Competition Scouting

[Ryan Dognaux]

Ohh okay so we participated in WASH last year. Gotchya. We'll be at Pheonix, Chicago, and Richmond, so some of us should definately be able to help with it again. I never knew about the live video thing SOAP provides, so that's pretty cool, I'll have to remember that.
Also, more people scouting means a greater accuracy, sooo .. the more the better, correct?

[jonathan lall]

Maybe this would interest you?

[Jack]

Originally posted by jonathan lall
Maybe this would interest you?

What? That takes you to this same thread.

[Brandon Martus]

It takes you to the 'Developing a standard..' thread that you started..

[Todd Derbyshire]

The key to scouting is determine what you think of the robot and not believe anything that the owner of the robot is telling you. It is human nature for a person to believe that they have a good robot. After all we all spend six weeks on our bots and no matter how miserable the performance is they still believe they have a good "bot," Here is how my team proceeded to work scouting for the past year. I will give this process to you on a day by day process. There were two separate methods that my team used I will give one for teams of a relative size and one of a team of six!!! Yes six people is all we needed to scout the UTC Regional and we finished second and were amongst the top 8 throughout the qualification rounds until the last match.
Thursday: arrive at competition EARLY!!!! Walk around the pit and send team members out with five teams to scout. My team had I believe at the team around 40 plus kids so we had alot of data collected. Also going into the event you should have three-five team members that are known as the strategist. These are the people in charge of making sure the info is right on the sheet and to compile all the information that night onto index cards. I believe we just used works but we made the program. The questions that we asked were what motors were you driving on, speed and etc.
Friday: Qualification rounds begin and here is where some problems with team's scouting begin. See there are many times where team will give a scout a inaccurate report. For they don't want to reveal there strategy quite yet. This happens and lemme tell you its quite prominent amongst teams and I practice it too. I don't want everyone to know what I'm doing unless your with me. And as for the statement: By using WASH we put our name out and get recognized is very false and absurd. Don't get me wrong I love the program however, I would much prefer to see what a robot can do then hear what a robot can do and take my chances. Just because a computer is telling you that a robot is rated a 2 on speed doesn't explain when that same robot beats you to the goals. Getting back to what is suppose to be happening again is you have to prime scouts watching matches and making a list of top bots. Last year there was two catergories Hybrids, and goal-grabbers. Using this list compile a top 10 picks for a draft. Then make another list of five teams that you believe are underated and you can steal in the second round of the draft.
Saturday: Have your scouts watch the matches exclusively for changes on the robot and make sure you get an infirmary report from all the teams that are on your list. This is how the chaos is managed with a large team.

For a team of only six people...(Yah six people including the drivers!!)
Thursday- Divide all the teams by six and begin the interview process before practice rounds start. During practice rounds watch two of there matches and then compile all the teams info into a notecard.
Friday- Drive team and strategist go into pit while two people watch the matches and check info. Once again changes are made to the scouting reports for accuracy. That night compile the list and repeat the process for Saturday by checking on all the teams on your list.
The key to making the six person team work is focus and knowing what to look for.

Lastly, another idea for scouts out there is if you know who you are playing ahead of time watch their previous match. You never know what you will find out from it.
Ok I'm done this post is way too long

[Gope]

When videoing, I would reccomend not videoing every match. We just video really good teams that we will be picking as teammates in finals, or that will be picking us, or that we need to work out a strategy to beat. You have to remember that there are alot of matches and if you video each one and watch them again, then you dont get any sleep(BAD).

Also, multiple angels on some matches(ones with 3 or more really good robots) is a must.

Originally posted by Gope
When videoing, I would reccomend not videoing every match. We just video really good teams that we will be picking as teammates in finals, or that will be picking us, or that we need to work out a strategy to beat. You have to remember that there are alot of matches and if you video each one and watch them again, then you dont get any sleep(BAD).

Also, multiple angels on some matches(ones with 3 or more really good robots) is a must.

What defines a good team though? Just because a team doesnt have a reputation of being a powerhouse doesnt mean they arent good. Scouters need to be sitting and observing every match possible. Shift off on the hours and have them record notes and fill out little forms for each match. You can take the results of each match and compare them for who you think could be a good partner.

Reputation isnt everything...

