Advice for new strategist?

[Potatoes]

Hello, I am the newly appointed strategist for team 5136, the MechaPirates.

2014 was our rookie year, and we did not have a clearly defined "strategist" position. This year, I will be in charge of of the strategy team, running tests to see who makes the best drivers, lead the scouting team at competition, and be the one checking the robot every step of the way to make sure it complies with the rulebook. I've never done any of these jobs before, does anyone have any advice on how to handle the 2015 season?

[Michael Hill]

Karthik has given a talk for several years now called Effective FIRST Strategies. It's definitely worth a watch: https://www.youtube.com/watch?v=smWy7FQ8jLE

[Jay O'Donnell]

The best things you can do as a strategist in my opinion are:
-Analyze multiple ways to play the game and make sure your design team builds to be able to play multiple ways.
-know the limitations and abilities of all six robots on the field before a match.
-Try to be the team that goes to strategy meetings with a plan, instead of the team that just goes with whatever an alliance partner says, and don't be afraid to stand up for your opinion if an alliance partner wants to run a strategy that you don't like.
-Make a pick list no matter where you are seeded. Also know enough about every robot at the event to make a good judgment call during alliance selections
-Make sure you have scouting data that you will be able to easily look at and compare teams with.

[Max Boord]

Take both subjective and non-subjective scouting data to strategy meetings. It helps verify what teams are capable of and prevents over/ under estimation of teams capabilities.

[Tyler2517]

Drive teams matter more then you would ever think. And reliability and redundancy are a great thing over would work like gods but only 2/3 times.

[TDav540]

Karthik has given a talk for several years now called Effective FIRST Strategies. It's definitely worth a watch: https://www.youtube.com/watch?v=smWy7FQ8jLE

Following this video is extremely useful. It provides strong insight into every aspect of the position you have described.

Additionally, I would also recommend that you take as much relevant data as you can during scouting. Its better to have it and not need it than need it and not have it.

Lastly, make sure your scouters are having fun. I know that some years, the lead strategists have made scouting horrific for some rookies. It shouldn't be that hard, but just make sure you are catering to their needs some and not working them around the clock.

Good luck!

[GKrotkov]

Originally Posted by Michael Hill

Karthik has given a talk for several years now called Effective FIRST Strategies. It's definitely worth a watch: https://www.youtube.com/watch?v=smWy7FQ8jLE

Yes. This presentation is beautiful, and helps immensely. The section I found most useful for scouting/strategy is about 31:54 - 51:32. However, the whole thing is useful.

Also, some advice I can give:

- "Checking the robot every step of the way to make sure it complies with the rulebook." I certainly don't know how your team operates, but it's probably best if everyone is familiar enough with the rules and regulations enough that this isn't a needed responsibility. Of course, checking again can't hurt, so I suppose this is a partially garbage point.

- General Scouting/Strategy: Get a team and, go over the match scouting sheets with the people that will be scouting with you beforehand, so they can ask questions, like: "If they pick it up, drop it, and pick it up again, do we count that as two intakes?". Also, take some time after kickoff to analyze the rulebook and design strategies before you design a robot to fulfill those strategies.


- Pit Scouting: Do not scout functions. Or rather, do not scout functions with a finality in mind. Simply asking (using this year as an example) if they can shoot, how many points they score in auto, and if they can intake the ball is pointless. Your match data will tell you this. Pit scouting can be used for a variety of other purposes, and is best when filling in gaps that the match data can't. Some I've tried include scouting the drive team (not really successful as I've tried it), scouting strategy (diagrams work wonders), or scouting drive train (a function, but good to know. It's hard to scout this during matches). However, early in the event, it's the best data you have. Try not to exceed one page, there's a lot of teams to talk to.

