Now that was a fight! As is Battlebots tradition, whenever a member from one Battlebots team leaves to become captain of their own robot, they must fight their previous team at some point in their first season. Our 2019 teammate Bunny came to Battlebots 2020 with her own robot, Malice, who we took on this past week in our second fight of the season. To clear up any confusion: there is absolutely no bad blood between us. Bunny has been part of the robot scene long before Shatter! was even a flicker of an idea. We're so happy she was able to captain her own team this season, she really deserves some time in the spotlight herself after being a part of other teams for so long.
In the end, this was a memorable fight for sure. If you haven't watched the fight yet, turn back now, because there are massive spoilers ahead. If you're ready for the post fight breakdown, hit the jump to keep reading.
It's impossible to go through this post fight report without addressing the judges decision - but I don't want to turn this blog into a line by line detailed rules breakdown and salt fest. There is one big reason for that: it doesn't actually matter what I think. To reiterate a bit of what I said in my post fight breakdown, what matters is how the judges judge things. What we need to do is learn from this fight so that next time we get in the arena we can make the changes necessary to come out with a win.
With that said, the way we'll walk through this fight is by reviewing our pre-fight strategy, and then breaking down what we were thinking as the fight went along, why we did what we did, the lessons we learned from this fight, and how we will change our strategies moving forward to hopefully prevent another tough loss in the future.
If you haven't seen it yet, here was our prefight video breakdown:
To sum it up, we're using the same front configuration from our Ghost Raptor fight (indeed, literally the same piece of UHMW), the same hammerhead (good ol' rusty), and a slight tweak for this week: our shorter arm (lil' stubbs aka stubbs mcgee). We've kept the bearing protectors that we made for the Ghost Raptor fight - even though Malice is unlikely to ride up that high, we have the weight and there is no reason to remove them.
The shorter arm is for three reasons:
Overall it's a very similar strategy to Ghost Raptor, but with one slight change. We were very worried Ghost Raptor would ride up the wedge and hit the bearing on our hammer, so we fired into their blade from afar at the first chance we got. For Malice, we wanted to see how their disc reacted with our front armor.
When we fight robots with big wide teeth, they tend to bounce off the plastic. However, Malice used their "purple pain" weapon, which is more like two skinny discs. Would they bounce off gently, or would they tear into the armor and send us each flying? We wanted to approach them and see how things went.
If we were gently bouncing off without much engagement, we'd stay on them, slow the disc down, push them around, and 'wait for a good hit.' However if Malice's teeth were getting big enough hits, we'd fire the hammer sooner - we wouldn't want the big hits to tilt a potential judges decision in Malice's favor. We can take big hits all day long - but generally that looks bad.
So if big hits were happening we'd try to take out the weapon sooner. Our weapon tends to work well against horizontals from past encounters - up until this point we were 4-0 against big horizontal spinning weapons: knocking all of their weapons out in 1 hit.
The first hit was nice and small, but the second was a bit too big for my liking, so we decided on the third approach we'd go for their weapon. We knew Malice had a lot of holes, in both the frame and the disc - the idea was to jam it up and see what happens.
One thing I want to say though before we move on to the hit - look how cool those triangles look while we spin around. Love that color-shift! You can also see a major benefit of the mecanums here - 100% efficient turning. We're able to turn back and face them extremely quickly.
Anyway, yeah, then this happened:
When people use the phrase "the hit heard 'round the world" I'm 99% sure they're referring to this and not the 1964 AFL championship.
This was a BIG hit in person. It's hard to get across the sound/impact of these hits on TV but this was nuts. Tons of sparks flying, and a belt going sky high - not to mention the something slamming into the wall somewhere.
There was a bit of a process happening here in my head. We saw a huge hit, and we saw a belt. At that point we knew Malice's weapon was done. That was great news. We saw the arm was still able to retract on our robot which was awesome... and then we noticed the hammerhead was missing (guess that's what slammed into the wall).
Well, the good news is we've now taken out 5 of 5 horizontal spinners in one hit.
I did some quick calculus in my head at this point - damage is scored on 3 categories: Functional, Effectiveness, Cosmetic. Cosmetic is mostly treated as a tiebreaker category, and functional is the most important. I figured that since their weapon was completely non-functional and non-effective, that our still-functional and mildly-effective weapon would win us the damage scoring. With damage won, we only needed to win one more category. Since we had an operational weapon, winning aggression seemed pretty easy. If we could win damage and aggression, and make sure not to overwhelmingly lose control, we'd have the fight in the bag.
