[platypus]

I posted these thoughts after day 1 at the BAE Granite State Regional:

My thoughts after scouting all day Friday:

1. Shooting is extremely difficult outside of autonomous. Autonomous is vital this year for the very fact that it's the only time your robot is not covered. After autonomous, the other bots are all over you if you're any good, and since a gentle knock is enough to make you miss every shot, a full push will keep you from ever scoring in manual.

As a side-note, anyone without an autonomous would be very wise to write one that just moves forward, because, as proven by my team (166), knocking a team that shoots well in autonomous can win games by stopping a 20-30 point swing (and often changes the winner of the autonomous bonus).

2. Any robot with a scoring autonomous is probably worth picking (with a few, very obvious exceptions). Any pusher that dumps its load in autonomous is strong, possibly picking, and any shooter that can score consistently well in autonomous is almost definitely picking. Additionally, if a team can score in autonomous, then it's a safe bet that they can score decently enough in manual (though shooters will have difficulty doing so unless your opponents actually leave them unblocked). If you're picking teams THEN FOR GOD'S SAKE, PICK THESE TEAMS BEFORE THEY'RE ALL GONE.

3. Autonomous is big. REALLY BIG. So big, that this is my third blurb points out just how big it is. These 10 seconds tend to sway the entire game.

4. The scores are LOW in this game. I expect them to be much higher on Saturday during the elimination rounds, but during the qualifying rounds, scores were generally no higher than 30-40, and anything above 50 almost always won the game. This means that getting all three robots up the ramp is indeed very valuable, though let it be known that the ramp is VERY defend-able and the ramp bonus rarely swings games unless all three robots climb it for the full 25 points.

5. Despite the fact that scores are low, penalties rarely swing games. All the relevant ones are 5 points, and unless team incurs a lot of them (by doing something silly like not having a backbot for a while), then they usually don't change the outcome of the game. Last year, penalties were swinging games left and right because they weren't particularly hard to incur (especially that 30-point one for knocking a robot in the loading zone). THIS YEAR IS DIFFERENT.

6. Push bots are good if they work, though many push bots have difficulty unloading balls and (believe it or not) accurately. The ones that work well though are consistently (something shooters have trouble with) scoring 10-20 points (which is very sizable in many matches). In the absence of an accurate shooter in autonomous, they can win autonomous easily (with an autonomous mode so easy to code that even the bad ones have it) for the 10-point bonus.

7. Balls are everywhere, because human players that don't have to load bots up try to score in the corner goals as fast as they can (and sometimes score 5 or more points if they're lucky). Balls are literally everywhere, so if you have an efficient gatherer, you're in luck, and if you manually load your bot, you're at a disadvantage.

8. Pushing power is important but not as much as one would expect. Since you only need to knock shooters off target, you don't need to have the world's strongest robot. It sure helps, but you can easily get away with just being strong in general.

9. Your 10 ball starting balls are important. Use them well. No matter how efficiently you can load your bot, you'll probably never collect more than 5-10 (and that's really stretching it) balls regardless of which way you load.

10. Small play mistakes often become big, game-changing mistakes. 'nuff said.

11. Most shooting autonomous modes simply move forward and shoot. No fancy sensors or cameras needed. There is a real problem with consistency and the fact that lining the robot up poorly (and that doesn't take much) can make you score 2/10 instead of 8/10, but in general, these simple autonomous modes have the effect of giving the opponent a nearly insurmountable disadvantage score-wise.

After Saturday, here are my additional comments/revisions:

1. Like I mentioned after Friday, autonomous is so important for the very reason that you aren't covered. The only exception to what I posted before is what team 1276 did: they climbed the opponent's ramp and shot from there. This gave them perfect accuracy, while at the same time making them undependable. Many teams could not climb the ramp at all, and while they pretty much all could in the elimination rounds, none were really built to climb it easily (it's just such a deceptively hard task). They had absolutely forever to take their shots but at the same time did not need long to line up and take them. As far as I know, they only had a defensive autonomous mode, but it didn't matter because they just picked shooters with strong autonomous modes for their alliance, which let them interfere with the opponent's autonomous while their alliance took its own shots. They would be the only team to score anything of note in the rest of the game (as a result, their alliance swept in all their elimination matches easily to win the regional). My congratulations to team 1276 on solving this years game!

2. If this regional taught me anything about what it takes to end up picking on Saturday, it's simply that you need to be consistent. You need not score an awful lot, just that you must be a constant contributor. The number 1 seed at the Granite State Regional (151) was, to be brutally honest, not a team that I would even imagine picking. They were a human-loaded push-bot that tried to score via a low-mounted wheel on the side of their bot, and would generally only score a maximum of about 5 balls in the corner goals over the entire match (their alliance was eliminated in the quarterfinals 2-1). The key was that they would score 2-3 points in autonomous, play strong on defence, and climb the ramp every single game. This consistently scored a respectable number of points, and got them to the number 1 seed. My congratulations to team 151!

3. I was wrong in assuming that the scores would be much higher on Saturday than Friday. They may have been by a thin margin, but by no amount great enough to change the strategy of the game. The points I made after Friday pretty much all still apply.

4. I saw the games in the elimination rounds play out two distinct ways:
a. Both teams would similarly in autonomous, and the ten point bonus became a difficult but surmountable advantage for one of the alliances. The competition would come down to whether or not the other team could sneak through enough points, and getting up the ramp could sometimes swing games (usually to secure games, the alliance winning would have one robot defend their opponent's ramp, and it always worked). Usually, the team that won autonomous won these games.
b. One team scores 20-30 in autonomous while the other scores almost nothing. The winning team has a 25-30+ point advantage, and wins handily. In fact, the team that loses autonomous tends to defend so poorly (their moral seems understandably crushed, plus they are forced to focus on offense too much to have even the slightest chance) that the winners often win by a 40-50+ point margin and the only penalty that means anything is disqualification (which actually happened in the semifinals to one alliance).

It is my bold prediction that while teams will get better on offence, teams will similarly get better on defence, so most qualifying and nearly all elimination games will continue to follow these models for the entire season.







That pretty much wraps it all up. Good luck to all the teams competing at upcoming regionals. My team (166) will be at Atlanta, hopefully doing better than we did this weekend (we finally have a shooting autonomous just about worked out ). Please feel free to comment on my observations.