How do you scout Defense?

[kstl99]

[quote=BJC;947922]This year the best kind of defense is to have a robot that can kick all the balls out of the far zone.

I agree with BJC. The faster you can get the balls out the better for your defense and your offense. I have seen many robots block the opponent effectively only to have another come into the zone and score the balls that have accumulated into the unblocked goal.

[fuzzy1718]

Just to keep me sane, let me recap. Some of you are saying that kicking balls out is the best defense. Others are saying there is nothing you can keep track of with numbers about defense, and yet others are saying to compare avg balls scored to what was actually scored in the match for the team that is being defended. Did I miss anything?

Oh and Lil Lavery, if you think defensive teams are a dime a dozen, then you obviously havn't been paying close enough attention. Picking the right defensive bot has made and broken many finals alliances. These teams were diamonds in the rough at the districts my team has been at, and we are kicking ourselves for not picking them.

[Sumathi]

It depends on the type of robot you use, and your game strategy. For example we play offense so we work around that. Anyways things to scout for would be speed, sturdienss, manuverability, how they move balls across the field, the drivers themselves, what the robot was built for, also add on qualities that your drives feel are necessary in a defender.

[carbuff2228]

Originally Posted by fuzzy1718

Just to keep me sane, let me recap. Some of you are saying that kicking balls out is the best defense. Others are saying there is nothing you can keep track of with numbers about defense, and yet others are saying to compare avg balls scored to what was actually scored in the match for the team that is being defended. Did I miss anything?

Oh and Lil Lavery, if you think defensive teams are a dime a dozen, then you obviously havn't been paying close enough attention. Picking the right defensive bot has made and broken many finals alliances. These teams were diamonds in the rough at the districts my team has been at, and we are kicking ourselves for not picking them.

I totally agree with you as our teams scouting captain and strategist the 3rd pick can make or break the alliance. And you have to pick the right defensive bot that works best for you because my idea for defense is a bot that can keep up with the lead offensive bot and to be able to push balls in occasionally thats my idea

[Zach O]

Our team has found no reliable way to scout defense. I know several team's that scout defense based on their scouters ranking them, but that leaves grey area in your data. You can't be sure that 2 people would rank the same robot the same "defensive ranking score". I think the best way to scout defense is just to have someone out there watching and taking notes. Do they consistently play defense, or just when they get stuck with it? And how well do they play defense when they do? The only way not to get mixed results from this is if a singular person, or a small group of 2-3, watches the matches and takes notes.

[ExTexan]

Others are saying there is nothing you can keep track of with numbers about defense

I often lean far to the side of relying on subjectivity but I have so often seen decisions made on complete sets of data that do not correlate with the results of the decision.

There may be something you can keep track of that would correlate to the effectiveness of a particular robot but it would be far too complex and there is not enough input for it to be significant, IMHO.

Assume the case of the great defensive robot, fast, durable, good traction and great driving team. They may bang around a single offensive robot with 3 balls in the zone, keep them from scoring, yet leave them and race to the other side to stop a midfield shooter who is lining up a shot. Meanwhile the offensive robot can't even push the balls in unprotected in the time they were gone. Yet if that single offensive robot was a good scorer, maybe they don't race to block the single midfield shot. How can that excellent coaching decision be ranked?

Too many variables in the game play and quality of robots in an average game. I agree that a person or small group should become qualified to make subjective decisions on defensive capabilities.

[brasspenny]

Team 997 Mentor here,

In NW Regional we won, picking first in qualifying and choosing the #26 and #55 seeded bots out of a field of 61.

Unheard of perhaps, but our scouting said these were the best bots who fit our "quantified" scouting system.

Scouting was fuzzy logic based. Logging bot attributes and then weighting them for each position in an Excel spreadsheet.

Data logged for a defender were: Clearing balls in autonomous, how many times blocking forwards who controlled balls, clearing balls in teleop, kicking two bumps, being able to go to mid field, knocking over opponents. I think there was another one but I don't have the scouting with me. (We also logged offensive data but was not weighted heavily for defensive bot)

These were the attributes we were looking for in a goalie and we were flabbergasted when we won based on the seeding of the picks.

I was very proud of our alliance captain for going with the scouting report versus picking higher seeds. It took guts to pass on the "big name" teams to go with the hard work that scouting had put in. It really pulled the team together.

I think this is 997's strength. We are a way way small market team (Motto since inception has been: Poverty in Motion). Can only afford one regional this year (and last) but we have won, from the first or second seeded position, 3 or the last 6 Regionals we have attended.

The reason is scouting and teamwork. We choose good partners and we play well with others. Also having a tank bot and driving well go a long way.

al

"Not to know but through search to find a way"
"If you are going to battle, best to have a tank on hand"

[The Lucas]

Originally Posted by Lil' Lavery

No offense to any teams, but most "defensive" teams are a dime a dozen. There will be a couple of standouts in terms of defense, but they'll be easy to pick out. The next group of teams will be almost identical in terms of what they're going to be able to give you on the field, and I highly suggest that you pick based on reliability and robustness (which is pretty hard to scout effectively).

Originally Posted by fuzzy1718

Oh and Lil Lavery, if you think defensive teams are a dime a dozen, then you obviously havn't been paying close enough attention. Picking the right defensive bot has made and broken many finals alliances. These teams were diamonds in the rough at the districts my team has been at, and we are kicking ourselves for not picking them.

I think you misunderstood Sean. He said there are a few elite defenders that stand out. The rest of the (average) defenders are similar in defensive skill and he suggests you use reliability to differentiate them.