A lot of people see hammer robots as control robots. I've always thought of our robot as a targeted-damage robot, more of a "sniper" robot. We try to be aggressive, and we use small targeted KE attacks to take out functional parts of our opponents quickly and efficiently. We don't tear off armor panels or crack frames in half - but we can burn out your drive by bending internal shafts, or take out your weapon by hitting the belt, etc. We're designed to win with functional damage and aggression - so that's what we set out to do. We'd try to keep control even - but we aren't designed to be a pushybot. Malice's Brazilian-special wheels (made custom by the Uarrior team) should easily outgrip our mecanum wheels.
So our strategy at this point was pretty simple - seem as aggressive as possible. This primarily relied upon whacking them repeatedly with the hammer arm. This doesn't do a lot of damage - but we believed we'd already won that category, and the shear number of hits we could deliver with the hammer should easily win aggression. So that's what we did - over 70 times. If you include misses, we're up around 80 hammer fires in one fight (it's a bit tedious to count these so I am not sure on the exact number). Deadblow famously had 112 "pecks" of the weapon vs Pressure Drop - so we've still got a bit to go to get up to their rate of fire. Grant set a high bar.
Of course, the problem is we are hitting Malice's AR500 strut with our titanium arm. AR500 is a hardened steel, and titanium is just not as hard of a metal. Mostly what this was doing was denting our arm. But make no mistake - these aren't love taps. That's a 10 lbs titanium arm, with an additoinal 10 lbs of mass around the hub, swinging at 60mph. Each one of those hits has as much energy as a major league baseball player swinging for the fences. It would easily kill someone - but Malice is a tough robot. The fact that they held up so well is a credit to their robot (and the defensive capabilities of AR500). If we'd been facing a lesser robot, these blows would have been caving in their top.
Pictured: Hammer-less-arm spark generator
Lest we forget, this is a manually controlled hammer. Pull the trigger, arm moves until you stop pulling the trigger - and every time you hold the trigger for a hair too long the hammer arm keeps firing into the stops. When this happens, there are mechanical friction clutches that will spin in place. These limit the torque of the system and keep the whole thing from exploding. However, if you do this a lot, the clutches eventually get hot and start smoking. With the hammer arm being so light and firing so fast without the hammerhead attached, it's nearly impossible to keep this from happening. The result is the clutches do indeed start to smoke at one point in the fight:
We simply stop firing the hammer for a bit, and the smoke goes away. This was not motors or controllers or batteries or any electronics at all. It's the torque limiting friction clutches that keep the hammer arm from breaking itself - they always tend to smoke a bit if we get up above 50 fires in one fight. It's just how they tend to work (maybe we'll add some fans next year to cool them off). Of course, no one knows what the smoke is except us, so the announcers and judges are left to make their own conclusions.
Regardless of needing to slow down on the weapon, we kept driving at full speed (not sure why the announcers thought we were slowing down) and had some nifty moves over the course of the fight. We knew we most likely couldn't win control, but that didn't mean we gave up driving. There was a moment immediately after we severed the belt on Malice, where we actually got stuck on top of the belt. So for a small amount of time after the initial hit, we were unable to get maximum traction, but once we got all our wheels on the ground again, we did have some fun with it.
I was particularly proud of the above pirouette where we were able to get them into the screws after their dodge of my initial attempt. This pirouette maneuver is a fun use of the mecanum wheels which may not be obvious at first glance.
Another moment I was proud of, is this one:
It was really satisfying to be able to push someone like this, as it's something the internet is constantly telling us that we cannot do. Particularly impressive here as well is that we aren't just wedging them or anything - their disc it engaged into our UHMW and this is a pure pushing battle. We were so, so close to getting them up and over the screws here too. Their right wheel just clipped the screw housing though, killing some of the speed and keeping them from getting pulled totally on top. Darn!
In addition to getting them into the screws a couple times, we also got them with a pulverizer.
But getting back to the fact that we're actually able to push someone. As noted many times on the show and on our blog, we have mecanum wheels. This gives us the nifty ability to drive sideways, at the cost of traction. Physically, there is no way for mecanum wheels to have more than 71% of the pushing power of a regular wheel - all other things being equal. However, due to real-life issues of manufacturing tiny rollers, you lose even more. You just can't get them as sticky as a big rubber wheel, so really we're at 50% of the pushing power of most well designed robots. And Malice is a well designed robot. Why then, are we able to actually push them clear across the arena into the screws? The answer, is magnets.
Magnets: how do they work?!
