Help us launch the network!
 

We're trying to erect a FIRST scouting network database and we need your help to do it. The benefits of a network are many, but it requires a lot of work, and we think that starting this thing this early in the year will give us time to work out the issues involved in the program.

One of the main areas that we need assistance in is knowing how teams do their scouting currently. As every team has many different needs, approaches, and expectations in their scouting, we're trying to make the software flexible enough to support those needs.

If you could leave us some feedback on how your scouting works, please help us out. Even if you aren't on scouting for your team, you probably know a lot more than you think. Just answer a question or two, or if you want to, answer more. Your feedback is critical to make this project a success.

Post away!

What is the main focus of your scouting, is it for strategy? For picking partners in finals? A little of both? Something else entirely?

What information is important for your scouts to collect? Drive train, arm info, etc.

How much set-up time goes into preparation for your team's scouting?

How many people are usually involved in the scouting, and how much time are they expected to put in?

How many teams do you scout at your regional?

When is your scouting usually "done" by? Is it ready on Friday for the first match, or is it only used in finals? Somewhere in between?

What methods do you use for your scouting? Randomly scribbled notes? Fill-in the blank scout sheets? Statistical analysis? Match Records? Something else? Extra feedback here would be very helpful.

How much of your scouting is done on a computer? (Includes those print-offs)


How effective/successful is your scouting?

BTW, we also need a whole bunch of help with other parts of the project, if you can think of it we probably have a place for you :) PM, AIM, or e-mail me for more information.

Thanks for all your help people!

-Chris

Re: Help us launch the network!
 

I see you decided to post. :D Yeah this involves more than just ideas but physical help. The idea is to have networks (I'm making a wireless network for VCU) in the pits where people can post info about their robot as well as talk to pit teams and ask for help. It would act like a notice board where people can put information. Were also looking at setting up this for the internet so everyone can view information about teams and their robots.

Re: Help us launch the network!
 

Historically my team has scouted to findout about other teams drive trians and arm/extra devices. The purpose of our scouting is to plan for up coming matches and to also see what innovative designs other teams are coming up with. We normally create a worksheet style form with questions and an area for answers and notes. We have attempted to keep records of matches but this information was not really used. The best info we get from match records is how other teams are playing the game and revise strategy if needed.

Every student on our team has a school supplied iBook, will be Dells this year, which makes a database system very appealing. Infact for last seasons competition we considered writting our own software and setting up a network at VCU but more pressing projects(building the robot) ruled it out.
Cyberguy34000, have you already started development? Will the database be MySQL or similar? Are you creating custom desktop frontends or will it be a website?
If I have the time I would love to help with any coding that needs to be done.

Re: Help us launch the network!
 

A little of both, usually.

What the team does or does not do in a match. The technical stuff takes a bit of a back seat if you use field scouting...

I think I put about 2 hours of work in on our sheet, then others revised it, so I don't know exactly.

As many as we can get! Actually about six or seven people spend as much time as they can spare and call for relief when they leave.

All of them, usually. If you are at one of the competitions 330 is at, you are scouted at least once.

We use some of the data to find out about our partners and opponents, but mostly for alliance selection. We have enough on Friday night to start strategizing, but we don't finish until sometime Saturday.

Fill in the blank scout sheets that last several matches (five matches, this year). Comments are made as necessary.

The match reports go to someone with a laptop and are entered into a spreadsheet. Print-offs are not done, the data on which teams we want to choose is hand-written by our team captain after it is compiled.

Very effective. We usually have a list that matches how things turn out with reasonable accuracy.

Re: Help us launch the network!
 

UPDATE: Uh, wow I didn't read your origional post, I thought you were talking about a WiFi based, multi-team, integrated scouting system... where did I get that? I think it's a good idea regardless...

I'm game for contributing to this project, and I think Perl would be ideally suited. It's the language for CGI applications (and that would be the easiest way to make this work across OS's) and it has excellent modules (perl term for libraries) for XML and practically every DB system out there.

The best way to run this would be to have people going to a regional sign up ahead of time for an account. Account registration could be done on the spot too. You could have the program automatically assign each scout a block of teams to cover, and when they have all been covered at least once, scouts can roam about as they choose.

I'm getting giddy just thinking about how incredibly cool this would be.

