On Perceived Strength of Defense

[sinewav]

Defense in 2021 is all about good sweeping. Actually, it's always been this way, but now it's much more obvious. Sweepers are the first line of defense and the longer they hold off an attack the more time your offense has to break the enemy zone. But there is another component to the strength of a defense, and that is perception of strength.

In the golden age of Fortress tactics, protection from holes was everything. Double defense was one way, and eventually we invented the sweep-box. These were incredibly resilient defenses that could give your offense ample time to attack. But there was a downside. If a defense is perceived as being "too strong," enemy attackers simply avoid it for easier targets. This was obvious in Ladle 128 where Paradigm relied exclusively on double defense. Occasionally they did manage a quick hole, but mostly their four attackers were met with five defenders. Perhaps if Force's outer defense seemed cutable it could entice attackers to stick around? Regardless, Paradigm's strong double-def provided no advantage at all in Ladle 128.

The standard 2+1 defense doesn't have the perceptual problem of "too strong." Sweepers will eventually give way and expose the defender, so the question changes from "should I bother attacking" to "how long will it take?" Let's take a look at ideal sweeping.

Right from the split, sweepers should offset their timing as they pan back and forth from the center to the Rim Wall. Good sweepers can dig, making it very risky to out-grind them at the center or Rim. Good sweepers shrink slowly, but they do shrink, and eventually enemies will get around them when they can no longer closely follow their trail. There are a few sweepers that can keep this up for a several minutes, giving their offense plenty of time to mount an attack. But what if we could double or triple the amount of time it takes for sweepers to expose their defender?

This is an idea I had and experimented with in pickup. I noticed that good sweepers will eventually hug the defender until it's no longer manageable to seal the center and Rim Wall. The front of the zone is especially a problem. So, I decided to try a heavily overlapped step defense that just barely enclosed the zone. Once reaching the final size it was fairly easy to maintain without shrinking too much. Sweepers had plenty of room to keep blocking. Even though this was in pickup it worked out great and here is why I think this is a viable defense.

First, the defense shrinks along with the sweepers as necessary (the front is the first to collapse, almost immediately. In every trial, the sweepers naturally started to wrap around the corners and then back of the defense. It was like having a double defense without the perceptual problems that come with it.

Second, even though I was completely uncuttable, a small, tight defense has the perception of being vulnerable. When attackers finally got passed the sweepers they would try to gank space that wasn't really there and eventually got tangled up and died.

Third, the small, "pre-shrunk" defense and close proximity of sweepers made it fairly easy to transition to double-def if needed.

Fourth, as the defense shrank, it naturally turned into a type of 3-person Sumo def. There were no awkward spaces from holes or cuts - each member of defense was essentially controlling the same space from when they started the round. The defense took a long time to break down, a time probably comparable to a sweep-box.

Just to be clear, this defense is basically a stalling tactic without the perception of being one. It's probably only slightly more flexible than a double-def. Ultimately, Fortress is a numbers game about survival, so you still needs to get kills.

Thinking back to the early 2010s I seem to recall coming up with this strategy for a team I was on. It may not have worked back then because we simply didn't have the skills to pull it off. But given that everyone is much better I wonder if subtle changes like this can make a difference to defense without changing perceptions?

[Kronkleberry]

A few weeks ago we saw you doing this in pickup and a few of us were discussing it on VC. We attributed a lot of the success of it to the fact that apple and wolf were sweeping (which supports your main point about having good sweepers) and a lackluster attack effort from the opposing team. It seemed pretty effective, but it would definitely be cool to see this philosophy under a real stress test.

Some potential flaws I noticed are:
1. It's sustained reliance on having 3 players at def. Any strategy that requires the attention of 3 players to block 1 zone for that long of a time puts your attack at a disadvantage. Sweepers are more locked into their role of sweeping and are less available to attack/gank/hole/follow-up center when necessary. You're affording the attack extra time to get in the zone but depriving them of the personnel to do so.

2. It forces sweepers to be less proactive. A lot of the success of sweeping comes from baiting attackers and/or exploiting their mistakes in order to get a man advantage. This is about being sneaky. You might show weakness, make them think they have a good opportunity to make a play on the defense, then pull the rug out from under them. Or you might see the attacker get inside their own tail while trying to set up an attack and then you put them in a box, and strangle them until they run out of space. Allowing sweepers to be in the open field is making the most of your good sweepers instead of reducing their role to "seal middle, seal rim, seal middle, seal rim". In turn, this will force attackers to take fewer risks (die less), and just wait out the inevitable shrink.

