Is Defense Back in a Big (Bad) Way?

[Mr. Lim]

Just wondering how many teams out there have had a chance to simulate real-match quality defense yet?

Team 610 has been running Open Practice sessions for Toronto area teams for the past two weeks, and we've had the pleasure of having several teams run both with and against us.

This past weekend was the first time we felt the robot was "ready" to go up against near-competition-level defense. We had our good friends at 4476 match up against us, as well as our alumni driving our 2013 World Championship robot "Taz" on defense.

What we learned is that defense can be really effective this year, even to the point that it could negatively impact the enjoyment of the game - both to participants and spectators alike. I like defense in games, but this year seems to be on a higher tier than most.

I'm not one to make predictions, but I can see a savvy defense robot posting a disturbingly low GAA (Goals Against Average - for you hockey fans), especially during early week quals. I can see a robot like this shutting out all but the absolute best, elite level robots at a competition. Think: having a competition with 40 teams, and say only 4 of them can actually score any boulders if quality defense is played against them.

We often measure a robot/team's effectiveness by their net effect on score. Stopping the opponent from scoring points is just as impactful as scoring points for your own alliance. Seriously disrupting opposing boulder points from being scored seems... well... pretty easy.

Lastly, breaking defenses alone is proving to be very hard on the robots. When adding tough defense robots into the mix, I think a lot of teams will have difficulty keeping their robots functioning for an entire competition. This isn't the same kind of bumper-to-bumper contact we've seen in past games. Even with really good teams/drivers/robots, we had robots driving into/onto/over each other. So many drivetrains are built to tackle the tall and bumpy defenses, inadvertently driving over an opposing robot's bumpers and into their robot was definitely something that happened too often.

I'm a bit worried that small events might struggle to field a full complement of working robots for the elimination rounds. Robots with fully working drivetrains might be rare on the backend of the serpentine draft.

I'm a bit worried that boulder scores might be awkwardly low in any match where there's a defense robot - unless one of the elite teams is on the field.

Is anyone else as worried about defense as me? Or am I the only one?

[Jay O'Donnell]

You're not alone.

I think most events, up to and including the championship event, will be decided by being able to get low goal scored past the defense and keeping your robot from breaking.

I can say without question many events will have some limping robots by the time finals comes around.

[PayneTrain]

Shawn, you might have been simulating competition-quality defense (which no team ever does enough of... our brutish 2013 robot is going to be getting some extra miles these next few weeks from yours truly, I assure you) but can you say you were simulating any and all likely competition-quality courtyard scenarios?

[EricH]

You know, we all complained about the lack of defense last year.

So the GDC gave us defense AND DEFENSES. Two years worth of bumper rookies in one, two years worth of defense in one. I think this is going to be Breakaway levels of robot defense, sandwiched in between the Stronghold Defenses.


This one's gonna be interesting and crazy.

[IronicDeadBird]

I desperately hope teams have the mind to bring in a lot of replacement parts. Last year defense was complained about in the game that was pretty much player vs environment, now we have a year where its player vs player vs environment.

[mrnoble]

An overbuilt robot is a must this season. I agree strongly that teams (or even groups of students; this sport has a very short rotation of veterans after all) who only know last year will be punch drunk when they experience this year's game. I am very concerned about broken bots on the field and in the pits.

[jeremylee]

Since only one defender is allowed to guard thier castle, I think teamwork can help overcome this problem. But I agree with the comments about a rough game, the pits are going to be crazy busy this year.

[IronicDeadBird]

Since only one defender is allowed to guard thier castle, I think teamwork can help overcome this problem. But I agree with the comments about a rough game, the pits are going to be crazy busy this year.

Doesn't mean you can only defend from the castle though

[Edxu]

Shawn, you might have been simulating competition-quality defense (which no team ever does enough of... our brutish 2013 robot is going to be getting some extra miles these next few weeks from yours truly, I assure you) but can you say you were simulating any and all likely competition-quality courtyard scenarios?

We went through several scenarios that we worked out with other teams at our practice, including individual cycling and feeding cycles among other things.

The defensive robot mostly played zone denial upon entry of the attacker from neutral to courtyard, then tried to disrupt them from reaching their shooting position with a mixture of head-on pinning and t-boning. After the opponent makes their shot, the defensive robot then player denial again, disrupting the attacker from leaving the courtyard.

From standing behind the driver station on both sides (attacking and defending), I observed that the Middle Driver Station and Right Side Driver Station have absolutely no vision through the glass of the Secret Passage or the left-side goal when trying to play defense. Conversely from the Left Driver Station, the driver has little to no vision of the right-side goal and the spy area.

