Game Design and Forced Defense in Qualifications

The discussion in this thread got me thinking about a perennial problem (arguably) with FRC games and the FRC experience.

Let’s say you’re attending your only event of the season, and you’re excited to compete with the robot you and your team worked hard on. You’re confident that your robot can score a lot of game pieces, although it’s definitely a work-in-progress (FRC is hard, and I think most of us have come to an event with a robot that isn’t quite functioning yet at some point).

In your first few qualification matches,

  • your robot experiences some issues and is unable to score
  • You keep working on your robot’s scoring mechanisms but you keep experiencing weird issues
  • In your next strategy session, your partners ask you to play defense. You know that if your robot is playing up to it’s potential, you have a better shot of winning the match, and you really think that’s your best shot of becoming an alliance captain/getting picked/winning the event, etc, but you play defense in that match to be a good alliance partner.
  • Every match after that, you’ll have partners asking you if you can play defense.

It can be hard to start playing offense again once you’ve been designated a “defense robot”.

In games where 2 Offense 1 Defense is the primary strategy, you usually need someone in each match to play defense. It isn’t bad to be a defense robot, on the contrary, it’s an important and often fun to play role, but it can feel demoralizing to have spent time on so many mechanisms to use in teleop and not get to use them.

I propose that this is an issue with FRC Game Design, rather than something essential to the FRC experience. Games should be designed such that Triple Offense is equally valid or slightly better strategy than 2O/1D, and teams going for the lowest possible objective (low goal or equivalent) should be able to contribute in that respect.

Thoughts on specific games (links it somewhat to the game tier list discourse):

Games that are designed well to avoid this issue/poorly

Games that do this well:

  • FIRST Stronghold (2016)
  • FIRST Steamworks (2017)
  • Aerial Assist (2014)
  • Ultimate Ascent (2013)
  • Infinite Recharge (2020)

Games that do this okay :

  • FIRST Power Up (2018)

Games that do this poorly:

  • Destination: Deep Space (2019)
  • Rapid React (2022)
  • Rebound Rumble (2012)

Egregious Offender:

  • Recycle Rush (2015)

Tell me what you think? Is this a real issue, or just a perceived issue? Is it an issue with game design or with the way we approach game strategy or something else?

13 Likes

I feel like most competive games as a whole would suffer if 3 scoring is the meta. You get a much less interesting game to play where the outcome of every match is more or less preordained by the time robots are consistent. I can see your perspective but in your own post you outline the scenario where your robot was incapable of offense for a large portion of the event, so the chances of pivoting successfully to offense isn’t really related to the fact that you played defense but the fact that your robot wasn’t fully functional.

  • For the excitement aspect, watching 3v3 offense just isn’t very interesting, I wouldn’t want to watch a basketball game where the teams just took turns scoring, or play a game like recycle rush (which is essentially what’s being described). Defense allows for more dynamic sports like matches, where teams who did not build as well still have a way to make a meaningful and exciting contribution.
7 Likes

In terms of game design, it comes down to a question of bottle necks - which can be separated into two portions, game piece availability and scoring location availability

With sufficient game pieces, and sufficient locations to obtain them from, you don’t have to worry about being starved for game pieces - there’s plenty to go around. In theory, 11 game pieces split between 3 robots (which can each only hold 2 at a time) may seem like enough - but then you have to take into account how long it takes for the game pieces to be fed back into the field, along with the cycle time for teams to score - it turns out it’s not enough!

With sufficient locations to score from, you don’t have to worry about waiting for people to move out of your way. This largely comes into play with pick-and-place games, as they have limited locations to score from. Shooting games, while there may be one or two preferred spots to shot from, can generally be scored from a large number of areas.

And for the oddball when it comes to bottlenecks: Field layout. It’s one that’s hard to quantify, because it’s impacted largely by other aspects of the game, but it can play a huge role.

Personally, I think games like 2014 really highlight how both offense and defense can be integrated into a game, without making anyone feel relegated to one or the other.

3 Likes

IMO, well designed games have relatively open methods of scoring and multiple methods of “playing defense”. 2011 was a great example of this in my opinion. Tubes had a secure loading zone and could be thrown on the field as a time saver but higher risk maneuver. Scoring also had a bit of a protected zone and defense.

Defense could be via the typical bump and grind… block and push… but you could also steal undefended pieces thrown into the open field. And if so inclined, you could be extra sneaky and steal 1 particular type of game piece which in turn limited the “logos” that could be made which created a difficult deficit.

I liked that style of defense, as instead of just bump and grind, you were using a collector in order to grab and manipulate game pieces. When that is allowed, I think it can add an extra dimension to the game play.

2022 had potential for this. Some teams would grab and hold a couple of the opponents cargo this year, but it was not as effective as 2011 could be.