3. In practice we saw the shrinking happen pretty fast.. to the point where the extra time generated for your attack didn't work and it left you with a broken def. The sweepers had to hold the sides of the zone and basically turned into a 3 man sumo-def. With a def setup committed to shrinking, you put a time limit on the feasibility of this strategy. Sure, you might not be getting holed or cut as much, and a 3-man sumo def (with wolf and apple, especially) is super hard to out-gank, I think similar results can be achieved without committing two of the best players on the team to a sumo def.

4. I think you can do this without a perma-shrinking def. Good sweepers can and will block center<->rim for as long as they can anyway. If I have a good defender inside my sweep, I will sometimes let attackers in the sweep because I know they're more likely to die on the def, die in the space I leave them between the def, or just get in the way of the other attackers. If someone like ampz/tx/nanu, who is more likely to kill an attacker (or help the sweep trap an attacker) than get cut, has their walls covered 360º by sweepers, this is also reducing their effectiveness at their own position.

I think this is a viable strategy given a couple conditions: your sweepers are good, your defense is prone to getting cut, and your attack can get the job done with 3 players. It does pretty much exactly what you claim it do though, stalling without seeming like you're stalling. There is definitely an application for this strategy somewhere in the fortress meta.

I've also only played fortress for 1 year so what do I know.

[Word]

But given that everyone is much better

Are you sure about that? I don't mean to crap all over your thoughts and say everything used to be better etc., but I do sincerely feel like the general playing standard is much lower because fewer people play. Then again I don't play that often anymore. It seems to me like almost any high- or low-concept strategy that used to be a little far-out is now looking pretty good if you compare it to the general chaos surrounding it. Back in the 2010s we always went back to the step-def because it proved to be the most reliable shape, and while attackers know that any defense is a stalling tactic (it's actually the same in most historic occupations, the occupied always lose unless they get help from outside, haha), they are also constantly aware that their own defense won't last forever, which creates additional pressure to gank as soon as the risk seems worth the taking. That's applicable to any defense and any attacker IMO. If you, the defender, leave even the tiniest, scrawniest bit of zone "un-enclosed" by a wall, usually at least two people will instantly try to get there without thinking.

Back in the day, this specific defense would've seemed like a potential risk to me, with the sweepers being used to hole if they're getting that close for such a long time, and I think on the few occasions I saw something similar, that's what happened most of the time. So while I doubt it's strategical merit, the gif makes it look aesthetically pleasing!

[F0RC3]

Sweeps are certainly one form of strengthening the defense, so is a double defense. They are both valid tactics, and I would argue there is no clear winner. They both provide their advantages and disadvantages. And to say our double defense provided us with no advantages is certainly disingenuous, and in my opinion you are willfully ignoring our advantages simply because you do not like the double defense tactic.

With double def you now gain a fourth attacker. Even if the enemy team has most of their players back with a well timed, well placed hole 3 players could gank before the opposing team has a chance for 2 people to get in. Also when holing, if a team successfully gets in the double defender can immediately come up from the double def, putting all the strain on their on the enemies defense. As if they go attack with any more than 1 person you will gank. The double def is vulnerable to holes, however in the case they hole you you still have a defense and its almost like a +1 situation if their hole is anything but excellent (this assumes holing at even numbers). Our double def also has some vulnerabilities such as center attack, however I think this can be mitigated by using various center traps, and is not really inherently the fault of double def, but was our fault as a team.

Having multiple sweepers obviously has the benefit of having multiple people back to protect the defense in case someone gets in. and allows for some easier options of stalling people like a sweep box. But anything other than a sweepbox and its really only going to take one minute max to get passed the sweepers and reach the defense in some way, regardless of the illustrations you posted. In any of the situations it will never be perfectly symmetrical. You can either dig past on the rim wall when they are at the front to gain likely around half the defenses side wall. Or in the one where they somehow manage to cover both front and back all the way with triple binds, they will shrink too fast (but more than likely they wont do it perfectly and you can just dig past on the rim. In addition in your illustrations without a perfect turn on the front block, someone is almost certainly going to dig through the center block (maybe not in the beginning of the round against ampz or deso, but after a rotation or so when their rubber has obviously decreased? definitely). The real benefit of sweepers is being able to kill attackers as they prepare to hole, gaining a player advantage through the kills, and that's one of the failures that we as a team needed to work on. Staying alive, avoid the sweepers, hole well, and gank.