I think that strategy will make or break a team's offensive strategy this year, as without cameras, the defending robot is at a huge disadvantage, and if the defensive robot is known before the match begins, a strategy can be made to exploit it.

[PayneTrain]

I don't think you gave me the specific answer I was looking for, but that's ok.

I would definitely agree that use of cameras in this game is potentially part of the foundation for a good strategy. Teams should take any opportunities within the bounds of the rules and Q&A rulings to maximize vision for their drivers.

[waialua359]

An effective scoring robot capable of handling defense, is a robot that never has to fully go into the opponents courtyard (subject to defense) when scoring.

Shawn,
great thread.
This is 2010 all over again in a very big way.

[mjc49]

An effective scoring robot capable of handling defense, is a robot that never has to fully go into the opponents courtyard (subject to defense) when scoring.

Shawn,
great thread.
This is 2010 all over again in a very big way.

This is our teams take away from this past weekend as well. I'm really looking forward to seeing alliance strategies, both defensive and offensive.

[Sperkowsky]

we played some defense and got some defense played on us at the Suffield Shakedown. Defense is something that is really needed this year. You can destroy a high goal shooters chances with good defense.

Also warning to teams with those giant 12"+ wheels. Smaller robots will flip you. It may be an accident maybe not and it is a foul but beware.

[JamesCH95]

I am interested to see how the lack of 'tall' robots impacts defensive play this year. It strikes me that a box with a stick or a box with a wall will be a very effective non-chassis-contacting defensive robot.

It also seems likely that teams who can shoot from the nominal protection of the outer works will have significant protection from said short robots.

An effective scoring robot capable of handling defense, is a robot that never has to fully go into the opponents courtyard (subject to defense) when scoring.

Shawn,
great thread.
This is 2010 all over again in a very big way.

A lot of care must be taken in this strategy - if a robot releases control of a boulder and does not fully enter their opponents courtyard they will draw a foul. However, darting into the courtyard after the boulder has been shot should be quite easy.

[AdamHeard]

I hope it is.

RIP Low shooters w/o a plan for this.

[Rangel(kf7fdb)]

My prediction is that defense will not be as effective against elite teams who designed to play the game to be very hard to defend against but extremely effective against everyone else. Shooting from the outerworks and over any defenders is a great way to negate defense. This shooting ability will be a tiny minority though. Everyone else is going to face Aerial Assault but even worse since the goals are narrower.

[barn34]

I definitely agree on the potential for defense to really play a big factor this year. This is the first game I can remember where if you don't have a certain minimum score, you can't access a lot of the point potential an alliance can obtain from capture/climb and that's not really figuring in how many points you stop from keeping them under 8 goals scored. If you force them to go low goal, that's also 2.5 times less efficient than 1 high goal scored so they need to score 3 times to your one to come out ahead on point potential.

These are just some of the reasons we've intentionally kept our design flexible to cater to the defensive evolution of the game this year. This could prove to be FCS blockers from 2013 and goalie poles from 2014 on steroids, or some tandem offensive strategies could evolve that consistently prove to knive through defense attempts like they were toilet paper. It'll be interesting to watch the evolution, that's for sure.

[mathking]

Here is my take, based on an admittedly small sample set of one practice competition. But also on a bunch of simulated games using remotely controlled Lego Mindstorm robots. Defense in the courtyard is very effective one on one. Not nearly as effective when the offensive robots cooperate. If a robot can either score from the protection of still being in the defense or it can score from the batter then good cooperative tactics can cut down on the effectiveness of defense in the courtyard. I think that during qualification matches, particularly in the first few weeks, defenders will have more of an advantage. But once teams figure out how to block and pick and roll it will be tougher.

I think simply blocking one of the defenses may be the most effective way to slow opponents down. Particularly since I see a lot of alliances with 1-3 defenses they can't cross in qualifications. In our simulated games slowing opponents as they come out of your courtyard and again as they try to cross back in was just about as effective as defense inside the courtyard. (Caveat: We were playing with a bunch of bots that had a "race to the batter and score" offensive strategy, so once they got around a courtyard defender they were hard to stop.) This will be one of the games where it will be hard for the great robots to dominate on their own. I think effective scouting is going to be absolutely imperative. You need to select the defenses that will most inconvenience your opponents and then use a strategy that takes advantage of your selections.

[ASmith1675]

I'm having a tough time reading how this game is going to play out defensively. In 2014 it was clear defense was going to be a critical part of the game as only 1 game element per alliance ostensibly forced 2/3rds of all the robots on the field into playing defense (since there wasn't much else to do).