[Chris Nowak]

All hes saying is that its not a valuable use of time to scout robots that would be a bad pairing for you in the finals. For instance, in 2003, it would not have been useful to scout a ball pickup robot that couldnt get any goals if you were a ball pickup robot who couldnt manipulate goals because there would be no way to manipulate goals with them as your partner. You would scout out the teams which most complemented your style of play. Of course, you also do scout the ones you are up against or have to be paired with in qualifying matches.

[Mike Soukup]

Originally posted by Todd Derbyshire
And as for the statement: By using WASH we put our name out and get recognized is very false and absurd. Don't get me wrong I love the program however, I would much prefer to see what a robot can do then hear what a robot can do and take my chances. Just because a computer is telling you that a robot is rated a 2 on speed doesn't explain when that same robot beats you to the goals.

Agreed, to an extent. Teams should absolutely have their own scouts and know what each bot is capable of. Reading what a bot can do isn't the same as seeing it in action. But it's valuable to have the data too. I think you're confused, WASH doesn't gather data by asking teams what their bot is capable of, the scouts watch matches and record what the bot does on the field is actually capable of. I've insisted that data be collected this way from the beginning. I too don't trust or believe what anyone tells me in the pits. Most people (unknowingly) overstate the abilities of their bot.

Mike

Originally posted by Chris Nowak
All hes saying is that its not a valuable use of time to scout robots that would be a bad pairing for you in the finals. For instance, in 2003, it would not have been useful to scout a ball pickup robot that couldnt get any goals if you were a ball pickup robot who couldnt manipulate goals because there would be no way to manipulate goals with them as your partner. You would scout out the teams which most complemented your style of play. Of course, you also do scout the ones you are up against or have to be paired with in qualifying matches.

In previous years, there have been teams hiding somethings they can secretly do. In the finals of the Great Lakes Regional of 2000, we were beating the Chief Delphi allience, until they went to our goal and stole balls out of it. Nobody had any idea that they had the ability or driver skill to do such a thing. It proved very valuable and won the regional for them. In other words, just because they do not show the ability to do something right away doesnt mean they cant. Thats the sole purpose of scouts. Ask around, you might learn something that other teams don't know and pick a team that could be a high asset that nobody knows about.

[Jack]

Regarding how data is collected. Yes, having data that is dependend on match data / an actual test is much better than having a team tell you what they can do. However, keep in mind that teams might have 'hidden powers' that they don't tell anyone about. BUT... if you gave them a place where they could post what their robot can do, then they might at least say that they have something special.

Also, with general data. Even if you do have subjective data, if you have lots of people putting in a vote, then the overall answer would be pretty good. But... yes, match data is better, and once again, the more people that add to it, the better it will be.

For the 2003 FIRST season, www.feds201.com will be doing something similar to what other teams have done in the past, but with many improvements and at one cental site (With options to include data from other sites too though).

(Read my first posts and other scouting ones for the beginning of the plan) As said before, the 'Team Profile' would all teams like 47 to say that they have a hidden defence device, but not what it is. This would assist the scouters at the competition because they could look at the robot and figure out what the 'special feature' is. Also, the picture section allows teams an easy way to display photos of their robot/team. The team forum section allows comments to be made by anyone so that to say something about the robot, you don't need to register

However, I have also made many addtitions to my original plan. As posted, we know that every team will make their robot sound the best. For that reason, just having the 'Team Profile' section in not good. On the left side of the team profile, there will be a quick poll section. This will allow registered members to vote on key items.

Next, there will be a 'match data' section. This will allow members to enter data about each match that occures and what each robot did thought a number of pre-defined (and mabye one 'free') fields. Yet... we all know that having to sit around in front of a computer after a regional (due to lack of internet access) and fill in team data into an HTML form is not fun. Thus... the next feature.

Rather than entering in match data via. a HTML form, scouters can dump data from their palms into a CSV file. Then they can upload that CVS file and the data will automatically be entered for you. For the teams w/o palms/laptops: You still will be able to upload data via. the HTML form, but you could also create a CSV file in Excel or Notepad, and then upload that file.

Once again, because there are few internet connections at regionals, you will be able to 'dump' data on the teams for each regional before you go. The data would include the Team Profile, the poll results, and all of the match data. This could be 'dumped' in a CSV file so that it could easly be uploaded to your palm, or printed out in a table format from Excel. For those wanting to use the data on their website or just would like a different format, you will be able to dump in XML format.

