Chief Delphi

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)
-   General Forum (http://www.chiefdelphi.com/forums/forumdisplay.php?f=16)
-   -   IRI 2016 Predictions (http://www.chiefdelphi.com/forums/showthread.php?t=149079)

Gregor 03-07-2016 12:24

Re: IRI 2016 Predictions
 
Quote:

Originally Posted by Kevin Leonard (Post 1595381)
I'd run a defensive robot if I thought they could lower the output of my opponents more than my third offense bot would add to my own offense.

That's genius!

TheBoulderite 09-07-2016 10:10

Re: IRI 2016 Predictions
 
Quote:

Originally Posted by NonStopScouting (Post 1595375)
Questions for the IRI community: [list][*]Which is more valuable: Single auton high goal or scaling? They are equivalent points wise, but are they the same value for picking?

As most of the teams on the list have a single high goal auton or better, you could say that scaling points will help, but the consistency of the auton will be most important.

On the other hand, my team's alliance experienced firsthand the power of scaling points during the Carson eliminations, and you could argue that scaling points will have a large impact on who wins the match.

hutchMN 09-07-2016 14:32

Re: IRI 2016 Predictions
 
Will a no-footprint scale add significant value to a team at IRI?

fishing_cat 09-07-2016 19:30

Re: IRI 2016 Predictions
 
Quote:

Originally Posted by hutchMN (Post 1596019)
Will a no-footprint scale add significant value to a team at IRI?

What does "no-footprint scale" mean? Ty

troy_dietz 09-07-2016 19:40

Re: IRI 2016 Predictions
 
Quote:

Originally Posted by fishing_cat (Post 1596051)
What does "no-footprint scale" mean? Ty

I would assume it refers to a robot that does not "scrape" the wall of the tower as it scales, allowing robots to scale on either side without risking a robot being pushed off. (allowing a triple scale to be much more feasible) 148, 1114, 330 and several other robots scaled away from the wall this year.

Brian Maher 09-07-2016 19:45

Re: IRI 2016 Predictions
 
Quote:

Originally Posted by fishing_cat (Post 1596051)
What does "no-footprint scale" mean? Ty

Having just recently watched one of 2502's matches, I believe this means climbing in a way that the bottom of the robot is re-oriented to points sideways away from the tower.

Quote:

Originally Posted by hutchMN (Post 1596019)
Will a no-footprint scale add significant value to a team at IRI?

All things equal, I think this kind of scale is more valuable than a simple hang. However, I would take a consistent hang over an inconsistent no-footprint scale, unless I already have two climbers on my alliance and a triple scale would not otherwise be possible.

fishing_cat 09-07-2016 20:52

Re: IRI 2016 Predictions
 
Quote:

Originally Posted by BMSOTM (Post 1596055)
Having just recently watched one of 2502's matches, I believe this means climbing in a way that the bottom of the robot is re-oriented to points sideways away from the tower.



All things equal, I think this kind of scale is more valuable than a simple hang. However, I would take a consistent hang over an inconsistent no-footprint scale, unless I already have two climbers on my alliance and a triple scale would not otherwise be possible.

Ah, thank you for this info!

Brian Maher 09-07-2016 21:26

Re: IRI 2016 Predictions
 
I'm a bit late on responding to NonStopScouting's discussion questions, but here are my thoughts:
Quote:

Originally Posted by NonStopScouting (Post 1595375)
  • Which is more valuable: Single auton high goal or scaling? They are equivalent points wise, but are they the same value for picking?

My answer here is whichever is more consistent. Controlling for accuracy, I would vote high goal auto due to:
  • +1 tower damage
  • the points are locked in at the beginning of the match. A robot dying mid-match may lose the climb but cannot lose auto points
  • not scaling leaves time for an additional high goal or two, potentially making the high goal auto worth five points more than a scale without high goal auto.

Quote:

Originally Posted by NonStopScouting (Post 1595375)
  • How does batter, outerworks, turret, and midfield shooting stack up? Is a midfield turret bot more likely to be picked than an outerworks bot?