The supply of this tier of non-elite elim-worthy defenders typically exceeds the demand (remaining slots) at most events. That is why the second round of drafts are very unpredictable. If you work on preparing a list for your team, you probably watch each draft pick and at least once think "Team XXXX is good, but I had Team YYYY a little higher". Chances are when it come to your second pick you will have 2 bots that you rated about the same to choose from. Occasionally there is that "I can't believe they are still available, steal of the draft pick " (see Bedford Express in Detroit or my favourite 56 in Philly '09) but usually it is just a another pick. A quality 3rd bot is important for any alliance, no one is disputing that. However, you don't need to hit a home run with that pick to win the regional.

[Lil' Lavery]

Originally Posted by fuzzy1718

Oh and Lil Lavery, if you think defensive teams are a dime a dozen, then you obviously havn't been paying close enough attention. Picking the right defensive bot has made and broken many finals alliances. These teams were diamonds in the rough at the districts my team has been at, and we are kicking ourselves for not picking them.

Look, I've been doing this for years. I attended four regionals in person this year, competing in two of them. I know what I'm talking about.

There are always a couple standout defensive teams, but if you can't pick them out, you've got some serious scouting issues. They're the teams who wreak havoc on even the best scorers, and it's pretty easy to figure out who they are.

Past them, it becomes a very hairy and the field has tons of parity. It doesn't take a lot of unique devices to fill the "typical" physical defensive role. Any team with good/experienced drives, a robust drivetrain, and some basic intuition of how to play defense can fill this role. Yes, some will be better than others, but differences between these will often be minute.

However, you need reasonable expectations and good instruction for these teams. There are styles of defense for fast and nimble bots and there are styles of defense for powerful and forceful bots. Don't try and make one play the other style. And don't expect any of these "pack members" to shut down elite scoring machines.

Most importantly, look at the strategy involved here. Many of the best alliances aren't just sticking their 3rd pick in the defensive zone and leaving them alone. The best alliances know what roles each of their bots can fill, and build the best alliance strategy around all three robots (even if it's not an individual team's preferred role).

[thefro526]

Defense is a fun subject. I've found that almost any team can play defense, but very few can do it well.

This year, we picked a defensive robot for our Alliance in Philadelphia based on how well they were able to play defense on us. When first hand knowledge of a defender isn't available (I.E. we didn't play against any) then we go look at who defended powerhouses in the most effective and least risky way possible. I say least risky because defense can go bad really fast, and a defender can rack up a lot of penalties and cards pretty quickly if they over do it.

[efoote868]

1. (From the stands) Figure out how well they can control their robot.
-A robot that looks well controlled and drives the way you think it should (qualitative), or how long it takes them to accomplish (simple) tasks (quantitative).

2. (From the pits) Figure out their gearing situation, their max theoretical speed.
-If they have a "slow" robot, they will not be good at defense.

3. (From the pits) Figure out what type of wheels they have.
-What style of defense do you need? Do you need them to pin other robots around and act like a "bully", or do you need them to play a different style?

[Dancin103]

Originally Posted by ICanCountTo19

The problem is that it depends on the team they are defending. Defending HoT is totally different than defending 397.

Defended this team #### to only scoring this many balls ## then compare that to the average number that that team scores? If it is significantly less then you know they are a good defender. If there was a robot problem on the offensive team it should be noted so that their already lower score for the match is considered.

I don't know that you would be able to quantify this to the degree they are looking for. Granted you can quantify it, whether or not you would get a good out put for a statistical analysis on defensive robots is another question.

You could look at a robots power output and the strength of their push and shove and what their speed can do in terms of how many blocks and blocked points they can output in a match. In terms of creating good defensive output in an analysis, I know you are only looking for quantitative, but in terms of qualitative you could do, can a robot block 2 robots or 1 robot. A yes would be 2 robots with a 1 value and a no would be a 0 value in your data. Back to talking in terms of qualitative you could look at whether the robot can score from the defensive zone, if so how many points a match.

Just a few ideas for you, as you can tell analysis and databases are my thing. My final project for a class is creating a monster scouting database. The professor is going to think I am nuts but that's fine by me, everyone calls me the crazy blonde anyway. I hope this helps, and was just my two cents.

[XaulZan11]

Originally Posted by fuzzy1718

Some of you are saying that kicking balls out is the best defense.

I wouldn't consider this defense. I do agree that having 3 teams just kick ball the close zone can be the preferred strategy, but I wouldn't consider it a 'defensive strategy'. Like, I wouldn't consider Otis Smith and the Giants offense that held the ball for over 40 minutes in Super Bowl 25 'defenders'. Yes, their brilliant ball control was sucessful in keeping the Bills no-huddle offense off the field for most of the game, but didn't defend anyone. If a team is looking for the 3rd robot to play defense, they are looking for robots to fulfill a different strategy: a defensive one, not a ball-controlled offensive approach.

Several people have hinted at or said that defense is only for their opponents' close zone. So, far this year we have won matches having our defenders starting and playing in each zone. If anything, I think we have had the most sucess when trying to keep an elite robot in their far zone.

[XaulZan11]

After reading some posts, it seems that some people are mistaking this thread for "how do you scout teams to play in the far zone" (see posts #10, 16, and 22) instead of "how do you scout teams that play defense".

[Chris is me]

Originally Posted by XaulZan11

After reading some posts, it seems that some people are mistaking this thread for "how do you scout teams to play in the far zone" (see posts #10, 16, and 22) instead of "how do you scout teams that play defense".

I think "defense" is poorly defined. All robots should be playing defense under my definition, which is anything that slows or denies a scoring opportunity for the other alliance.