ok this year was our rookie year and i saw alot of teams with a complete(or near complete) list of every other team with their average points per game…we tried doing the same thing but ended up with like a jumbled mess and some teams were entered many times…
i personally thought excel would be a good enough program… but that didn’t work too well
so any suggestions on how t keep large amounts of data organized and easy to recall when needed…
Many teams used PHP and MySQL based systems. Some teams use FileMaker databases. Some teams use paper based systems and keep file boxes.
Personally, I think Excel is a fantastic program for scouting. Anyone can use it easily.
I go down the left side with team numbers, and across the top with matches: “Team XXXX’s 1st Match, Team XXXX’s 2nd match,…” For each match, I have a set of say, 10 data points. Autonomous lines, laps, hurdles, ball knocks, ball place, etc. So if each team plays 10 matches, I would have 100 columns of data. Personally, I prefer not to sort or average data. I like to spend hours evaluating it on my own, learning about each team, and seeing what their strengths and weaknesses are. In addition to the numerical data, I am a big fan of subjective comments. I sometimes use colors to highlight things of great importance. Then I create my own subjective rankings. I’ll be honest here, a little bit of the “ranking” process is not pretty. Friday night, I create a “possible” list, an “absolutely not” list, and a “if our possible list isn’t long enough” list. The “absolutely not” list usually ends up the longest. There are many robots that just do not fit our game strategy. An interesting note is that teams can actually end up on my “absolutely not” list for reasons not related to their robots at all, such as a complete disregard for gracious professionalism. The machine itself is not the only factor in our subjective rankings. After creating these lists, we rank the possible list by order of preference based on several factors. On Saturday morning, we closely watch all the teams on the “Possible” list, and make any adjustments in our rankings, and any eliminations from or additions to the list. We spend about a minute or two in each team’s pit looking at their robot up close and observing their team, to make sure there aren’t any reasons we shouldn’t pick them. This year, our ~27 team “Possible List” friday night was accurate enough where we had to make only two preference ranking adjustments, and no additions or eliminations. Another interesting note, win/loss and qualification ranking plays absolutely ZERO role in our picking preference. We don’t even record win/loss. Win/Loss is based on alliance partners and opponents. If you need to rely on these for your scouting, I’d venture to say your scouting does not go in depth enough.
In data collection, I developed a system this year that I believe worked absolutely fantastic, that I haven’t ever heard of before. I was able to scout all of the robots in every match we watched with only the assistance of two other people. I call it the “verbal call-out” scouting system, and i believe it has many advantages over other systems. I had one person keep their eyes fixed to the red alliance, and one person keep their eyes fixed to the blue alliance. These people did not take notes and did not look at computers. They were always watching. By this method, only very rarely was there ever a scoring action that they did not see. I sat at a laptop, with a simple Word document open, and a line for each team. My eyes were glued to the laptop in times of intense scoring, and at other times served as an auxiliary check in case the observers missed anything. The two observers would yell out to me every time anything happened on the field: “968 Hurdle. 60 Line Cross. 233 Ball Knock. XXXX dead on field” etc. With each verbal call-out, I would move my mouse cursor next to the team number they called, and depress one or two keys on the keyboard for each action. autonomous line was “a”, hurdle was “H”, autonomous knock was “AK”, teleoperated knock was “TK”, line cross was “L”, place at the end was “P”. So, at the end of each match, for each of the 6 teams I’d have a string of characters that looked something like aAKaaTKLHHLHP. Then I’d quickly count up the total each item in my head, and input it into the proper cells in the spreadsheet. This was the only tedious and inefficient part of the process. However, we were consistently able to accurately scout a majority of the qualification matches on Newton by this method, using only three people total.
Ultimately, you’d have the keystrokes automagically alter the database. But that is more advanced than my skills.
A final note. Perhaps the most valuable tool in scouting is taking a high quality photograph of each robot. Makes it much easier Friday night, rather than saying “yeah, that one with the orange bumpers.”
You might try the Scouting subforum for some examples. I think some teams (like 1114) have posted data sheets in the Whitepapers, too. Personally, I’ve set up a few sheets and rigged Excel to do all the math for me once data’s been entered. At least one other team has it do that as the data is entered, but I’m not sure how to do that.
Here is the link to the thread about the scouting system our team uses. It is excel / labview.
http://www.chiefdelphi.com/forums/showthread.php?t=65300
The one advantage I notice is that it makes it fun for the team to scout, which is good since everyone gets a turn.
In case you haven’t found them yet, here are some excel based scouting tools that are available in the whitepapers that successfully conquer the organization problem.
http://www.chiefdelphi.com/media/papers/1916
http://www.chiefdelphi.com/media/papers/1954
http://www.chiefdelphi.com/media/papers/2110
http://www.chiefdelphi.com/media/papers/1806
A key thing to keep in mind as you select a technology is what functionality you want to achieve with it. Tracking text based comments will drive you to a significantly different decision than numerical data that suggests a pick list to you. If we can get a little more information on your vision of the application we can probably make some more relevant suggestions.
Personally, I think that excel is more than adequate for most scouting tasks; but then again, I’m an excel junkie. 