This needs to be balanced though because game piece starvation (complete starvation) is a boring game to watch and play. It would be an interesting exercise to play this years game with 1 or 2 less game pieces as there is likely a breakpoint where slightly fewer pieces would not completely starve the game, but would dramatically change its dynamics.

2 Likes

One way the GDC can mitigate unwanted defensive play is through the judicious use of clearly marked and universally visible protected zones.

5 Likes

Couple of points:

  1. Having a defensive role as part of a competitive game allows for a wider variety of skills to be valuable. As has been discussed elsewhere, some teams struggle to field a robot that can do much more than drive. It seems to me that having the game design allow for defensive tasks in addition to offensive tasks provides a way that these teams can participate in a match. Without the ability to play defense, you would be telling those 3 or 4 teams at each tournament that were not able to complete much more than a drivebase that they should just sit on the sidelines. I think that would be unfortunate and un-inspiring for those teams.

  2. defensive tasks are not just good for those teams that struggled through build season. It is also a way for higher tier teams to salvage a match after a breakdown. In our first match at Worlds this year (Newton, Q8), our vision system stopped working after a hard hit. This effectively disabled our ability to score cargo. Our drivers recognized this failure and started playing defense against the opposing alliances scorers. This allowed our alliance to win the match. We also experienced matches this year where a belt came off the intake or some other breakdown occurred that we were able to salvage by switching to a defensive role.

  3. Defense is a highly valuable part of many other competitive sports. Without defense, the game becomes more of a skills competition and there is very little reason to have both alliances on the field at the same time. While skills competitions can be exciting (think of individual sports at the Olympics or the NBA’s slam dunk contest), it is a very different style of competition than team sports.

  4. Defense creates additional design considerations that provide more of a challenge to all teams. Designing a robot that can maneuver well and avoid defensive strategies is more challenging than designing a robot that simply performs a skill when un-defended.

1 Like

I hope teams know that they can push back on that in qualifications. In our experience some conversation about who needs what usually solves the problem. If we’re asked to play defense and we say “but we really want to try out our shooter this time” it’s almost always been accepted by the alliance. Once or twice another team countered with, “but we really need the win RP” and we’ll help out with that instead if we can. But I don’t think anyone has ever tried to force us to play defense in qualifications. Everyone has always been cordial to us in qualifications. I know it would be largely anecdotal, but is this really that much of an issue in qualifications matches?

Like the OP in the referenced thread, we’ve occasionally come across some teams that acted like jerks in eliminations and “forced” us to play defense. But that’s their “right” if you want to call it that, and we have learned to brush it off. They could have picked someone else. We knew when being picked it would likely be for defense, but it’s always a better experience to be included in decisions.

5 Likes

I definitely don’t like games with no defense. But I think games where defense is an option, rather than feeling mandatory, are a better experience for low-to-mid level teams.

4 Likes

I think defense should be considered a key aspect of an FRC game. It adds a level of strategic depth that is difficult to replicate. I do also recognize the reality of teams who want to show off something besides their drivetrain and their base driver’s skill feeling pressured by alliance partners to “just” play defense. One way that the design of a game can help mitigate this is by creating opportunities for robots to perform offensive tasks in a defense way.

This year, acquiring cargo was a task that could serve an offensive purpose (allowing the robot to score their alliance’s cargo in the hub) as well as a a defensive purpose (denying the other alliance access to their cargo). If the game designers wanted to take this a step further, they could have added a second “dummy hub” that didn’t score any points but still kept cargo out of play for a period of time. That would have allowed a team playing defense to still leverage their capability to score cargo, just in a defensive way.

The best example of this overlap that comes to my mind is 2018. The primary task that most robots were designed for was scoring a cube on the switch and/or scale. That task was inherently both offensive and defensive, since it increased your alliance’s scoring potential while also decreasing the other alliance’s scoring potential.

This overlap seems to be strongest in pick-and-place games where both alliances are scoring on a common field element. The other two games that stand out to me from this perspective are 2007 and 2005. In 2007, a single well-paced tube could disrupt the other alliance’s ability to create a high-scoring string of tubes. Similarly, in 2005 a single tetra scored in a defensive position could negate multiple rows that had been built by the other alliance.

3 Likes

You’re both basically correct, but in different ways. I do think that sometimes high-level teams get into a offense-only/defense-only mindset that isn’t as helpful as it could be. As @wgorgen points out, this is something that really good teams (like his) don’t do. Instead they work with more flexible concepts of defense that allow for teams to perform both defensive and offensive operations as the situation warrants. There were a fair number of times this season where our team persuaded our alliance (in both qualifications and elims) to play what we termed “zone defense”. Since we were basically zoning the field anyway to keep scorers from interfering with each other, we used that to facilitate a shifting defense strategy on good opposing scorers, so that no matter where they went they were defended but all of our alliance robots could also score whenever the opportunity presented itself. He can attest that we did this to his team in more than one match at FNC DCMP, often to good effect (though, of course, they still beat us in the finals. :grin:) There are a lot of ways that even low-scoring teams can add in to a more flexible defensive strategy, as long as their allied teams are willing to think creatively about it.