- Match Scouting: Make a sheet and print it out beforehand. Number them beforehand as well. In fact, print out two copies of each and label them both. The worst possible thing for a scout is to think they have enough sheets printed for each team at the event, and then try desperately to find a printer at the competition while chicken-scratching data for three different robots onto the back of scrap paper (So much gratitude towards 708 right now). It should be one sheet, 12 x whatever-you-think-is-necessary. Be sure to add in attempts as well as successes for functions. Looking at a sheet and seeing that team xxxx shot (n) high goals in one match is good, but looking and realizing they missed (2n) in the same match might shed a new light on that data.

[pabeekm]

make sure your design team builds to be able to play multiple ways.

Sorry if I'm nitpicking, but I respectfully disagree with this note; in my three years in FRC I've seen only teams with a lot of resources/experience pull this off well. I've far more often seen teams fail because they took on too much when building their robot, have mechanisms that turn out to be hardly used in competition, or waste lots of time on field performing the roles that they won't be most valuable for in eliminations. I'd advise spending the first few days of build season doing heavy analysis of the rules to find what strategy you will get the most out of before even thinking about the design of the robot, then, build primarily for that strategy; being able to do one thing very well will likely get you more than being able to do multiple things adequately.

Make your pick list the night before alliance selection. You'll almost definitely be too busy to do it the morning of, and by that time you should already have a pretty good idea of what everybody can do and how well. This also means you can lighten up on the work for your scouters: reduce your scouting team to 2 or 3 people, give them a list of who's most important to watch, and have them just write down notes on any relevant happenings on field. Then, look their notes over briefly before alliance selections and adjust your picks accordingly.

As said above, always make a pick list no matter where you're ranked; if you're picked in the first round you can help out your alliance captain with the third pick. Also mark off the teams who are captains/are being picked during selections, because picks move fast and your captain might not be sure who's left. Compiling the list will also help reinforce in your memory who can do what, which helps a ton if you have to do some analysis on the fly.

[RunawayEngineer]

Originally Posted by Potatoes

This year, I will be in charge of of the strategy team, running tests to see who makes the best drivers, lead the scouting team at competition, and be the one checking the robot every step of the way to make sure it complies with the rulebook. I've never done any of these jobs before, does anyone have any advice on how to handle the 2015 season?

I would be wary of being your team's robot inspector while trying to be the scouting lead. If your robot goes under any changes at all, then you will be spending a lot of time in the pits - i.e. away from the action and information on the field. Your best bet would be to delegate either coordinating the scouts or inspecting the robot to someone else because you can't be in 2 places at once.
The most valuable thing that a strategist can gain - IMO - is direct observation of the matches at your regional. I always watch the first ~10 matches in a row without doing anything else because then I get to see at least 1 match for every machine (with some exceptions). With just some notes from those matches, I can provide the main points that the drive team needs to know for all of their qualifications. Of course, robots will change throughout the event, but match data, scouting data, and pit scouting go a long way to staying updated.
Watching matches also gives you information that you can't get anywhere else. For instance, this year it was extremely important to know what fouls were being called and for what - which changes from event to event and even from ref to ref. This informs you on what risks your drive team should take during a match. In theory, your scouting team will get the fouls recorded, but the context is extremely important.

[cadandcookies]

Originally Posted by pabeekm

Sorry if I'm nitpicking, but I respectfully disagree with this note; in my three years in FRC I've seen only teams with a lot of resources/experience pull this off well. I've far more often seen teams fail because they took on too much when building their robot, have mechanisms that turn out to be hardly used in competition, or waste lots of time on field performing the roles that they won't be most valuable for in eliminations. I'd advise spending the first few days of build season doing heavy analysis of the rules to find what strategy you will get the most out of before even thinking about the design of the robot, then, build primarily for that strategy; being able to do one thing very well will likely get you more than being able to do multiple things adequately.

I respectfully disagree with your respectful disagreement. There's a distinction between building to play multiple roles and designing a "multitool" robot. This year, for example, a relatively simple robot with a roller intake and a surgical tubing catapult was able to play a large portion of roles in the game.

There is a huge difference between a mechanically complex robot and a strategically-- most teams can't pull off a mechanically complex robot well, but the more you can do with a simple mechanism the better off you are. True flexibility is something I think is often undervalued, and even less achieved.

