Scouting 2016

[The Doctor]

Originally Posted by cdogcdl

And some friendly advice for pit scouting: don't believe what people say, just what you see. If they really want you to believe they can cross 'X' defense, have them show you.

Yeah, this year I've been designing a scouting webapp that is focused 100% on on-field scouting. I'm pretty sure a light amount of pit scouting would be really all that's useful, since teams will surely try to promote themselves. They might also have different luck under ideal conditions, as opposed to an actual match.

[emimadi]

Do you all think that Practice Field scouting would be beneficial? Also, make sure you're scouts are well trained and probably want to be there all day on Thursdays.

[Kevin Leonard]

Originally Posted by emimadi

Do you all think that Practice Field scouting would be beneficial? Also, make sure you're scouts are well trained and probably want to be there all day on Thursdays.

Only for training new scouts. Teams aren't ready on practice day and their efforts on the field aren't indicative whatsoever of their final performance.

[XaulZan11]

Originally Posted by Kevin Leonard

Teams aren't ready on practice day and their efforts on the field aren't indicative whatsoever of their final performance.

True, but it is probably the best way to determine how they will preform Friday morning when you have no 'real' matches to base strategy off of.

[ldsedam]

A big thing this year to scout as well is driver performance. Definitely a more qualitative form of data, but something that could prove useful. Being a game with cramped space, a driver's ability to work around the defense could prove to be super useful information when going into elimination picks. The flip side of that, a robot could have the ability to perform well, but in the hands of the wrong drivers this year, being a more defense heavy game, teams can perform poorly

[evanperryg]

Originally Posted by chandrew

Are you referring specifically to pit scouting or to match scouting as well? In general data should be more quantitative such as number of high goal hits vs. number of high goal misses, number of defenses breached, etc... Instead of simple yes/no questions.

Not necessarily. The qualitative notes we take match-to-match are an extremely valuable asset, as they allow us to recall certain robots more precisely. Especially if there's a lot of teams who look almost the same statistically, certain observations can be the difference between a good pick and a bad pick.

Personally, I think the most important metric this season (second to average scoring and consistency, of course) will be high goal consistency. It's a difficult objective, as it requires consistent positioning and consistent shooter angle/velocity, but it's worth great points, in a game where no one basic objective is worth much more than any other.

[jajabinx124]

Originally Posted by Kevin Leonard

Only for training new scouts. Teams aren't ready on practice day and their efforts on the field aren't indicative whatsoever of their final performance.

Teams aren't ready, but you can still get a "feel" of the competition. Usually I just walk around the pits, watch performances on the practice fields, etc. As the team strategist, its beneficial for me so I get a feel of what type of robots I may see at the competition during practice day.

Also if your team uses a electronic scouting system it's always super beneficial to test out your system to make sure it runs as you want it to during practice day.

[tindleroot]

Originally Posted by Kevin Leonard

Only for training new scouts. Teams aren't ready on practice day and their efforts on the field aren't indicative whatsoever of their final performance.

But you can see what defenses they (think they) can cross over.

[The Doctor]

Originally Posted by emimadi

Do you all think that Practice Field scouting would be beneficial? Also, make sure you're scouts are well trained and probably want to be there all day on Thursdays.

It could be more useful than pit scouting, but it's still very different conditions than an actual match. What I'd recommend is doing some pit/practice scouting at first, but relying on match scouting after enough is compiled.

[Robojacks4674]

im new to FRC, coming from VEX, and scouting wasn't anywhere nearly as in-depth as it is here. What do you recommend for an interface. im willing to develop applications, use web-based, iOS, Android. Anything.


EDIT: I plan on developing a design similar to http://www.chiefdelphi.com/media/img...485ea6c2_l.jpg by user 'The Doctor' using Microsoft Virtual Studio, and I'm curious at what methods I could use as to organize this between six scouts. only focusing on field scouting, perhaps with an editable notes field for pit scouting. willing to use LTE as a viable method of sharing the information.

[philso]

You would want to speak with the drive team to ensure that you can work with them.

Three of us were pit scouting to look for robots to cheesecake at the Lone Star Regional last year. A team that "made the top 70% possible" told our team member to "go away because they were busy" after answering one or two questions about their robot. Needless to say, we made sure that we did not pick them. Our second pick was a team that ranked lower but demonstrated clearly that they wanted to work with us. Our team member doing the pit scouting is a pretty mild and polite guy so I doubt that he said anything to offend them.

[Just Doug]

I think the basics about this game are pretty well covered here. So I'm going to throw in my two bits about pit scouting.