It obviously depends on the number/accuracy of high goals scored, but my general ranking looks something like this:
  1. Un-defendable outerworks shooters. While I haven't personally seen any of the IRI teams demonstrate such a shot, I've talked to a few who claim to be able to shoot over a max height defender with 15" extension from the outer works.
  2. Low catapult or tall outerworks shooters who are difficult to defend.
  3. Batter shooters, because of how difficult they are to defend once they make it to the batter
  4. Low non-catapult outer works shooters, because of how easy it can be to shut them down by sticking a defender with a wall in front of them.
  5. Non-turret courtyard shooters, because of how easy it is for a defender to push them around
I have chosen not to list turret shooters on this ranking because I think their value ranges from 2-4 depending on their time to aim and their ability to evade defenders by means such as outerworks, batter, or 179/254/1241-style wall shots.

Quote:

Originally Posted by NonStopScouting (Post 1595375)
  • At what point would you put in a defense bot? For example there isn't much point using courtyard defense against three tall outerworks bots. Would you still run another type of defense or would you only run defense against 3 midfield shooters?

Like many questions, this one can be addressed with math. I calculated the teleop high goal component OPR for each of the teams in attendance, rounded to the nearest integer, and then simuated "alliances" assuming the robots who scored the most high goals seed highest, teams pick the team available who scores the most high goals, and there are no declines:
Code:

1) 8 8 5
2) 6 6 4
3) 6 6 4
4) 6 6 4
5) 5 5 4
6) 5 5 4
7) 5 5 4
8) 5 5 4

A few observations:
  • The top alliances (1-4) will likely be able to capture with only two robots
  • The lower alliances (5-8) may have a hard time capturing, even with all three robots on offense.
  • The weakest shooter on a triple-offense strategy will average 4-5 high goals
From here, it can be concluded that defense is advisable if:
  • an alliance can capture with two offense robots
  • the defender can deny their opponents the capture OR deprive their opponents of 4-5 high goals
Based on this, playing defense will likely be advantageous for the top few alliances, and the lower alliances will need to have all three robots on offense to keep up with captures.
This simulation is obviously an over-simplification, ignoring autonomous goals and scaling points, but it gives a decent sense for the scoring ability of these alliances and how viable defense will be.

XaulZan11 09-07-2016 21:50

Re: IRI 2016 Predictions
 
There's been a couple teams showcasing their new 2 ball autonomous modes. Is there any team (that doesn't have a consistent 1 ball) planning a defensive autonomous mode? I think it would be cool (and valuable) to see a team drive over the defense, drop their ball off and then driving back to the midline to disrupt an opponents 2 ball auto.

Kevin Leonard 10-07-2016 19:57

Re: IRI 2016 Predictions
 
Quote:

Originally Posted by XaulZan11 (Post 1596079)
There's been a couple teams showcasing their new 2 ball autonomous modes. Is there any team (that doesn't have a consistent 1 ball) planning a defensive autonomous mode? I think it would be cool (and valuable) to see a team drive over the defense, drop their ball off and then driving back to the midline to disrupt an opponents 2 ball auto.

Alternatively, don't start with a ball, and first thing go and pick up the center ball.

That being said, teams often show things prior to competition that don't work while in the competition. We'll see if these autonomous modes actually work, and whether all the teams will be showing up on their A-game (they so rarely do).

I'm excited for another great IRI!

Cash4587 10-07-2016 20:03

Re: IRI 2016 Predictions
 
We will likely have a 2 ball block ready for IRI. It is easy for us to do with our robot needing no modification to implement dropping the ball, then picking up the one on the center line.

Liam Fay 10-07-2016 23:54

Re: IRI 2016 Predictions
 
My predictions are as follows:

2056 118 1024 868
133 195 2451 494
1241 3620 3130 5460
225 179 3824 71
67 1114 2052 4587
330 16 27 3641
2481 33 217 5254
3683 2771 2013 1806

Banderoonies 11-07-2016 08:03

Re: IRI 2016 Predictions
 
Anyone know where IRI will be webcast from?

Collin Fultz 11-07-2016 09:02

Re: IRI 2016 Predictions
 
Quote:

Originally Posted by Banderoonies (Post 1596218)
Anyone know where IRI will be webcast from?


http://livestream.com/indianafirst

BrennanB 11-07-2016 13:19

Re: IRI 2016 Predictions
 
0-1 - 5 ball autons will go 5/5.

0 - 5 ball autos will affect the outcome of the match.

#predictions


All times are GMT -5. The time now is 02:00.

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