And in my opinion that where the folly of sweepers comes in. The defense is extremely vulnerable to holes in comparison to a double defense. Sure a double defense may be more easily holed, but you always have the inside def with most of the zone, disallowing a gank. But with the right tactics, a hole can almost instant gank a normal defense. I had an idea to be able to setup a hole that will be hard to block and get people in rapidly, but I had only shown it to over the day before and we didn't have time to work on it. A fast gank wouldn't necessarily be needed if your team has better sumoers though, but would still be preferable.

With being able to fully setup our double def and the right holing tactics, we would have had a chance at beating mym. But simply put that was the first time I had ever setup double def at the beginning of the round, Johnny was doing a reverse bottle out of habit when a normal one would have been more efficient. We didn't practice various possible center traps to disallow centering when double defending. We died a lot for no reason. But in no way was that the fault of the double defense tactic. The double defense tactic is solid, it provided us with a rather large advantage that we squandered in many situations by bad turns all around on our team.

In summation. Both defense tactics have their advantages and disadvantages. Double defense does not put you in an immediately losing situation. Better hole tactics allowing for fast ganks are required for double defense to be at its strongest, which we did not have practiced. And we just need to turn better. But none of this is the fault of double defense, an actually great tactic. I am also sure I missed some advantages and disadvantages of both,

As for the defense that you tried in pickup, I think if you ask around the general consensus will be that the vast majority of people would think your normal defense is better. In my opinion, it relies too much on the sweepers. Having good sweepers is great, and they are an asset to the team, but to rely on them so heavily is an overall bad idea.

[sinewav]

Thanks for reading my article. The purpose was to Illustrate that defensive strategies are not purely about effectiveness, but also how they are perceived, which is an important factor. In 2021 it's basically assumed that you're putting really great players in sweep since 1) it's probably the most important role these days and 2) the sweepers are your "2nd wave" of attacks. But even if your sweepers aren't great, giving them room to enclose the defender has a similar benefit as double defense without discouraging your opponent from attacking.

But given that everyone is much better

Are you sure about that?

Absolutely sure. I'm constantly amazed at the things people can do today. You should come play, Word!

[blondie]

Korean Meta was further ahead of things strategically than any team I've personally played on. There were defensive and attacking concepts there I'm not sure any other team has ever used (maybe they have in past months, but seems like not judging from this thread, idk) and it felt like we were just scratching the surface.

I wrote this at the time (December 2020) but never published:

Forming Gossip Girl
In the beginning of October, I cocast Ladle 118 on Syllabear's twitch stream. Afterward I began talking to Ampz about his team and his experience with Evolve, who had just lost in the finals to Twixted Arrows. He felt like Twixted Arrows were very beatable but unforced errors had doomed Evolve. At the time I had only returned for a few weeks, so I asked him about players and rosters. We agreed very quickly that we wanted to play pretty standard, no double defense or complicated tactical stuff, certainly no sumo defense, which had doomed at least one of his previous teams. To play standard, we wanted a roster than could be confident in broken zones, that could afford to give up holes and then play "post-hole" situations. We thought overwhelming talent is always the first approach a team should explore because overwhelming talent is just the best way to win fortress matches.

Our list got pretty narrow pretty quickly; there were only so many skilled available players. Appleseed, Mr, and Jericho came on board pretty quickly. Apple had little interest in fort and needed some convincing, but Mr was motivated to fight and win and brought him along. We tried to convince Koala to take a ladle off from Evolve, but his interest was never really there and, anyway, he believed in the squad he had at Evolve. DGM soon rose to be a good candidate as a sixth once he started getting active again, with Nate and vein eventually rounding out the squad.