This year a highly effective defender might be able to partially shut down scoring in the tower, since it is a relatively small area to defend. At the same time, I think an offensive blitz strategy can and will easily overwhelm the forced single defender. In my opinion the most effective defenders will be the ones that can force turnovers and deposit boulders in the passage. This will make the opposing alliance "waste" cycles going back through the defenses to grab more boulders.

[Justin Montois]

I think one thing that's missing in this thread is the fact that the opposing alliance will (I think) be focusing more on getting their RP's than on playing defense.

From the Week 0 events, putting 8 balls into the tower was rare from what I saw. Without those 8 Balls, that's a Ranking Point left on the table. Devastating to teams trying to seed high. Further, taking down the defenses quickly enough to leave enough time to score 8 balls must be a team effort. Failure to breach your opponents outer works would again leave a Ranking Point on the table and again devastating to teams trying to seed high.

In most cases, if you play well enough to get your two RP's in the match, the additional RP's from winning the match will most likely follow ensuring a high seed.

I think the nature of both alliances trying to get those all important ranking points will limit defense and create a fun, high scoring game to watch. I hope.

[Joe Johnson]

Good discussion of defense, but I think that this is going to be a penalty nightmare for many defensive robots. When can they touch a robot, when can't they?

The early weeks of the season are going to see so many penalties due to teams playing aggressive defense around the defenses.

It will sort itself out by weeks 3 or 4 but not so much in those early weeks.

Dr. Joe J.

[pfreivald]

I think one thing that's missing in this thread is the fact that the opposing alliance will (I think) be focusing more on getting their RP's than on playing defense.

I agree. An average of four cycles for each of two robots (damage defense + scoring on the tower) looks to be significantly harder than an average of three cycles each for three robots.

That said, on any alliance where you know you're not going to breach and/or capture, that defense can really obliterate a game plan.

[Josh Goodman]

That said, on any alliance where you know you're not going to breach and/or capture, that defense can really obliterate a game plan.

Or, let's say, against an alliance that you know is capable of doing both.

[pfreivald]

Or, let's say, against an alliance that you know is capable of doing both.

In elims, sure. I'm not sure that in quals denying their qualifying points is priority over maximizing your own game points.

[Rangel(kf7fdb)]

In elims, sure. I'm not sure that in quals denying their qualifying points is priority over maximizing your own game points.

Depends how the alliance thinks. Is risking 1 breach point a better option than risking 2 win points. I would imagine there would be no defense only if both alliances believed they could score more than the other by focusing on offense. If an alliance feels weaker on scoring, playing defense to secure the 2 win points may be more worth it than focusing on breach and capture points.

[Josh Goodman]

In elims, sure. I'm not sure that in quals denying their qualifying points is priority over maximizing your own game points.

I think that depends on the situation. I'm more talking about trying to maximize your RP by snagging the W. If they're capable of a breach and a capture, there's a very good chance (in the quals) that solidifies a win. If you can shut down a fantastic scorer with a mediocre scorer playing defense enough to prevent your opponents from capturing, you have a much better chance of winning that match.

(Sorry, thought process was stupid convoluted and deserved more than a one sentence response.)

[pandamonium]

I think one thing that's missing in this thread is the fact that the opposing alliance will (I think) be focusing more on getting their RP's than on playing defense.


I totally agree with this. It may be able to contain a shooter and lower point output but will it be enough to change the outcome of a match? The other thing that is missing here is the defense free end period and autonomous mode. A top level robot that can shoot 2 balls in auto and a few during the end of the match.

[AdamHeard]

I totally agree with this. It may be able to contain a shooter and lower point output but will it be enough to change the outcome of a match? The other thing that is missing here is the defense free end period and autonomous mode. A top level robot that can shoot 2 balls in auto and a few during the end of the match.

No ranking points in elims.

Also a win is worth the most in quals (as I doubt many events will see consistant breach and captures early in season).

[PayneTrain]

No ranking points in elims.

Also a win is worth the most in quals (as I doubt many events will see consistent breach and captures early in season).

I don't think the 20+25 bonuses in elims are necessarily small potatoes, for what it's worth. Maybe I'm wrong.

I hope teams considered how often are they would get 2 rank points in a match they lose in their strategy.

[tindleroot]

I've been noticing similar trends about defense (and maybe Karthik has too?).

As far as I've seen low, ranged shooting robots are going to be extremely common since those designs can easily cross the low bar and shoot in the high goal. However, there is an inherent flaw in that design in that given how long it takes vision (or manual alignment) to line up a shot, those robots can easily be blocked by a tall robot or by a pool noodle blocker. And, if a robot has a decent drivetrain but a poor scoring mechanism, they can play this role in a match and be extremely effective.