4 Likes

Both 2016 and 2020 games allowed for interesting starvation strategies. Allowing for poaching also can be an interesting strategy that a third robot can participate in.

2 Likes

I’ve always wondered how a game like 2011 would’ve been if each robot on alliance was only allowed to score one shape of tube. You’d probably want to change the number of game pieces, scoring design, maybe even field layout somewhat. But I think it could be quite interesting. Most robots probably still wouldn’t have a “favorite” shape, so you could still decide match by match. (Or maybe something else is symmetrical, like one loading zone is under a low bar and another over a bump.)

Defense would be very viable, but having 3 scorers would be specifically rewarded. Or maybe don’t do it by shape but by goal, e.g. only one robot per peg level (add more pegs). Even crazier, maybe you only do these restrictions in elims. I’d want to change the scoring to not screw over random alliances in quals: more like typical 3-bot endgames and less like 2017 climb roulette.

they could have added a second “dummy hub” that didn’t score any points but still kept cargo out of play for a period of time.

This hub already existed in the Terminal. A single rule change that allowed human players to possess a certain number of opposing cargo could have allowed for this strategy.

3 Likes

My point was to provide a way for a team to use the full capability of their system for scoring Cargo in the Hub, which they likely spent significant resources on and want to be able to show off, in a defensive capacity. The Terminal could certainly be turned into a “Dummy Hub” but it would need some additional structure to fully mimic the Hub.

It absolutely has to do with game design, Kevin, although I also think there are key circumstances that make sending a robot to play D the best option for the alliance that are harder to solve through game design. These are circumstances such as:

  1. The field size being 27’ x 54’ - Six full size robots on this size field, equally spread out, is not a problem. When all 3 robots are playing offense and trying to exist in the same areas as each other, though, it becomes easy to choke out your own alliance members. Even on the most open of fields (I think to 2013 as a great example), you still had choke points and robots needing to maneuver around their alliance partners if everyone wasn’t perfectly in sync.

  2. Teams aren’t good drivers - Fits in with reason #1, when a team is bad at driving, they bounce around the field, knocking into field elements, game pieces, and most often, their own alliance partners. Sometimes the best thing for the alliance to do is send that team somewhere else, away from where the more capable scoring robots are. Short of sitting the robot in the corner, the best option is to have this team play defense and crowd the opponents’ zones.

  3. Game piece scarcity - A little bit of this is game design, but it exists to some degree in every game. You could pump every game full of a bunch of game pieces, but it eliminates a lot of strategic value. I’ve played matches in plenty of games where there are technically enough game pieces to go around, but some game pieces are easier / better to acquire than others, and the alliance is better off giving those game pieces to the best scoring robot(s) on the alliance, leaving the less-skilled robot(s) to struggle getting any game piece (especially when the skilled robots are so fast at getting game pieces that they just swoop in front of the less-skilled robots and steal the game pieces they’re going for). At some point the most productive option is for that team to play defense.

  4. Reliability - Often times teams show up in matches and they’re unreliable in what they can do. Happens all the time. In match strategy, it’s not uncommon to prefer reliable performance, and when it comes to going for something reliable, that often means prohibiting teams from trying out new things or playing a more risky strategy. This could mean play defense.

4 Likes

Strong disagree that Aerial Assist should be considered good here. I actually think this was one of the worst games to be a poor offensive robot due to only one game piece. A few issues:

  • Being asked to forgo auto to give an alliance partner a multi ball auto
  • Being asked to not touch the ball for risk of the ball getting stuck in a partner

I had a qual partner at champs that year absolutely refuse to let our other partner touch the ball even once. They wanted them to just park in front of a low goal the entire match. This is the only time in my FRC career I can remember DNPing a team because their drive team was so toxic.

I AM NOT REALLY ADVOCATING FOR THIS TO HAPPEN: ITS A THOUGHT EXPERIMENT!

This was just something that occurred to me as I thought about all the Defense questions coming out this year.

What if “Offense” and “Defense” became like Red and Blue. For example in 3 of your 10-12 qualification matches you were the “Designated Defender” and therefore it became part of the design Challenge? Rotate defense through the field.

4 Likes

The autonomous ball issue is definitely real with 2014, but the fact that getting 3 assist cycles was necessary to win competitive matches meant figuring out how to get your alliance partners involved was a priority in most strategy sessions I was involved in.

1 Like

2019 comes to mind when thinking of “forced” defense. There were many examples of capable alliances in Playoffs or Qualifiers with triple offense as clean strategy.