Gossip Girl in Ladle 119
Our plan for Ladle 119 was relatively simple. We would play standard. Sweeper's job was to get kills, attack's job was to use numerical advantages to take good risks and win rounds. We wanted to be rock solid in defensive sumos so we built from the back and put a winning WST team in positions 4, 5, 6, with Apple, Mr, and myself. The attack was a pretty balanced three-headed monster with DGM, Jericho, and Ampz. I felt like Ampz gave us good sustained attacks in longer rounds, DGM and Jericho could could hole or center interchangeably. Nate and Vein were also great substitute options. We won a friendly round-robin scrimmage without any matches being particularly close with the downside of putting Twixted Arrows on notice of a new challenger to their dominant five-Ladle win streak. Still, this was confidence building and showed the basic concept of the team was going to work.

In Ladle 119, everything went to plan until we were up by 30 points in the first match of the finals. With something like a 80-46 score, we started throwing rounds to Twixted Arrows. We couldn't block late center, people were suiciding early, we lost longer sumos. Twixted Arrows don't storm back but rather march: they slow things down, they grind out rounds with their experience, cohesion, discipline, and belief. We make all the mistakes and they take the first match.

The second match we stick to our plan and blitz them. "Don't let them dictate the pace" is the gamecall. We don't allow that and we expose their sweeping with quick holing round after round, winning the match by sixty or seventy points.

Match 3, Twixted Xats finally goes to something we'd anticipated and thought we were prepared for: counterholing. Rather than sit back and try to defend, they just hole us as fast or faster than we can hole them. Our whole plan was to play post-hole and win these sumos, but we lose them consistently. We lose some fast because sweepers are caught outside and we're ganked. We lose some slow because we can't maintain numbers in long crowded rounds. Facing a sizable deficit and unable to find another move in the bag, grinding and blocking problems creep back in and we completely unravel in the second half of the third match. Twixted Arrows wins their sixth straight.

Learning from 2nd place in Ladle 119
That evening I was not pleased. It seemed like we had basically lost for all the same reasons Ampz had seen Evolve lose the ladle prior: unforced errors, lack of discipline. Twixted Arrows had made the winning adjustment and we didn't have an answer. Still, it was tempting to run things back and avoid overreaction. After all we had blown a huge lead in the first match and completely dominated the second. We'd been close.

The problem was the next finals was in a European server. I didn't feel like our current roster or approach was good enough or, more importantly, mentally tough enough to climb back in a month in harsher conditions. Furthermore, the way the whole event had played out exposed this formulation of the team as being, ultimately, front-runners. We ran up the score when things were going well, but we choked when things got close in the first match and crumbled when facing an early deficit in the third. We had floated to the finals having won every match we'd played: four in the scrimmages a few weeks earlier, and then every match in the swiss. Coming back a month later with chinks in the armor and this shared experience of failure meant that if things got tough all the doubts and memories of failure from the previous ladle would resurface. The team had been designed to have overwhelming talent, but it was doubtful we had managed to assemble that. We might have been the most talented team that ladle, but overwhelming talent requires a margin of skill that obviously we did not attain.

The biggest trap of the "overwhelming talent" approach is when you think you have it, but you actually don't. Then you are playing with an approach that basically claims to win by default, and when default doesn't work, your team has no other options to go to. We had some team philosophy, but in reality it was just fancy default fortress. Players had clear roles and goals. And it was almost good enough. Talented team with clean and clear standard approach.

Our post-hole approach had failed dramatically when put under the pressure of counterholing. The WST-champion sweeper and defense combo hadn't done well at winning defensive sumos. While some of team were confident and wanted to run Gossip Girl back in Ladle 120, other members had doubts.

What had worked well was the attack. The quick holing and pressure we created had been effective. We didn't think anyone had managed to stomp on Twixted Arrows like we had in that second match. Jericho had some good ideas about holing that we hadn't been able to coordinate but seemed like things we should be doing. (Another downside of the "overwhelming talent" is that if you believe you can win when both teams play standard, you are unlikely to spend any time preparing or coordinating anything.)

[... omitted section here about roster recruiting for Ladle 120 ...]

So we had a roster that could clearly win if things broke right. We had also healthfully divorced ourselves from the notion of overwhelming talent.

