Pit Scouting Questions?

[logank013]

Originally Posted by Jarren Harkema

What weight are people giving to pit scouting data if they collect it? .

Basically, we only use the quantitative data for the first day of Qualifications. It helps us decide which defenses would be the best to put out. Day 2 qualifications and Eliminations, we have the same exact defense predictor but use our stand scouting data instead of pit data. We also use the pictures on day 2 to help us decide which teams would be the best to pick for our alliance (assuming we will be a captain because it is always better to be prepared than relying on being picked). Hope this sort of answered that question for anyone who was wondering.

[Taylor]

Stand scouting allows you to get to know the robot.
Pit scouting allows you to get to know the team.

Keep that in mind, and let that guide your pit scouting questionnaire.

[logank013]

Originally Posted by Boltman

RE: Scaling question in Pits

Yes= 1,
Maybe= 0.5
No =0

May I suggest a tweak to your assessment technique..

Visually witnessed in game=1000
Visually witnessed in practice or on tape=100

In pits....
Yes = 0.5
Maybe= 0
No=0

Haha. Read my post that is at the top of the second page (the one right above this post). We only use that pit data for day 1 defense selection. Day two, our defense predictor gets moved to Stand Data Mode. Then we use our day 1 quantitative data for day 2 (quals and elims)

[EricLeifermann]

Originally Posted by MichaelBick

We ask a lot of questions that aren't directly about the robot:

How many batteries do you have? (Will we need to lend them batteries?)
How many speeds do you have? (Helps us differentiate teams for defensive roles)
What programming language do you use? (Can we help them with programming if they need it)
Do you have pneumatics? (Can we cheesecake pneumatic mechanisms onto them?

At the end of the day teams can and will say what they designed the robot to do, not what it can actually do. They are not saying these things to be misleading, its what they designed the robot to do.

But at the end of the day if you don't do it on the field in an actual match you cannot do it.

1678 showed a 2 ball auto in a video but didn't do it in a match. Therefore they can't do a 2 ball. When they do it in a match(which I have every confidence they will) then I will say that they can do a 2 ball auto.

[Libby K]

Originally Posted by MichaelBick

We ask a lot of questions that aren't directly about the robot:

How many batteries do you have? (Will we need to lend them batteries?)
How many speeds do you have? (Helps us differentiate teams for defensive roles)
What programming language do you use? (Can we help them with programming if they need it)
Do you have pneumatics? (Can we cheesecake pneumatic mechanisms onto them?

Also, "What's your current weight/Are you open to cheesecake?"

[tindleroot]

One thing that would be good to ask since a picture does not make it apparent is "what defenses (if any) can you cross in autonomous?" This would allow you to prepare for collaboration in mutual qual matches, and prepare to pick defenses that they don't cross for any matches you have against them. Again, no promise that they will tell you the truth, but it's nice to have an idea anyways.

[mentos54]

What I've found to be useful in pit scouting is asking what the team prefers, not what they can do. As many others have said, a team can't be relied on to give accurate performance starts on their robot.

On the other hand, you can find important info on what their preffered role is. "Would you rather shoot or handle defenses" and "do you wish to focus on getting a breach or capture" gives you their preferences, which they won't be wrong about. This can help a lot when looking for alliance partners.

[JohnFogarty]

This is the pit scouting questionnaire that my teams use. You can use it as well if you choose.

https://drive.google.com/file/d/0B6z...ew?usp=sharing

[bennettj800]

sometimes robots can do certain things that they never get a chance to do (like when a portcullis is broken and nobody uses it, but its "gonna be fixed" for the last day, i.e. elim matches...)

Aye, There's ↑ the rub!

[BenjiSG]

I run pit scouting on our team. Our operations generally run like this:

On Day 0, I go around to every team's pit and get pictures of their robots. These pictures then get uploaded to a Flickr account (from my phone if the Internet's good enough, otherwise at home that night after the pits close). From there, the pictures get put onto spreadsheets which are printed and used to record pit scouting info, with each being for a different team and having their number and name on the spreadsheet.

On Day 1, I and typically at least one other pit scout divide up the teams and circle around the pits to ask questions. We have separate boxes for robot nickname, distinguishing characteristics, low/high goal shots, defense (against other robots), climbing, each of the various defenses, and goals and defenses in auto. Scouts aren't given a script, but I generally ask questions in the following manner:

• Does your robot have a nickname?
• Which defenses can you do?
• Can you shoot high/low?
• What's your drivetrain like? Are you good at pushing other robots around?
• Can you climb?
• What's your auto like?
• Anything else cool or interesting about your robot?

From there, the forms stay in the pit so that they can be accessible to drive team. As drive team plays matches, they use a separate box to record interactions with their alliances drive teams, plus anything else unique they notice.

When we go to rank teams for desirability, we use the pit scouting forms to see if they potentially have any other valuable skills that they advertised, or if drive team has problems with them, or whether they have abilities which have not been demanded in their matches.

This data usually advises decisions secondarily as a means to cross-reference, but they also serve as excellent look at a team at glance if anyone picks up on something and wants to check over a team's basic skills and robot image.

Your scouting shouldn't rely on data you get from the pit since many teams exaggerate/lie, but it can be a useful addition to your scouting process.

[bennettj800]

here's what my team has generally done:
[Click this link to seeee the sheet im talking about]

It is very specific information, but when asked in the right tones, the feedback is remarkable (yeses and no's are still information, are they not?)

after the first day of pit scouting, we have a lot of down time, so we all (we = scouts) go around the pits and write down one interesting feature that is unique to the teams' robots. If two teams share a feature, the one who works better trumps the other.

We also find our favorite bot in general and make predictions on who will be successful, who will not, based on the design aspects and performance on the practice field.

of course, prediction isnt everything... which is why we have match scouting!! the link to that stuff (which i made myseyuf) is [HERE] (in the case that anyone might want it...)

[evanperryg]

Originally Posted by EricLeifermann

I can get all those answers from the picture.

Not necessarily. There was a team at Midwest who's bot was so close to the ground that their wheels weren't even visible. At first glance, it seemed like they couldn't do any defenses, and even the low bar seemed questionable with how low that thing was. However, they were somehow able to do the group Ds, as well as the low bar. We use the pit scouting data for the initial sort (i.e. "who can do the moat?") then cross-reference that with actual crossing data. It just makes the whole process a bit faster.