You get mixed signals on pit scouting when you talk to other people about it. Many mid to upper tier teams don't do pit scouting. I personally think this is a mistake, but then again I would also argue that most teams that do pit scouting are doing it wrong. The big problem with doing it well is that it is really hard to teach others how to do it.

The general question and answer session that teams like to do isn't terribly useful. Teams tend to tell you what the design goals for the robot is, not what the existing robot is actually doing. It not that they are intentionally misleading you, but that they are giving in to human nature and being overly optimistic. However, asking questions can be useful when a team tells you it can't do something. Teams tend to be telling the truth when they do that.

What is more useful is getting pictures of the robots, so you can put a face to the name, so to speak. Also a visual inspection of a robot, or a teams pit can tell you a ton of information when you know what you are doing. I might be hard to identify a team's drivetrain while they are on the field, but standing next to a robot you can identify a good number of details. Drivetrain details are particularly useful in figuring out what happens when you get into a pushing match with their robot.

I like to walk the pits between matches to check out the general feel of teams we are allied with and going against. With a little experience you can figure out when a team is having trouble with a robot. If you have build guys that are twiddling their thumbs it is a great opportunity to volunteer their help. If you don't have spare manpower it is useful to adjust the individual match strategy.

It was mentioned earlier, but it is also a good chance for diplomacy. Remember that you are representing your team when you are wandering the pits. Behavior that is not in alignment with the spirit of gracious professionalism not only reflects on you, but on your team as well.

Depending on how many students you have working for the scouting team, I think it is safe to skip pit scouting in exchange for getting a student watching every single robot every single match. At least run pit scouting at a lower level, because numbers from match scouting is much easier to parse into useful information.

[Citrus Dad]

We are back to our basic 2014 scouting system with added tweaks. You can see two workshop sessions on how we set up our scouting and what's important here. Our scouting system whitepaper is here.

This year we are trying to implement a more quantitative measure of defensive effectiveness based on score differentials from predicted scores. We also use a separate set of "superscouts" to evaluate qualitative robot data in a ordinal comparative fashion across an alliance.

We'll be looking for effectiveness in crossing specific defenses and goal scoring. We'll have to see how defensive strategies evolve. I think these might depend in part on rules interpretations by referees.

[Citrus Dad]

Originally Posted by Just Doug

I think the basics about this game are pretty well covered here. So I'm going to throw in my two bits about pit scouting.

You get mixed signals on pit scouting when you talk to other people about it. Many mid to upper tier teams don't do pit scouting. I personally think this is a mistake, but then again I would also argue that most teams that do pit scouting are doing it wrong. The big problem with doing it well is that it is really hard to teach others how to do it.

The general question and answer session that teams like to do isn't terribly useful. Teams tend to tell you what the design goals for the robot is, not what the existing robot is actually doing. It not that they are intentionally misleading you, but that they are giving in to human nature and being overly optimistic. However, asking questions can be useful when a team tells you it can't do something. Teams tend to be telling the truth when they do that.

What is more useful is getting pictures of the robots, so you can put a face to the name, so to speak. Also a visual inspection of a robot, or a teams pit can tell you a ton of information when you know what you are doing. I might be hard to identify a team's drivetrain while they are on the field, but standing next to a robot you can identify a good number of details. Drivetrain details are particularly useful in figuring out what happens when you get into a pushing match with their robot.

I like to walk the pits between matches to check out the general feel of teams we are allied with and going against. With a little experience you can figure out when a team is having trouble with a robot. If you have build guys that are twiddling their thumbs it is a great opportunity to volunteer their help. If you don't have spare manpower it is useful to adjust the individual match strategy.

It was mentioned earlier, but it is also a good chance for diplomacy. Remember that you are representing your team when you are wandering the pits. Behavior that is not in alignment with the spirit of gracious professionalism not only reflects on you, but on your team as well.

Depending on how many students you have working for the scouting team, I think it is safe to skip pit scouting in exchange for getting a student watching every single robot every single match. At least run pit scouting at a lower level, because numbers from match scouting is much easier to parse into useful information.

We pit scout in two ways. First we get pictures so that our drive coach can remember the bots during our draft sessions and to settle any issues about physical traits and parameters after the fact. We have a phone app that organizes the pictures.

Second, we look for very specific traits. In 2013 it was adding blockers, in 2014 it was adding inbounding and testing with a stopwatch on inbound time. In 2015 it was "cheesecake cookability". Don't know this year's yet. But we found it was a GREAT opportunity to get to know other teams. We think they appreciated that we had very specific requests and requirements. And we found opportunities to help other teams. We now dedicate 1-2 mentors to just working with other teams at competition.