You could have a dedicated rotation of scouts in the stands recording scores, and they could be visible in real time! (Which reminds me, the Circuit Runners did something similar to this, ableit without wifi, and the FIRST brass came down on them hard for not marking their info as "unofficial") Each scout could record commentary on each team's performance ("Arm too weak!", "Excellent defense!", "Brilliant Automode!") subjectively. You could even have a quantitative means for recording match events. So you could record automode attempts/successes and such.

The big brother effect of this would be ridiculous. Your robots every success and failure would be recorded. Muhahahaha! :D

Re: Help us launch the network!
 

The intention was for both, but the reality was we only used to for picking partners most of the time, because we didnt have a good enough system of running the info back to the drive team until our 3rd regional. Then we used a few times (once even to call our alliance member on their abilities! eep!)
We do two sets of scouting, one is "pit scouting" where we get basic info on their robot and their claims of what they can do, the other is "match scouting" in which we have people watching the matches and noting everything they could about what the robots did, the scores, etc. We also ended up developing a third ranking system (I think it was based on 73's) where we had one person sit and watch nearly all the matches, and they would rate each team in about 15 different categories on an excel sheet, each sections were given weights, and then the total gave us our own internal rank. So we pretty much watched all the robots for what they could do, how many they could stack, how high they reached, etc. In addition to getting notes on robot description features.
There was quite a lot for us, especially because we learned after every regional. By nationals we just did pit scouting, the excel sheet, and a reduced version of the match scouting which mostly covered how many tetras could be stacked in one match by each robot and notes on it. So every time we headed off to competition, we spent probably 10 or more hours of prep.


Everyone on our team gets involved. There are about 5 people on the strategy/drive team plus myself that lead it up, the rest of the team is expected to sign up for times to do one of the three types of scouting we did. It depends on the number of people we had at each competition, but my guess is that everyone put in at least 3 hours of scouting.

All of them. At Rochester I think it was 32, Cleveland 43, Toronto 64 (I think).

I sort of mentioned this, we tried to get our pit scouting done on thursday during practice rounds. All thursday and friday we watched matches and did match & excel scouting, then saturday if we were in good standing (which we were at all 3 regionals) we continued scouting on saturday morning.
By the end, we found the scout sheets ineffective and cumbersome. The most useful was the excel sheet of rankings, second was the teams ranked by the number of tetras they actually scored in a match, third were the pit sheets to get a general idea of what the robot did. Along with the pit sheets, we also took photos of every robot so we could remember what they looked like. That was really valuable.
In the end it was 60% computer, and 30% pit sheets.
I would say ours worked well for us in the end (as we learned). We spent a lot of time after the first regional trying to go through our pounds of notes to handrank each team, and that kind of worked, but by cleveland and toronto we had a really good ranking system. We even used our rankings in toronto to help the second place team (we were selected by the first) to select their 3rd partner! :)


Cyberguy, this sounds like a really awesome idea, and I would bet my strategy/drive team would love to help. I will suggest it to them, it might be after school starts until they get back to you, but its a really great idea :) Keep up the good work!

Re: Help us launch the network!
 

You might check these links, many answers and suggestions talked about a lot.

http://www.chiefdelphi.com/forums/sh...ad.php?t=34772

http://www.chiefdelphi.com/forums/sh...ad.php?t=37693

http://www.chiefdelphi.com/forums/sh...ad.php?t=37216

http://www.chiefdelphi.com/forums/sh...ad.php?t=35916

http://www.chiefdelphi.com/forums/sh...ad.php?t=33457

http://www.chiefdelphi.com/forums/sh...ad.php?t=34430

And those are just on the first page in the "Scouting" forum. Check them out.

Scouting is for anything and everything and I feel it is an extremely important part of doing well during matches. It's for watching team's consistency, their strategy of play, to size up your own strategy, to know what a team can or can't do, regardless of what they tell you in the pits, to strategize with your partners and against your opponents, etc.

Ours is done on paper, we have tried other methods, but always come back to that. We do pit scouting (to get the technical aspects of every team's robot), and we do match scouting (to get the play aspects of a team and their robot). Match scouting sheets change each year to fit the game. Pit scouting is done on Thursday. Match scouting goes all day Friday and Saturday until elims. Depending on what type of scouting you do and how detailed you get, you can have one scout per robot on the field during any match.