However, the mechanics of the game and field lent itself more to always having a defender and a sense that someone more than likely had to do it.

  • Game pieces came from one area of the field: at or around the Human Player Stations. To keep the robots cycling to the HP efficient it was best to keep other teams away from the Cargo in this area so they didn’t mess with the flow.
  • The field was very crowded with how large the field elements were. See the Cargo Ship. It made more sense to remove your third offense robot to clear up your side of the field and send it to be a nuisance in their traffic flow.
  • It was possible to max out your scoring elements. This a rare, rare accomplishment, but at higher levels of play it would come up that 1 team could do the Rocket and some of the cargo ship on their side so why need a 3rd robot to support if the 2nd team was doing something similar on the other side.

All of these pointed the ideal strategy to one defender as the optimum strategy in a majority of scenarios.

I don’t think these were necessarily intended by the GDC compared to more obvious game tactics like Kevin mentions in 2014 with a three assist cycle requiring all robots to participate in a meaningful way.

1 Like

I have a lot of thoughts on the topic of defense in FRC that I will likely post in one of these threads, but I haven’t organized them well enough yet and I’m not sure which of the half-dozen related threads are the most appropriate yet.

But I do wanna address the list in the OP quickly first:

This list confuses me a lot. Some of these takes are wildly different than what I would say.

2018- This is the best game for triple-offense and low-goal value in qualifications in FRC history, bar none. It’s not even close. There were effectively three different low goals (near switch, far switch, vault) that teams could score in, one of which didn’t even require lifting cubes off the ground (something that many recent low goals have required). Every alliance wanted to control at least one of those low goals (their near switch), and typically also wanted at least three cubes in the vault. With the primary scorer’s often engaged in the battle for the scale, it left plenty of space and value on having an alliance member work the near switch and/or the vault. There were also large protected zones around the scale and the cube pyramid, both of which discouraged bumper-to-bumper defense in two key scoring areas. And given the scoring method, the best form of denying points (aka playing defense) was scoring cubes on the scale or far switch. In the 48 official qualification matches that 1712 played that season, I think our alliances decided on having a robot play physical defense exclusively in only 1 or 2 of them.

2016 - This game was a mixed bag, and whether or not it made sense to play defense in qualification really depended on where your alliance stood compared to the ranking point break points. If it was going to require all three of your robots to complete the breach and/or weaken the tower, you got to have three offense-oriented robots. If two (or even one) robot on your alliance could complete those tasks alone - then it often made sense to send a defender. The thing this game did really well is allowed points to be scored by just a drivetrain. That is something that FRC used to have fairly commonly, but basically doesn’t exist anymore outside of autonomous taxi points. It’s something that should be in more FRC games, even if its just the ability to push objects into a scoring position.

2014 - Ehhh, I understand where you’re coming from in that the inclusion of all three robots was involved in achieving maximum assist points, but that has plenty of downsides as well. There were tons of qualification matches in which a team was basically asked to sit nearly still and allow a ball to be inbounded thru their robot (or perhaps with a few foot roaming radius - in which they could play selective defense). Those cases are marginally better than 2015 third robots, but not drastically. And there were even more matches in which alliances simply opted for 2-assist cycles. People tend to block these out when they think about late-season competitions and elimination matches, but 2-assist cycles were very common for most of the season. And this isn’t even touching on the bounce-back pass that developed late in the season or the fact that to score multi-ball autonomous routines you had to take a ball away from your partners.

2019 - This was a pretty good game for triple offense. There were tons of scoring opportunities, and ranking points that alliances would often have to score collaboratively to achieve. Not to mention the simple act of needing hatches to score cargo incentivized collaborative scoring. Additionally, because you never had to cross mid-field to acquire game pieces and the extension limits put in place if you did cross mid-field, the game could essentially be played “in parallel” with 3 robots trying to outscore the other 3 robots if both alliances desired to. There were also plenty of game pieces to be shared readily. The only downsides were the lack of protected zones and the crowded field did enable defenders to be very effective, which then could make defense worthwhile. But typically if a team wasn’t comfortable playing defense, they could still contribute meaningfully while playing offense. Because defense was effective, I wouldn’t rate this as an ideal triple offense game, but I also wouldn’t put it as “poor.” Very far from it.

2017 - Like 2016, this depended on various break points. Did having a 3rd robot scoring increase your odds of achieving the next completed rotor? Then have them play offense. But given the increasingly large gaps between each rotor completion (and particularly between the 3rd and 4th rotor), there were often periods where having an additional robot scoring gears would literally result in zero additional points. And at those break points, defense was the sensible play. Additionally, the cross-field layout encouraged collisions, and each alliance having loading zones in only one corner of the field (with visibility to one exit path obscured by the far airship) made driving paths predictable and relatively easy for defenders to disrupt.

3 Likes