If you've been keeping up with this blog, or listening closely to the broadcast, you'll know we have magnets this season. We have some beefy permanent magnets mounted next to each of the 4 wheels. The front wheels, as shown above, actually have mounting positions for 2 magnets (one is empty in the video). These magnets, combined with the new 1/2" thick solid steel floor, allow us to get more downforce on the wheels. This extra downforce makes up for the reduced traction in the wheels, giving us back some of the pushing power we lost. Theoretically, we could even get to the point of having more pushing power than a regular wheeled robot which lacks magnets.
There is a bit of adjustability in the system, allowing us to turn down/up the magnet force. Since we aren't able to do any driving practice on the actual arena floors, we have no idea how the magnets will effect the system until we try it out in a fight, Because of this, we are very conservative with the magnets, and in the current limiting of our drive system (we don't want to get stuck or blow up our electronics). The first fight against Ghost Raptor we ran the minimum magnet force, and in this fight we turned it up a tad.
It seems to be working well here, getting us on par with Malice in pushing force, so maybe we'll turn it up again for the next fight...
Anyway, every time we hammered Malice with the arm, it pushed against Malice and decreased our downforce - losing our pushing power. We knew this would count against us in control - but again, we really wanted to nail aggression 100%.
As we all know, we lost the judges decision. Why? Who knows. They didn't release the scorecards for this fight (I wonder why?) and probably never will. So with that, we can only guess.
If we assume we won aggression and lost control, basically nullifying those categories, all we can imagine is that it came down to damage. If we keep our assumptions rolling, we think that we won damage on 1 card and lost on 2 other cards.
It seems that trading a hammer head for a weapon belt will not get us a win in damage points. Or possibly that the decorative triangles caused some to think our armor had been severely damaged. Going forward, we'll make sure to take this into account. We know we have to do more than just take out a weapon, we know that we have to be more careful with our weapon head, and we know we need to try harder to win the control points. Maybe we even need to get rid of the triangles so we don't appear as damaged?
We've always tried to play a very strategic game - but our reading of the judging criteria seems to be a bit off from how it's actually judged, and we need to adjust going forward.
We'll still be aggressive, but we'll also be a bit more controlled. Get ready for a leaner, meaner, more serious Shatter! going forward.
All that said, if we could go back in time and refight this specific fight - I'd probably keep the beginning the same. If we can get that one hit kill again, but change up the strategy a bit and use the arm as a spear, I think we can win the control battle. We can intersperse fewer hammer arm-hits to keep our aggression up, but by using the arm as a spear we can try to control them more. IF we get exactly the right angle, we could even potentially get a lift in. This could probably tilt the fight in our favor. Oh well, you live and you learn. It just wasn't a strategy we thought of in-the-moment.
With that, let's look at some post fight damage pictures. First up, Malice:
We still have no idea what exactly happened on that first hit. We can note a small ding on the pulley, but was that caused by a direct hammer blow? Was it the titanium arm fingers after the hammer came off? Was it the hammer or bolts themselves as they sheared? When and what took the belt off? We don't know, and the slow motion isn't slow enough to tell. What we do know is that at some point the hammer head went into the holes in the disc of Malice, and then was torn off as the disc spun and pressed the hammer into Malice's top frame member. We also know the belt came off very quickly. Some mysteries will live on.
In the end, Malice will need to replace their belt, probably that dinged up pulley, and add a fresh coat of paint on some of the parts.
Let's take a look at the damage on Shatter!
On Shatter! we lost one of our shorty arms. It's always a shame to lose an arm (titanium is expensive! - though less expensive if bought from our sponsor TMS Titanium) but we brought plenty of spares.
Outside of the hammer arm itself however, not much damage was done. We had some time consuming labor ahead of us to re-vinyl the decorative triangles, but the actual UHMW was reusable. There were lots of tiny dings on the front UHMW, but the vast majority of it was actually fine outside of the cosmetics. The frame behind the ablative was obviously unscathed, and there were no mystery gouges on our beautiful billet this time around (thanks again Prismier for the gorgeous frame).
The clutches, which we overheated a bit, had their friction-material replaced as a precaution before the next fight. We're never sure how much damage overheating actually does to them.
The hammer itself was reusable as well, and the rest of the robot was basically untouched. Ol' Rusty lives to see another day.
But lets talk a bit about the arm, and why the hammer head came off here, but not in other horizontal fights.
If you look closely at the hammer arm, it's actually bent, in addition to the broken fingers which hold the arm on. This was a major amount of force to do this kind of damage. We've thought a lot about this, and realized a bit of an issue with the design of the hammer connection.