Our scouting is effective. It is important to know other team's capabilities, strategy, and consistency. I feel it is not as effective to only know straight statistics on team's performances, as it is to know the things I mention in the first sentence. Standings are not as important because they could just be unlucky with partners, but have a very well performing robot of their own.

The key is, if you're in alliance picking, to choose a robot that is COMPATIBLE to your own, one that maybe does something you can't do. So you would figure out what that is, and you might scout those teams closer. Rankings, and objects scored won't tell you that kind of info... (not saying we don't use that info, but it's not always the highest priority info). I think most teams scout every other team, and maybe narrow it down the second day.

We never give up scouting. No matter what our standing is going into Saturday, you never know what will happen, and if a team chooses you, you could be a hero for having all kinds of detailed info on your opponents. Plus, it gets the students (and adults) familiar with other teams. You might just see those robots again someday.

Re: Help us launch the network!
 

That's because I made this thread about one.

The alpha of the alpha is available for testing here: http://niscout.zapto.org/scouting/. That isn't the most updated version, but it's the one that I currently have access to for viewing.

Re: Help us launch the network!
 

Wow, thanks for all of your info and your help people. It's been a big help. We're going to try and have a weekly AIM session to discuss progress, ideas and such with this project sometime soon, but the details still need to be worked out.

Oh a albiet of clarification on the how many question, I know most teams scouted all their bots in their regional, what I'm asking is how many bots were at that regional that you scouted :)

Thanks for posting guys, and I'll be contacting those interested in the project shortly with details. This is going to be really cool. Keep a-postin'!

-Chris

Re: Help us launch the network!
 

Well we've gotten quite a bit of interest and are now working on plans for the the next phase of the project. Currently we've got a website domain and server space, and our army ( I wish) of programmers are working on the coding.

We've gotten permission from Brandon to reopen the FSN sub-forum, so future posts on the project should be located there.

We still need help with many aspects of this, but things are rolling along quite nicely. Thanks for all of your help people!


-Chris

Re: Help us launch the network!
 

For the last couple of years we have been scouting using customized Excel spreadsheets. They have a plethora of input boxes and selection menus to enter all kinds of information from matches. Usually we have two people scouting and at the end of each match, the data is combined onto one, master laptop. Our focus in scouting is both for picking partners in finals and for strategy in qualifying matches. We typically only scout a robots performance: how many points were scored, did autonomous work, did the robot get stuck or tip over, etc. We will scout all the matches on Friday and Saturday up until the final matches at which point the drivers and coach use all the data to decide who we want to be allied with. The scouting is quite effective, but a bit cumbersome because the data has to be manually transferred to the master laptop.

This year we are looking into using a server and database type setup, awfully similar to what has been posted so far. We have not decided if we should continue the Microsoft route and use Access (easy to make pretty forms) or go all out and snag a Linux box with Apache, Perl, MySQL and build everything from the ground up on a web interface. In the efforts of not duplicating work, I would like to get in on this project. One concern I've had all along, how can we put together a wireless network large enough to encompass the pits and the stands around the playing field? Several wireless routers linked as repeaters?

Cheers!
Brandon

Re: Help us launch the network!
 

What is the main focus of your scouting, is it for strategy? For picking partners in finals? A little of both? Something else entirely?
Both

What information is important for your scouts to collect? Drive train, arm info, etc.
We collect detailed match data for each robot.

How much set-up time goes into preparation for your team's scouting?
20-40 man hours

How many people are usually involved in the scouting, and how much time are they expected to put in?
We have 2-3 developers and 4-6 scouts on matches.

How many teams do you scout at your regional?
All.

When is your scouting usually "done" by? Is it ready on Friday for the first match, or is it only used in finals? Somewhere in between?
We start the first match day.

What methods do you use for your scouting? Randomly scribbled notes? Fill-in the blank scout sheets? Statistical analysis? Match Records? Something else? Extra feedback here would be very helpful.
Statistical analysis

How much of your scouting is done on a computer? (Includes those print-offs)
All

How effective/successful is your scouting?
Our scouting allewed us to sweep the PAC NW regional undefeated.

FYI: There is a new scouting system based on stat analysis being set up, check it out at stampscouting.org