[BigJ]

Regarding being the "rules manager", I agree with RunawayEngineer here. Each person on the team (at least those involved with designing and building the robot) should understand the rules and the importance of following them.

[Andrew Schreiber]

Originally Posted by cadandcookies

I respectfully disagree with your respectful disagreement. There's a distinction between building to play multiple roles and designing a "multitool" robot. This year, for example, a relatively simple robot with a roller intake and a surgical tubing catapult was able to play a large portion of roles in the game.

There is a huge difference between a mechanically complex robot and a strategically-- most teams can't pull off a mechanically complex robot well, but the more you can do with a simple mechanism the better off you are. True flexibility is something I think is often undervalued, and even less achieved.

Nah, I'll respectfully agree with both you at the same time...


pabeekm is correct that focus on building a mechanically simple robot that excels at one role is a key to success. Assuming you can properly identify a need that you can excel at. For example: this year your average 4xxx team was not going to excel at scoring compared to top teams at the event. [Cue the plethora of NUH UH TEAM 4YYY WAS THE TOP SCORING TEAM AT THEIR EVENT that I really don't care about so save me the hassle of ignoring you] but the inbounding/trussing roles were pretty open. And what's more, they both required similar systems (notably, an intake) ...

In this case (and in ones similar to it) cadandcookies is correct. Identifying key mechanisms that can enable you to play multiple roles and focusing on them being reliable and effective is important. Many games have mechanisms like that. Usually they involve active control of game pieces. Focus on minimizing driver line up and timing. Big wide rollers work wonders, forklifts suck. [Cue the plethora of NUH UH TEAM ZZZZ WAS THE TOP SCORING TEAM AT THEIR EVENT AND THEY HAD A FORKLIFT that I really don't care about so save me the hassle of ignoring you AGAIN]. I'll leave the exercise of finding these mechanisms over the years to you, but I'll give you a hint: It's almost always involved in possessing the game piece. (I also don't care about the counterexamples someone is going to try to prove me wrong with).


Want the TL;DR of it? Keep it simple, effective, and easy to use.

[Citrus Dad]

Originally Posted by cadandcookies

I respectfully disagree with your respectful disagreement. There's a distinction between building to play multiple roles and designing a "multitool" robot. This year, for example, a relatively simple robot with a roller intake and a surgical tubing catapult was able to play a large portion of roles in the game.

There is a huge difference between a mechanically complex robot and a strategically-- most teams can't pull off a mechanically complex robot well, but the more you can do with a simple mechanism the better off you are. True flexibility is something I think is often undervalued, and even less achieved.

I agree with both points, but the single most important factor is choosing a set of achievable functions. Don't try as a rookie team to build 1114's 2008 mechanism! We did see a lot of effective simple catapults this year which were quite adequate during the qualification rounds. But pick a primary function appropriate for what role you think you will be playing in eliminations and make it as perfect as possible. 973's success as a goalie bot is a good illustration.

Which brings me to the second point - be prepared to give up your "favorite" function to join a successful alliance. 1640 had an excellent offensive robot, but we asked them to play inbounder and defender at which they were equally successful.

[Citrus Dad]

Originally Posted by Potatoes

Hello, I am the newly appointed strategist for team 5136, the MechaPirates.

And how foolish I look in my last post - of course 5136 was on our Newton alliance and knows exactly what role 1640 was playing!

5136 is another great illustration of how build the right kind of robot. Yes they could score during qualifications, but their value was in defending and inbounding which they did extremely well.

I'm not sure that you need much move advice!

[Firgreen]

Use the software that FIRST gives you/free software online. FIRST gave out five copies of Tableau 13 software last year, which I seldom saw teams use in the competition.It complements Excel as a game analysis software; it's game changing. Although it has its many advantages and disadvantages compared to Excel, it can save many hours which otherwise would have been used to construct visual basic macros/excel spreadsheets.

I can share with you additional information/what we did with it, if you'd like, just message me.

http://www.tableausoftware.com/first-robotics