When we first designed this hammer mounting system, it was 2016. 2016 was after the second season of the Battlebots reboot, and this may come as a surprise, but vertical spinners weren't really "meta" yet. Season 1 Biteforce was a control robot, and Tombstone had just won as a horizontal spinner. We designed Mega Melvin, our first hammer robot, to take on horizontal spinners. Now, I know we're in the Malice post-fight here, and Malice is a horizontal which just ripped off our hammer, but horizontal spinners usually apply a sideways load to the hammer. In this case, Malice managed to apply an upward force on the hammer tip. This type of loading is also very similar to what vertical spinners do (see Witch Doctor fight in 2019). Our hammer simply was never designed to take these forces. However the more we think about it, the more necessary it is that we take these into account.
The titanium fingers that go up through the hammerhead are great for taking shear. This covers all the forces our hammer applies to itself when firing into an opponent, and sideways loads from spinners. The arm itself is designed to flex for horizontal spinner hits, while remaining stiff when we deliver direct blows with the hammer. This flexure from sideways hits also deadens the impact and lowers the force applied to the hammer connection. It usually works great.
However, any upward force on the tip is transferred directly to the bolts holding it on. In an upward loading, it also does not flex the arm at all - the lack of flexure means the force applied to the hammerhead is not dampened at all. Malice put a direct upward loading on the tip, while pressing the hammer against their frame. It's basically a worst case.
We use three 1/2" grade 8 bolts - each one has a tensile rating of 17,000 lbs. It would make sense then that our hammer would take somewhere around 51,000 lbs to rip off. It turns out, after we looked more closely, that this isn't really the case. Our hammer head is actually a big lever. Here's what we think happens in an upward loading case:
The hammer actually pivots around that blue point at the back. That means a small force applied at the tip (shown in green) is multiplied into a large force (shown in red) applied only to the first bolt. It's not really sharing the load well - and the bolts pop one after the other like a zipper.
From running simulations and such, it turns out that it only takes something like 5,o00 lbs of upward force on the tip to pop the hammer off (lol, "only"). This is much less than the 50,000 lbs we'd assumed. Once the bolts popped, the hammer head then tore the rear two fingers of the arm off as it rotated out of place. Again, these weren't designed for bending loads - only shear loads, and so they failed.
In the Witch Doctor fight, the bolts were installed incorrectly, and popped very early, allowing the hammer to slide off without damaging the arm. Here though, the arm stayed attached a bit better and ripped the rear fingers out along with it as it went.
In our smaller 45 lbs hammer robot, Knock Off White - we've never lost the hammer bolts like this despite repeatedly firing it into all sorts of spinners. When looking back at how that hammer fails - the hammer itself is much thinner. Instead of popping all the bolts and flying off, the tip of the hammer simply bends upward, and a bent hammer is better than a removed hammer. In one fight we lost the tip of the hammer - but even still, half a hammer is better than no hammer. In a way, the extremely stiff and strong 1" thick solid AR500 Ol' Rusty hammer is possibly too strong for its own good.
The result of all this is not good for the current Shatter! hammer connection design, and we're looking into better methods to take this type of force moving forward. We're great at taking force through the hammer, and sideways of the hammer, but upwards and downwards at the tip is now a known weakpoint. With how many verticals exist today, which basically only apply an upward force to the hammer - and now with the possibility of horizontal spinners applying this loading as well (at least, 1 out of every 5 times or so) we need to take it seriously moving forward. But this isn't something we can do in the pits at Battlebots - it's something we'll need to do in the offseason.
Moving forward in 2020, we'll just need to be more careful.
Taking a step back from the post-fight damage and looking at the fight as a whole, we were really happy with the robot's performance overall, especially for a loss. It showed impressive pushing power for a mecanum robot - against a robot with super-sticky Brazilian wheels no less. We had a one-shot kill on a weapon once again, which is always memorable. The arm and drive lasted the full fight, being used consistently. The drive in particular was notable in that we barely even used much battery power, and none of the roller pins popped out (a major issue in our previous fight).
It's never fun to lose, but we learned a lot from this fight. We learned lessons we can use moving forward that should help us win judges decisions, and some useful real-life testing of the robot which will help us design a better version for 2021. We also proved out some more of our design choices, with the ablative armor working well once again, and the magnets starting to show their worth in the arena.
Keep watching to see who it is we fight next! This isn't a single elimination tournament (yet, at least) and we still have 1 more fight in the regular season before the top 32 is chosen. Hopefully we can win that next fight and make it in! Either way, we'll have another post-fight report here, and we'll have more pre-fight and pit content on our new youtube channel:
Of course you can also follow us on Instagram, Facebook and Twitter to see more bite sized content.
We'd like to thank our sponsors, without whom this wouldn't be possible: