How does your team scout?

[Vermeulen]

This year, we had some students build and code an optical scanner system, which was built in the week before our regional out of some scrap angle iron, wood, and a digital camera. By Champs, we had (almost) fine tuned the software, and made a scantron-like scouting form, and 80% of our picks were in the Curie finals. We might release the software at some point, too, after we get rid of some bugs.

[TEE]

I asked someone on team 67 how they scouted at the Michigan State Competition, and their scouting was the best I've seen, and the easiest to implement; try hunting one of them down sometime and asking about it (I'll probably explain more later)

[Zach O]

2337's scouting system has come a long way since we first started. Our data is also a combination of pit scouting and match scouting, although match scouting is much heavier then pit scouting. This year, I'd go around to every pit, take a picture of each team's robot, and ask questions we felt were important to display on a sheet for the drive team before every match (the only question we thought we needed was do you hang, and if so, how do you hang?). I'd input that into an Excel scouting program that we change every year to fit the game. After that, I'd either enter the match schedule, or have someone do that while I was off pit scouting. I then print sheets, which we scout 4 robots per sheet, and give each person 3 sheets per scouting shift, which they have on a schedule for when they are need to scout. They scout all the matches they need to, and I go to the stands every 4 matches, take 6-8 sheets (depending on if we were scouting human players), and enter them all. I kept up changing sheets and entering data all day with help from one assistant. Before all matches, I'd also deliver a sheet to the drivers, displaying all information they might need.

What is on our scouting sheets:

• Starting Position
• Did they move in autonomous
• Where there any penalties on the team
• Points scored in autonomous, and from where (near, mid, or far)
• Points scored in teleoperated, and from where (near, mid, or far)
• Did the robot attempt to hang? Were they successful? Did they suspend another robot?

What is on the sheet we give to the drive team:

• Predicted score
• Where all usually started (based on averages, just to show a trend)
• Average points scored during a match
• Average points earned during a match for every team (this would include hanging)

Early on, we used to measure time it took to hang, but we decided it was too inaccurate, and we just used a did they attempt in fail, or attempt in succeed, or not attempt at all. After we get all of this data, the computer would organize it into a top 24 teams (which in the beginning, we mainly picked from), but as we went on, we also came up with our list of the 24 teams we wanted. In Michigan, we'd have a list of teams that we wanted on Friday, and give it out to a few people in the stands. They'd take notes, and we'd go over them shortly before alliance selection and makes some decisions (for Atlanta, we made this list on Thursday and Friday, and watched teams both days).

That's just what we do. It's worked well so far, and we both won Ann Arbor with that system. If you want information on how we did it in previous years, or more information, private message or email me.

[angelawence11]

Sounds like your guys' scouting leaves a lot to the opinion of the scouter. On 195, we try to keep everything technical. If the robot can go under the tunnel, it gets a 1. If it doesnt, if gets a 0. If it goes over the bump, it gets a 1. if it cant, it gets a 0. each robot gets point for every ball it scores, minus one for every penalty.if it can hang, 1 point. if it doesnt, 0.
Basically, with a point system that can classify each robot, we can objectively judge each robot. In order to get scouters perspectives as well, each scouter could put comments on the robot he is scouting.

[Mop Iii Top]

Our robotics team is not very big, so we only have 1-2 people scouting. At Traverse City, I was the only one scouting. At GVSU, we had 3. At TC, what I did was sit down in the stands, get some snacks, and get a pencil and paper. I numbered the match numbers as I went down the page. For every match, I was looking for the robots with good offence, defence, lifter, auto, and good kickers. After all the qualifying matches were over, I looked at my page and put in the informantion on a excel spreadsheet. I rated the robot's abilities on a scale from 1 to 10 (1 being bad, 10 being good). I rated them on kicker strength, accuracy, far zone ability, middle zone ability, near zone ability, speed, autonomous, defence, aggressiveness, and a space on the side for extra notes. When it came to alliance selection, all we had to do was look on the spreadsheet and make a list from 1-? of the picks we would want on our alliance. We ended up as 3rd seed, but the captains of the first alliance picked us. This scouting system failed in the end. I did not write anything down on the lowest robots. When it came to picking our third alliance partner, we didn't have any information to help our alliance partner pick a third robot.
At GVSU, I tried to fix this problem by writing down things on the lower robots. I had my teamate bring up a listing of the places on his computer, and tell me the 2 lowest seeded robots in this match. This didn't really work either. Mabye the fact that I was volunteering the first day, I only had the second day of qualifying to scout. Hopefully next year our scouting system will be better.

[marcopants1]

If you don't know about team 111's scouting system called WildRank, follow this link http://www.chiefdelphi.com/forums/sh...d.php?t=129763. I read this and want to implement this system into our team.

[Chris is me]

Originally Posted by Vermeulen

By Champs, we had (almost) fine tuned the software, and made a scantron-like scouting form, and 80% of our picks were in the Curie finals. We might release the software at some point, too, after we get rid of some bugs.

You only had ~8 teams on your list?

[Vermeulen]

Originally Posted by Chris is me

You only had ~8 teams on your list?

No, we had 25 or so teams on the list. When I say "finals", I usually just mean championship rounds.

[Zach O]

Scouting is much more then just looking at the numbers from each match though. For alliance selection, it's knowing what exactly your team is looking for. This year, 2337 needed a scorer, and we designed ourselves to be the best midfielder possible (although there could have been many improvements). We hung consistently after Kettering, moved a few balls every match, and by MSC we decided to add a diverter. We were able to play great defense in the mid, shutting down many bots to my knowledge, and set up or opponents to score. This made us exactly what I high scoring team would be looking for; someone who could move balls, play defense, and get points by hanging. So when looking for an alliance partner, we generally were able to look at the top people in our list, but you also need to pay attention to those NOT on your list. Someone who might have great defense, fill an area your looking for, work well with your strategy, or that team that you think will get overlooked for a 2nd round pick.