Winning in quals is more important than the other ranking points since winning is worth 2 RP, plus you deny the opponent 2 ranking points, which helps you advance farther than the competition. I can guarantee that any smart alliance will do their best to shut down a high-potential scoring robot if it is easy to stop them, as is the case with the low-shooting robots. This will happen in both quals and elims, and it will definitely happen if the team in question is high-ranked.

[Brian C]

While your opponents defensive strategy is always something to be concerned with when planning your offensive moves and this years game does indeed provide for some interesting scenarios.

I believe this year will see a very steep learning curve in Robot Robustness for those teams that have never competed in a game like this. There are many teams out there that have never had to deal with conditions they will see this year. To me this harkens back a decade ago when bumpers were first optional and even further back when there was indeed a fair amount of robot to robot contact.

I expect to see lots of loose parts strewn about on the field and at least a battery or 2 or 3 :eek:

I have a hunch the broken parts box we keep on the scorers table at the competitions this year may need to be a tad bigger than in recent years.

[Keefe2471]

I've been noticing similar trends about defense (and maybe Karthik has too?).

As far as I've seen low, ranged shooting robots are going to be extremely common since those designs can easily cross the low bar and shoot in the high goal. However, there is an inherent flaw in that design in that given how long it takes vision (or manual alignment) to line up a shot, those robots can easily be blocked by a tall robot or by a pool noodle blocker. And, if a robot has a decent drivetrain but a poor scoring mechanism, they can play this role in a match and be extremely effective.

Winning in quals is more important than the other ranking points since winning is worth 2 RP, plus you deny the opponent 2 ranking points, which helps you advance farther than the competition. I can guarantee that any smart alliance will do their best to shut down a high-potential scoring robot if it is easy to stop them, as is the case with the low-shooting robots. This will happen in both quals and elims, and it will definitely happen if the team in question is high-ranked.

I agree with this. I would add that I don't think defense is going to be as difficult to play as people think. The Refs have (at any competition i was at) emphasized exactly how pinning will be called and what to look for to avoid committing the foul. The only other penalty that defenders will have to look out for is G43;

ROBOTS on the same half of the FIELD as their ALLIANCE TOWER may not interfere withopponent ROBOTS attempting to traverse OUTER WORKS (regardless of direction). A ROBOT isconsidered traversing the opponent’s OUTER WORKS if any part of its BUMPERS are within theopponent’s OUTER WORKS and no part of it is in their opponent’s SECRET PASSAGE.

If defense is played between the outer works and the tower (which will be an effective spot to block low shooters and interfere with short range shooters attempting to reach the batter), it will be played 10-15 feet in front of the alliance station. https://www.youtube.com/watch?v=acDqfZZLWV8 POV from a human player at a real field week zero if you want to see how easy it will be for drivers to see their robots. They should be able to see all the edges of the defenses marking the "safe zone". The vision advantage and the need to protect the secret passage from poaching should make a defense robot very powerful (if not required) in eliminations at most of the regionals and districts.

[Ginger Power]

I expect to see lots of loose parts strewn about on the field and at least a battery or 2 or 3 :eek:

Yep (http://www.twitch.tv/knightkrawler2052/v/48269560?t=01h18m55s)

[GeeTwo]

I hope teams considered how often are they would get 2 rank points in a match they lose in their strategy.

We definitely targeted our strategy at the non-zero-sum breach and capture rank points at least as heavily as the WLT points. We may well be in the minority.

One of our top design goals was a solo BREACH in case we show up to a match with partners who have negligible scoring capability. This was actually the most important reason we decided to do the low bar; it's easier and faster to cross solo than the category C doors.

[CTbiker105]

I am very concerned about broken bots on the field and in the pits.

Here, here. A lot of teams at week 0 events could easily manage getting over defense after defense, but that was largely without any opposing robots playing defense. I saw only a handful of bots that seemed robust enough to withstand the damage they'll most likely receive during an actual competition.

[MooreteP]

It's the Playoffs now, not Eliminations.
A kinder, gentler FIRST.:)

That being said, defense. The opposing Drive team will have the advantage with a very clear view of their actions.

Breaching the Outer Works will guarantee an RP.
Must do this, and can't be prevented from this task by defensive action.
Three Robots crossing a defense in Auto would be 30 points.
Crossing the rest of them in tele-op could be worth another 30 for 60 total.
You can't ignore these points that a a defense cannot stop. No boulder holding required.

Best score from scoring in the high goal on the tower is 60 in autonomous and 25 in tele-op for 85 points. (assuming no two boulder autonomous, just enough to weaken the tower, both of these may change, I know....) But with an effective defense, this becomes 10 in the low goal, which equals 70.

