So, why couldn't the human player (with much practice) toss the frisbee at the glove and let the robot dump the frisbee in the tower goal. <- This is being already suspended.
(I am not sure if this is legal as I haven't throughly looked at the rules.)

So basically use the "glove" as a backboard? This would be perfectly legal and would be a good strategy to try for 10 more points if no partner robots could score in the pyramid goal. Every match the human players might as well chuck their 6 colored frisbees at the pyramid goal to try to swing the match if they can, a robot backboard would increase the likelihood of making one of these shots. If we can get a climber working and go this route we might try that strategy...

Thanks for finding G30 Gregor. That definitely changes the game dynamic for shooter bots.

Grim Tuesday 11-01-2013 00:56

Re: Ultimate Ascent Strategies
 
Letting the human player lob them at it is a very interesting and feasible idea. If the 'good guy flag' as a friend of mine puts it requires a robot to do the scoring I don't think it's worth it. If the robot could be scoring 4 white discs in level 3 that's 12 pts vs 10 pts of colored discs.

PayneTrain 11-01-2013 01:48

Re: Ultimate Ascent Strategies - Scoring possibilities
 
Quote:

Originally Posted by Gregor (Post 1212675)
G30
Regardless of who initiates the contact, a ROBOT may not contact an opponent ROBOT

contacting its PYRAMID or
touching the carpet in its LOADING ZONE.


Violation: FOUL. If purposeful or consequential, TECHNICAL FOUL. If an opponent's CLIMB is affected, each affected opponent ROBOT will be awarded points for a successful Level 3 CLIMB.

Great news for people who want to get robots in a sweet spot around or in the pyramid, bad news for dumper-defenders.

Gregor 11-01-2013 01:50

Re: Ultimate Ascent Strategies
 
Did teams really not know about G30? It is a key rule to this game.

PayneTrain 11-01-2013 02:25

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by Gregor (Post 1212853)
Did teams really not know about G30? It is a key rule to this game.

I know since that's integral to my proposed strategy, but 1:48 AM me is kinda an idiot.

CalTran 11-01-2013 09:11

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by PayneTrain (Post 1212860)
I know since that's integral to my proposed strategy, but 1:48 AM me is kinda an idiot.

But that's Robowrangler time...:D

G30 should make for interesting plays, both in shooting, and in crossing the field to get to the feeding station.

Donut 11-01-2013 23:45

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by Gregor (Post 1212853)
Did teams really not know about G30? It is a key rule to this game.

I think everyone was aware of it in the sense of "don't touch any robots that are near the pyramid or loading." I think some of us missed the connection that teams being in contact with the pyramid aren't necessarily there to climb.

mbarina 05-02-2013 16:05

Re: Ultimate Ascent Strategies
 
Can robots start hanging on the pyramid before the last 30 seconds

PayneTrain 05-02-2013 16:15

Re: Ultimate Ascent Strategies
 
Quote:

Originally Posted by mbarina (Post 1228216)
Can robots start hanging on the pyramid before the last 30 seconds

(From the FRC Q&A)
Quote:

Q30 Q. At what time may the robot begin climbing the pyramid? The manual doesn't seem to specify.

A. There are no Rules dictating when a ROBOT may begin to CLIMB the PYRAMID.

DMike 17-02-2013 10:56

Re: Ultimate Ascent Strategies
 
This is a long thread and this thought might be contained within, sorry if I am repeating. Could a strong auto shooter in a defensive manner prevent a def/climber bot from ever reaching the pyramid for it's climb? The auto shooter would score 18 in auto 12 in tele and prevent a 30 point climb. essentially scoring 60 points.

Citrus Dad 29-04-2013 16:32

Re: Ultimate Ascent Strategies - Scoring possibilities
 
Quote:

Originally Posted by Citrus Dad (Post 1211244)
Here's an assessment of this year's game's scoring:

Parameters:
- There are 72 potential points in autonomous.
- With 45 discs, the maximum points from fed discs is 135 points.
- Assuming that the colored discs are used at the end of the game, and most likely dumped directly in the goal after climbing the pyramid, an alliance can score 120 points on the pyramid. A single team can score 50 points by climbing and placing 4 discs. This procedure might even be easily automated.

Observations about likely outcomes:
- A single robot can probably score about 36-45 points in autonomous in cooperation with another robot feeding it. But most robots will score only 12 at best.
- We can expect roughly similar (or worse) teleop accuracy as last year's game. The best teams were scoring no more than 60-70 points in teleop, and most were less than 40.
- Unknown points can be scored with discs "owned" by the opposing alliance, but most likely, these will not be additive to the maximum available points.
- The floor will be littered with discs late in the game. This may affect strategies and tactics for the endgame.
- The best alliances will score 120 points in their endgame.

Given the parameters and observations, I expect the following scoring breakdown in the Einstein round:

- Autonomous: 36-45 points:
- Teleop: 60-70 points
- Pyramid: 120 points
- Total: 216-235 points

Note that the pyramid strategy has a high probability of reaching a full score in every match (unlike balancing 3 robots) once the climbing and placing tasks are mastered. On the other hand, the teleop scores will be highly variable and therefore riskier.

While the point breakdowns were off, and autonomous was important than pyramid climbing, the totals were pretty accurate. Our alliance led by Team 1678 had the top score in Einstein of 233.


All times are GMT -5. The time now is 00:38.

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