Being able to upload/download mass ammounts of data of coures requires that a standard is created. Once the game is announced, i'll post here to create a standard. The standard will need to decide what data will be collected (Ex: What fields), and in what order will the data be uploaded/dumped. Of course, If though the season we find that a field has little value or we need to add a field, that will be able to be easily changed. From the rumor that this years game will involve 'flying' objects, I think we'll have quite a fun/tough time decideing on what needs to be collected.

If you have any suggestions on how to make this better than what i have, please post. I want the design to include input from the people that will be using it most. I also want the system to also attract non-scouters who are just interested in robots and other teams. However, please wait to suggest fields untill after the game is announced.

Thanks for reading this really long message

Jack (www.feds201.com)

[Stephen Kowski]

Since our scouting program was mentioned I thought I'd give it another cheap plug here. We plan to have it packaged and be able to distribute it to teams that want it(essentially managing their own data if they like), or they can migrate their info with ours. This year it will be compatible with Pocket PC and the Palm OS, not just Palm. This information we will be making this information available online viewable by ASP.NET, and of course all the old features will be updated for the new game...picture database, mini drawing board of the field, etc etc. So not only online, but in your handheld you'll have everyones info.

[Jack]

Originally posted by Stephen Kowski
Since our scouting program was mentioned I thought I'd give it another cheap plug here. We plan to have it packaged and be able to distribute it to teams that want it(essentially managing their own data if they like), or they can migrate their info with ours. This year it will be compatible with Pocket PC and the Palm OS, not just Palm. This information we will be making this information available online viewable by ASP.NET, and of course all the old features will be updated for the new game...picture database, mini drawing board of the field, etc etc. So not only online, but in your handheld you'll have everyones info.

That sounds like a great idea!

Because of the design of two systems with the idea of sharing data in mind, both systems will be able to share the data. Of course, we (and all the scouters) will need to deicde on what data we want collected. That should be one of the first things that we do once the game is announced.

[soap108]

Originally posted by David Kelly
they do real time video, stats and scoring at the competitions that SO@P goes to. there's lots of interesting and useful data that they compile.

Video
------------
At the Florida Regional we plan to video record every match and post on the web. If we obtain an Inet connection at the event (likely) then the files will be uploaded ASAP. If not, then they get uploaded by the following Tuesday. This year the video format will be MPEG4 and all matches will have audio as well. We tape the FIRST Video feed, meaning the 2:00 clock and any other info that gets super-imposed (i hear the match number + maybe the unofficial score will be on the video too) is what we capture.

We have same plans for Midwest Regional and hopefully Nationals. Additionally at MidWest, we hope to stream the event over the Inet. Last year when we did this we recorded over 15,000 hits!!!

Last year we created a user guide and setup procedure so that another team at a regional out of our reach could capture video just like us, provided they upload the video to our site. This proved very successful. The SOAP website was able to house video from Great Lakes, New England, and St. Louis thanks to partner teams 111, 65, 67, 177, and 343. We hope they will be willing to provide similar service for all of us this year too!


Match-Scouting Data
-------------
SOAP got out of the watch-a-match-and-write-what-happened business with the onset of WASH. Team 111 heads up that effort along with a dozen other teams.


Non Match-Scouting Data
----------------------
What we still do though is provide an on-line u-fill-it-in form. This "SOAP Robot Profile" is filled in by anyone about anyone. It has fields for "how many balls/boxes can you pick up?" ; "what are your strengths / weaknesses" ; "What is your robot weight" ; etc.... The best we can do for accuracy is require a checkbox to be selected which states that the person entering the data did so with "gracious professionalism".


Scoring Data
--------------
In 2002, SOAP was able to obtain the offical scoring data from several events complete with the details (#balls, #goals, etc...)
Hopefully this will again be the case this year. Once the data is entered into our db, stats and charts are immediately available. Also we compute ranking/seeding just like FIRST.



Regards,
KA-108 www.soap108.com

Team 108 - SigmaC@T
Motorola & Dillard HS & Taravella HS
Plantation, FL

[1337 /\/\4573|2]

I think its a bad idea to try and rely on other team's scouting efforts.

I remember last year at a competiton 9 other people and i went out to scout, and that was our only task.

It was really helpful when the actual people who were doing the scouting could go back and clarify some data that might have been misleading to others.

Basically what |'m trying to say is that data collected by one's own team could be more helpful than data taken from another source. Also, when a person collects data, he/she can include data that is relevent to his/her team (ie would this robot be a good pick for a final alliance)