Discovering the Plan for Korean Meta
What happened next was maybe the worst thing that could have happened: I had a week off from work for Thanksgiving holiday. That was a lot of time to overthink things. And I definitely did. Jeri's major idea was what he called quick-gank-holing. It's not something any of us had invented, we all knew that, but he was pushing us to take the idea more seriously. Basically the concept was to hole and immediately have two people enter the hole. This put a lot of pressure on the defense because the threat of immediately ganking after a hole is so immediate. The faster your gank happens, the more the enemy has to fear your holing. They can't try to play post-hole sweeping like we had tried the previous ladle.

I was thinking about sweeping against quick-gank-holing. The sweepers can play post-hole, but only if they are positioned well enough to prevent the quick-gank-hole. I started thinking about how they would do that. Where would they have to position themselves? How would they set up their tail to keep attackers split apart? I started watching recordings of sweeping. I tried to experiment with different patterns in pick-up.

I had thought I understood the strategic principles of defense. I had even recorded myself talking about it. I had thought my understanding was slightly ahead of the general fortress understanding. I thought I knew why double defenses and sweepboxes didn't help teams win, even while so obviously made defenses stronger. But the more I watched recordings with my theoretical model in mind, the more I realized it didn't make any sense.

I had thought that the purpose of a defense was to generate a numerical advantage for the offense to spend on attacking, whether that meant taking risks cutting or just holing. The sweepers were supposed to be this sort of player advantage engine powering up the attack. Kind of like adding to the team's risk credit balance. A kill gives the team a full death to work with in terms of risk-taking they can afford. The problem was there wasn't any time for them to do this because the attack needed to take risks immediately or hole as quickly as possible. I thought I could explain this with the concept of debt: attackers spend risk credit before the sweepers generate it, basically gambitting material in the form of risk with the belief that their sweepers will make it up on the back end. But really the sweepers didn't have any time to accomplish what I claimed they were out to accomplish. More often, a good team functioned when its attack recouped its own risk deficit. The pressure or crowded zone scenario was the type of environment to get kills. You had the threat of ganking to bait or even force opponents into tight spots.

So what was a sweeper? Well by my previous definition, the one I'd given to my teammates in Ladle 119, sweepers had to prevent or recover from holes, they had to never die, and they had to generate kills. But now I was realizing they had to generate kills quickly to have any value compared to just holing and crowding and attack. This "quickly" part was what made it obvious that this positional definition was unrealistic. How can you have positional discipline to defend against holing or play post-hole, the durability to never die, and the playmaking aggression to quickly generate kills? It just makes no sense. Maybe the best sweepers can peak a few rounds a match to achieve these strategic goals, but it's plainly not realistic to achieve consistently. You can't build a strategy around that position because that position doesn't exist.

My first instinct upon realizing this wasn't great either. It seemed like sweeping was a fool's errand, so I toyed with just eliminating the position from our plans. We could just double defend, send four attackers, avoid this impossible position of sweeper. I had even won ladles in the past with that exact plan, so I had some good memories of being the outer defender. But I also knew what would happen. Teams would just drop everyone back. They would have one attacker keeping our two defenders busy while they played 5v4 against our attackers. So we would have four attackers but be outnumbered. What's the point of having an extra attacker if now you're facing even more opponents?

That, it turned out, was the key realization. To succeed, an attack needs to deliver a payload to the opponent's zone. The payload is relative to the number of defenders. If you want to win immediately against a single defender, you need a payload of three. One to hole, two to gank. The necessary payload against two defensive players should be four attackers: three alive after holing to gank against the two defensive players. Theoretically, a sweeper and defender should be stable against three attackers. If the attackers hole, they only have two gankers. As long as the defensive team gets enough space and doesn't get ganked right away, they should be fine. So maybe we could just get rid of our other sweeper. One sweeper was theoretically enough to defend against three attackers, the other sweeper could just go up.

In fact, I realized, the other sweeper needed to go up in order for the defense to work. If we had four attackers, they couldn't afford to send any of their sweepers up because that would give us that 4v2 payload that would be an instant win. By going up fast, our fourth attacker was pinning their sweepers to a defensive role. Their attack could hole, but they wouldn't be able to transition anyone forward and they'd be stuck with a relatively inert 2v2. Meanwhile, we could hole fast or slow, but either way we'd have a more dynamic 3v3.

