Defense Type and Position Selection

After getting through one competition and getting the chance to pick defenses in many matches against several different combinations of alliances I was wondering what strategies teams are using for selecting defenses and defense position.

What is everyone’s first choice if there is no scouting data on teams. What do more teams struggle with? Our thought process was that if all else is equal, put the RockWall, Drawbridge, Cheval, and Ramparts on the field, but we just felt that those are the defenses more teams would struggle with, it wasn’t based on data. Does anyone have some good data on which defenses from each category tend to lead to less breaches?

Then when we are thinking about field vision the choice of Portcullis vs. Cheval and Drawbridge vs. Sallyport came in to play.

What other considerations do people have for their selections of defenses?

The other thing that I have been thinking about is defense position. The things that came into play when we were selecting defenses was field vision, but also what routes teams would have to take to do their scoring. We wanted to make their routes as long as possible and create as much congestion as possible for the opposing alliance. The only other thought we had was to put the sallyport in position 5 whenever we put it on the field. The logic is that it is much harder to get into the sallyport if you don’t have as much space to approach from the side.

Do people have any other thoughts on picking defense position or anything at all with regards to defense selection?

At NYC our go-to “no-data” defenses were sally port, moat, rock wall, and cheval de frise. It seems that teams had a much harder time collaborating on the sally port than the draw bridge, and often got caught on the moat/rock wall. The cheval de frise also seemed to be more challenging for teams than the portcullis. I also usually made sure to put the sally port in position 2 so teams couldn’t bowl over the position 2 defense.

I haven’t developed a good opinion on defence position, but for defence selection, this is an excellent resource. If you scroll down to “Year-Specific Statistics” on that page, it has data on the number of defences damaged of each type throughout the year. Based on this data, the statistically hardest to breach defence picks are:

Cheval De Frise
Ramparts or Moat – data inconclusive as of week 3
Rock Wall

For the argument between the Ramparts or the Moat, I tend to believe that the Moat is more likely to cause damage to enemy robots (i.e. comms lost, mechanisms breaking) whereas the Ramparts takes longer. If picking blindly, I personally would pick the Ramparts.

As for the Sally Port vs the Drawbridge, the drawbridge is statistically crossed much less often. In spite of this, teams have been picking the Sally Port more than double the amount that they pick the Drawbridge. In the end, it comes down to preference and scouting.

Of course, those defence picks are based solely on statistics. There will definitely be cases in practice where certain alliances will be completely trumped by the Portcullis, or for which the Drawbridge is easily managed. But if it comes down to a blind pick, those are my recommendations.

At NYC i noticed more teams not being able to do the port, than the Chivel. plus the sally really hurts the vision of small robots but allows you to still see larger bots.
The moat and rockwall really depends on spacing of the wheels, because thats that messes people up.
the rough terrain vs rockwall really depends on center of gravity of the robot/ ground clearance.

Also i feel the drawbridge requires more robots to have to work together than the sally port. Since the bridge is really flismy it is harder to open.

I would make another argument for Ramparts being better because they tend to turn robots as they drive over. Many teams can’t cross the Ramparts in autonomous because of this, which cuts 3 points off of your opponents’ score (2 point reach and 5 point teleop cross vs 10 point auto cross).

Yes! The position of defenses is not thought about nearly enough at events I’ve been to. Generally, here’s Dawgma’s process (oversimplification, of course, but there’s no one good way to pick for every match):

  1. Figure out what defenses are toughest for opponents to crack, preferably by match data. If unavailable, use pit scouting. If that’s not conclusive, go with the CdF, Ramparts, Sally Port, and Rock Wall.
  2. Position the defenses such that the defenses that one opponent will likely cross is right next to the one the other will cross in order to funnel your opponents. I consider long routes a bonus. Often if you’re faced with an alliance of 2 cyclers, what you can do is figure out which will use what defense, an alternate as best you can, depending on audience selection.
  3. Consider whether there are conditions drastic enough to merit use of the drawbridge. Are you severly outgunned in boulders? Make it so that your opponent’s big shooter can’t see when they’re lining up to shoot.
  4. If you wanted to use the drawbridge, think again. Are you sure? Are you really sure? Check with your drive coach.

After having gone through a regional- I don’t give a Flying Toaster about defense selection.

Unless my opponents are known to be really bad at something in particular, I’m probably going to put the Moat, the Rock Wall, the CDF, and the Sallyport in almost every time for the purposes of my own visibility.

If my opponents are bad enough that their alliance can’t get a breach done regardless of what I throw at them, I’m going to win the match anyway, at which point who cares what defenses I place?

People seem to get stuck on the moat and ramparts equally, and the CDF is both more difficult and easier to see over. The Sallyport might occasionally be replaced with the drawbridge if I really don’t want my opponents to have an easy 10 points sitting on the table.

Defense selection is cool, but lends very little strategic value to the game once you get to a point where every robot at the event can get a breach solo.
(And if you can’t get a breach solo regardless of the defenses placed, you probably messed up your strategic design from day 1, because robots near the bottom of our picklist at Tech Valley got solo breaches when their partners lost comm. There are a few exceptions to this, but for the most part this holds.)

I’m going to ignore the first part of the statement and focus on the second half.

All events are different. Just because at one event the bottom of your picklist has teams that can solo breach doesn’t mean everyone’s does. Honestly, at both of our events so far, we haven’t been able to put anyone beneath the 8th position at our picklist who can solo breach, simply because that’s who could solo breach at our events.

But my main point is this: just because a team can’t solo breach doesn’t mean they fail to be a valuable partner. Does it make it much, much easier to select someone who can? Of course, having that flexibility is great. But if I have a team who can score 6-7 boulders per match, and is only missing maybe one or two catagories of defenses, I think it’s worth considering. A special case? Maybe, but I’ve seen it more than a few times. They might not have failed their strategic design; likely, they might have just gone about it differently.

And the point about Auton by Paul Richardson is also a good one.

At the end, I did say there were exceptions. 4481 was at the Tech valley regional this past weekend, and they couldn’t solo breach, but they did hit 6 high goal shots in a match once. But they’re definitely the exception rather than the rule.

I might amend my original statement though- if you can’t solo breach when the Sallyport is on the field, you probably failed your strategic design from day 1. Not doing the low bar can make sense if you can do the bumpy stuff, the CDF/Portcullis, and do the sallyport backwards.

I can agree with that. With every robot being able to do the Sally Port with intelligent drivers, it basically allows you to choose a defense you don’t want to do (assuming you’re low bar capable).

Some smart teams put all the easy to cross defenses on one side so that robot auto modes can collide. Especially when the B or D defense is the audience selection. ::rtm::