How do you decide your strategy?

[Andrew Schreiber]

Talking for a moment on what Don said.

One of the methods I have been working on employing more and more is making a simple simulation of the game using excel. I then sit and play with strategies. Figure out how to score the most points as quickly as possible. Assume this is what the top tier teams are going to do. You now have three choices. Beat them at their own game, find how best to compliment this strategy, or find how to stop this strategy. Obviously, you will always have your oddballs (71 '02, 469 '10) but those are exceptions. Though, you should try to come up with these game breaker strategies if you can.

The reason I'm working on this is because I, stupidly, missed the importance of the minibot this year. I let me ego get in the way of good strategy when I said, "I don't like the concept of a minibot, scoring tubes is where the points are anyway" I had a bad case of the dumb and didn't use numbers to make my decisions. My lesson from last year was: Use numbers to determine the bare minimum needed to build a competitive robot and then rank every additional task according to value. Focus on the bare minimum first because it is better to have a functioning robot that does 1 task than a half done robot that does 4.

I will second the suggestion of reading Jim Zondag's paper. I also recommend JVN's white paper on Weighted Objective Tables, and Karthik's presentation on FRC Strategies. Also, as much as I hate to stroke his ego Chris Picone is surprisingly good at figuring out strategies, shoot him a PM and see if he will walk you through his strategy.

[JesseK]

We ask ourselves "What Would Karthik Do?", "What Would Beatty Do?" and "What Would Baker Do?" to every brainstorm item. This year we may add "What would 469 do and how would 254 implement it?"

Well, that's mostly true. The team does a whole host of things, including CAD drawings on the fly to answer questions that come up, playing the game with humans representing robots, and drawings on a field-backed whiteboard.

[gblake]

Originally Posted by Grim Tuesday

How does your team decide on the strategy you will be using for the game? Is it something pre-determined, something that a small council or subteam makes? Something your mentors do? Or is it a whole team decision?
...

In part, I help the students ensure that they are developing a strategy and not something else.

In my parlance, a strategy answers this question, "In situation X, we will do action Y.", for all useful values of X. An example is, when to bunt.

A strategy is not a single script for what to do during the entire match. It is rules for what to decide to do moment by moment as the context around your robot/team constantly changes. no single script (or small handful of scripts) is likely to be adequate.

Blake

1. Make sure you understand the game rules, penalties, animation, and think about any possible rule change to come later on.

After that, depending on the size of your team you start making decisions. If you have a very large team, then you may just want a few representatives of each subgroup displaying their ideas. If your team is in the lower numbers, then you can have an entire group discussion.

After that, write down all your ideas, and a rough draft of the robot, and start again, this time from a different point of view. Maybe now you think of an elevator with a gripper attachment instead of a roller claw, or maybe you think of a defensive robot. This helps in brainstorming ideas that would have been thought of days later once you have established a final design.

Make sure you know any possible robot roles. You could be designing an offensive robot, a defensive robot, a herder, or anything else that fits the bill.

Once you get at least 3 good ideas written out, go over the advantages and disadvantages in each. From there, say what you think will be necessary on a good robot, and what role you will be playing. After that, take the best parts of each design that will benefit your robot, and there you go.