Now 3v3 doesn't sound much different from 2v2. It is 50% more attackers, and it is a 50% more crowded zone. It is three players who needed to stay alive and in the zone for the opponent to be stable against two. That's 50% more failure points. When we hole, we had three players who could immediately rush in for a quick gank. Our chances of winning immediately after holing were higher because we had this third player to get to the 2v1 gank before enemy sweepers rushed in to stabilize. Fundamentally, in the chaotic transition moment when the defense is holed, the more space and the simpler things are, the easier it is to stabilize. The more crowded and more complex the tails are, the harder it is to stabilize. If the crowded offensive zones were where kills came from, then if offered a 3v3 attacking or a 2v2 attacking, you'd prefer the 3v3. How do you get to the 3v3? You need to get your fourth attacker forward faster than the enemy. We figured we might as well send that person forward as soon as possible. Don't wait for either team to hole or anything. Just go forward after splitting and making sure defense was set.

This also clarified the sweeper's role. The lone sweeper now just needed to make sure the 2v2 that ensued after the hole was stable. Make sure we don't get quick ganked. It turned out this was a lot easier to do when there was more space around the defense due to that other sweeper being gone. It was easier to navigate around and get into the zone. I sort of had a hunch that a lot of what sweepers were doing when things were crowded was useless. They were driving just to survive or maintain space in a crowded part of the grid. Take away one sweeper, and now the lone remaining sweeper has more space. More of their actions are useful or meaningful. So you take away a sweeper, but you don't lose a full sweeper's worth of effectiveness because now the other guy gets a lot more impactful because he has more space to work with.

The other element was that it was easier for opponents to hole. This meant they didn't spend time trying to kill the lone sweeper. They thought holing was going to win them the round. We were holing fast, surely they should be holing fast too. The problem was after they holed they only had two gankers and we had two defensive players. The normal situation where the rest of their team comes up wasn't happening because they couldn't leave to come up. And our attackers were schooled to not come back. Coming back would have relieved the pin on the enemy sweepers. It made the zone more crowded and less stable. No joke, return to reinforce defense and you end up making it weaker.

Essentially we got the 4 attackers that double defense promised but didn't deliver. Of course, opponents still had the option of dropping more players back like they do in double defense. We didn't think that would happen long term, although we were happy if it did. Fortress teams are very concrete. Very literal. They look at hard dynamics. They would see our two defensive players and think that was a vulnerability. Or they would see our quick holing and double defend. That happened a bit. We were happy to see that because we knew it was bad. Now we could just play 6v4. We sometimes didn't even bother attacking the double defense, begging the question: "If we don't attack at all, are you still going give us the respect of having two defenders?" [/b]

Korean Meta in Ladle 120
Strategically the ladle went as planned. We lost to deaths on the grind, not relieving center double teams, slides in American servers. We won on the strength of our plan. One sweeper, four attackers. Removing a sweeper was strengthening our attack and our defense. We were far enough ahead strategically to have a pretty big margin for some serious blunders. In the first match of the finals, I gifted the opponent about 30 points singlehandedly in the first three rounds. Once we got to the tiebreaker against Twixted Arrows in the European server, we felt like we had a huge edge. We had solved our sliding issues by simply getting back to Europe. We had already fixed the early death and center trap issues, which were the only reasons we had lost on the day. We felt confident and were playing pretty relaxed. A good novel strategy is a bit self-reinforcing. When the team sees it working, they all start to trust that it will continue working. Playing with confidence and without doubt is a huge edge. Meanwhile, our opponents kept changing their approach. They would try to ASAP hole us, they would drop an extra sweeper, they would double defend. None of which actually made a difference. our confidence at times bit us with some costly errors, but we pulled out the tiebreaker and then the next two matches in the finals. The plan had worked beautifully and I was very pleased with the team's overall cohesion and execution.

Learning from Ladle 120
It took me another couple days to fully realize what was going on though. I talked with Jeri on the day after about the next Ladle. We talked personnel, our need for bringing on at least two American players, and also strategic adjustments. We agreed teams would be better prepared for us, even it that just meant mirroring our approach. More likely it would be stuff like sweepbox or double defense. These latter options did force us to drop everyone into sweeping or midfield roles. But Jeri was all in favor of that. We speculated about how our holing threat should be present even if the actual players were back in midfield. They could charge up and counter attack.