If after that, you can capture the tower, which can't be stopped by defense the score increases by an additional 15 for 3 Robots on the batter, up to 45 if each of them can scale.

Defense IS back in a big GOOD way.

A good defense can prevent a tower capture, which could lead to a scoring advantage of 25 in the playoffs.

The math is making my head spin.
I am loving this game.
God Bless the referees.

[BrennanB]

I'm a bit worried that boulder scores might be awkwardly low in any match where there's a defense robot - unless one of the elite teams is on the field.


Basically this. But that's been pretty much what I thought from day 1. Any mediocre defense can shut down/significantly slow down the average team. It's even worse than 2010 though because of the close proximity of scoring objectives. If you are a close shooter/lowgoaler you have to outplay the defender significantly to have a chance of getting around them. There are however protected zones, and some teams will have had the foresight to tune/design/play the game on that idea.

There are still some clever things that teams can do to overcome this, and I don't think it's a fault of the game really, just adds some strategical depth to the matches. People will probably catch on to the easiest ones end of week 1/mid week 2.

What I am most curious about is the robot driving up robot action. Likely flips and such. What about these penalties? Robot frame rule?

There is plenty of defensive meta that people aren't thinking about, and as that slowly comes out I think people will have to adapt even further from the basic courtyard defense. Very excited to see what I think is a very even playing field for alliances, much like 2014 was. I suspect we shall see some very successful low seeded alliances taking out the big guys.

[waialua359]

I've been noticing similar trends about defense (and maybe Karthik has too?).

As far as I've seen low, ranged shooting robots are going to be extremely common since those designs can easily cross the low bar and shoot in the high goal. However, there is an inherent flaw in that design in that given how long it takes vision (or manual alignment) to line up a shot, those robots can easily be blocked by a tall robot or by a pool noodle blocker. And, if a robot has a decent drivetrain but a poor scoring mechanism, they can play this role in a match and be extremely effective.

Winning in quals is more important than the other ranking points since winning is worth 2 RP, plus you deny the opponent 2 ranking points, which helps you advance farther than the competition. I can guarantee that any smart alliance will do their best to shut down a high-potential scoring robot if it is easy to stop them, as is the case with the low-shooting robots. This will happen in both quals and elims, and it will definitely happen if the team in question is high-ranked.
You hit the nail on the head. I can see a team with a great linear shooter who is low, get shut down because they could not forsee this. This especially for teams that did not experience previous games such as 2010, 2012, and 2014.
However, more experienced teams that chose to go low understood this, and made sure a max 54" robot or pool noodle blocker was ineffective in blocking their "arching" shot.:)
148's reveal is an example of such a robot.

[Michael Rossi]

From the Pittsburgh practice field. The strategy was very much based around breaking all defenses then shooting. After all the defenses where broken what would happen was the 2 best shooters went to the opponents courtyard. While the third robot shuttled boulders underneath the lowbar while playing some light defense. So while defense is back if you have at least 2 robots in your opponents courtyard it should minimilize defensive effectiveness.

[Citrus Dad]

A couple of points of how defense differs this year from the 2012-14 period. The first is that it is much more difficult to flow from offensive to defensive ends of the field. A robot is less likely to spend time defending as part of their scoring cycle. With a good HP feed, they may not need to go beyond the neutral zone, and may be able to poach balls in the front court.

The second is the limit of one defender. A 3 on 1 attack will minimize defensive effectiveness. Yes the defender might focus on the best scoring robot, but the truth is that the better scorer will be able to outdrive the robot left behind to defend. (The only exception might be in the Champs Subdivision finals and on Einstein. But then all 3 offensive bots will be highly effective.)

We'll see how the defenses play out. This will be a much different flavor than recent games though.

[Conor Ryan]

Speaking of Penalties, with all the field elements, will the Referees have Line of Sight issues?

[IronicDeadBird]

A couple of points of how defense differs this year from the 2012-14 period. The first is that it is much more difficult to flow from offensive to defensive ends of the field. A robot is less likely to spend time defending as part of their scoring cycle. With a good HP feed, they may not need to go beyond the neutral zone, and may be able to poach balls in the front court.

The second is the limit of one defender. A 3 on 1 attack will minimize defensive effectiveness. Yes the defender might focus on the best scoring robot, but the truth is that the better scorer will be able to outdrive the robot left behind to defend. (The only exception might be in the Champs Subdivision finals and on Einstein. But then all 3 offensive bots will be highly effective.)

We'll see how the defenses play out. This will be a much different flavor than recent games though.

The 1 defender limit exists only in the courtyard.