But I wasn't sure. Wouldn't we just get counterholed? We would lose the race, the enemy was closer to our base. We had to drive across the grid to hole.


"No," Jeri said. "It will take time. They have to call it out, they have to decide who will hole, they have to drive there and do it. It will be crowded."

He was right. It would take time. That was all the edge we needed. I kept working on the idea over the next few days to figure out how to communicate it to the team in a clear and concise way. That took some tinkering. I kept exploring the idea.

I also felt strongly that intuitively it was based on the same principles as our Ladle 120 approach, even though it seemed outwardly extremely different. Instead of four attackers, didn't we have zero, after all? I decided no. Our Ladle 120 approach really had nothing to with four attackers and everything to do with attacking with a numbers advantage.

The way we did that was by transitioning a sweeper forward. We did it the whole ladle and enemy teams never really reacted by simply transitioning an attacker back to maintain parity. (Twixted Arrows sometimes had three sweepers, but they had already been experimenting with that in previous Ladles, and they would start with three sweepers rather than transitioning one back.) Why hadn't teams reacted in this simplest of ways? Well it seemed like such a concession. We have this flimsy looking one sweeper set-up, it feels awful to concede that you can't just hole it and send your third attacker back. I mean it really is like giving up. Teams just kept thinking they could punish our set-up and more or less kept failing.

Our midfield holing approach would generate numbers advantage by transitioning, rapidly, the whole payload forward at once. This meant our attackers could provide other value during most of the round, then counterattack at a key moment. As with our previous approach, this would create a critical moment where the enemy needs to make a decision without much time to make it. When they notice out counterattack, they are forced to react to it.

Now I started to realize the real advantages we had been generating. There were some concrete space or timing advantages, for sure, but we were really attacking opponent's lack of communication, their lack of understanding of the game, their lack of coordination. We were finding all these new things to attack. Meanwhile our opponents were still playing extremely concretely, looking only for huge material or tactical advantages that simply don't readily exist in organized games between even teams. We found a strategy that would generate a ton of small advantages in all sorts of different subtle ways. We had space and timing advantages, we were pressuring their whole team structure right down to their captain's leadership and playcalling skill.

Key Takeaways

• Sweeper position as traditionally (past 10 years) defined is basically incoherent
• You would rather defend a 2v2 sumo than a 3v3 sumo
• Effective defense requires pinning players to one end of the grid
• Space and speed and positioning define future threats. These are as important as present structure and walls (think cycle position and cycle potential to be in space rather than where your walls are)
• Attack things besides zones. There is plenty out there.

[sinewav]

Thanks for the great read. Your articles are always full of insight and humility. This type of analysis helps reinforce the idea that competition is a mental game as much skill (regarding timing and decision-making). One topic I don't think has been explored in depth is what "pacing" means in Fortress. What does it mean to change the pace of a match and how is this an advantage or disadvantage? Maybe you'll do a future podcast?

Sweeper position as traditionally (past 10 years) defined is basically incoherent

It does seem like the role of sweeper is overloaded these days and maybe it's time to redefine, and possibly rename, the position. Even positions 4 & 5 appear to have diverging duties at times. While reading your article, I wondered how much teams use sweepers as simply a form of insurance in case things go wrong.

[Lucifer]

I've said it before and I'll say it again: Sorry about only recently stumbling onto this thread. Didn't mean to wake the dead.

Also, Nothing discussed here is new. Not saying it's not worth discussing, just saying there's nothing really new here. We already knew stronger sweepers make a better defense. We also knew that under certain circumstances, a sweeper goes to a double defense because that's what the situation calls for. It's annoying when a sweeper doesn't do this when it's needed, and it's even more annoying when a sweeper *does* do this and it's not needed.

There are, on an 8 person team, two levels of sweepers. It's frustrated me that people don't play the midfield. It's all ball-hogs. Either you're attacking the other team's fortress, or you're hanging back and being a "sweeper". Ok, but that's a total of 4-5 people. On an 8 person team, you've still got half the team deciding to attack or hang back and be a "sweeper" and step on everybody's toes,

Play the midfield, Seriously, Your defense only has to be stronger than the attackers that make it there, so stop them from getting there in the first place.