Humanoid Robot Fights suck... BUT It's So Close to Perfection
Thanks, I Hate It
I love the show BattleBots. The idea of two robots in a ring with lots of intense destruction and fire was always an incredibly easy sell for me as a kid. I was so excited when they rebooted it and it was better than I ever remember it.
When I heard about the Ultimate Robot Knockout League, I thought that it could be very interesting! Sure, it might not have the intense battles that I’m used to. Nevertheless, I thought to myself I’m sure it will be incredibly engaging to watch humanoid robots fight in the ring.
How wrong I was.
I’m sure many people thought that it was amazing and that it’s truly a testament to modern robotics to see two robots not punch each other and instead break themselves. But quite frankly, the 2 1/2 hour live stream had 32 minutes of actual fighting and maybe 3 to 5 minutes of the robots successfully hitting each other.
This is the most boring sport I have currently watched that I think could become the most amazing sport of all time. But before that happens, I think that these fundamental changes need to be made in order for the sport to really takeoff and get good.
Issue 1: Depth perception
When you extend your arm all the way out to perform a punch, just how long is your reach? You might not know the exact answer, but you can generally place in space where your hand will go and where the punchable face is. The better your depth perception, the more effective your punch will be.
These robots currently have no depth perception. They recognize when their target is in front of them, but not how far away that target is. This is why when they predict a move, often times it will miss. They either attempt to perform a punch from too far away, perform a long kick and get too close, or perform some intricate move that causes them to bump into the other robot.
They also don’t have a good idea of what the other robot is going to try and do. They don’t necessarily need to know, but one of the best parts of combat is reading what move your opponent is telegraphing to you. A lot of the blocks and guards never really made sense. A lot of their fighting choices never worked out effectively. It didn’t feel much like a predictive model as it did just seeing the opponent and pressing a button and hoping it’s the correct move.
Issue 2: Shot Placement
In an MMA fight, why would someone attempt to hit the right side of your stomach? Because that’s where your liver is. A well placed shot to the liver can deliver a decisive knockout. Now, why would a robot hit another robot on the right side of its stomach? There is absolutely no reason.
One of the biggest failures of robot combat in the URKL is that where the robots hit makes no sense. The idea of a headshot is the only attempted move that makes sense. But so many of the more intricate moves would require something that makes sense to hit. Charging into a massive intricate kick that is far more likely to leave you on the ground than your opponent makes no sense. Only aiming for one or two sections on the robot makes no sense. Which leads into my final basic point.
Issue 3: They Aren't Fighting To Win
In the URKL, you get points by hitting your opponent and knocking them on the ground, forcing them to right themselves. It is incredibly clear that neither team had that goal in mind.
More often than not, the robots were attempting a very intricate move that caused them to knock over or fall off balance. It was very rare that there was any move that attempted to knock a robot off balance. Where were the leg sweeps? Where was the pushing? Where were any attacks to vital mechanical systems?
All of the moves they performed were incredibly flashy and would make a lot of sense in a fight where you’re attempting to look good. But the combat style they would need to adopt for this fight to be amazing is an incredibly different one than what we saw.
I Want Robot Destruction
My final point is that what I love about BattleBots is when a robot truly gets damaged. It’s incredibly expensive, and I understand why current humanoid robots are not likely to get completely destroyed. But that’s a big part of what I personally want to see coming out of robot fights. I would actually want them to be closer to gladiator than we would ever currently have. I want to see a mace in action. I want to see robot limbs get cut off or torn off. I want to see insane moves that humans can’t pull off.
Right now, it’s every other Boston dynamics demonstration about robot agility. And wow, that’s incredibly impressive, it’s incredibly boring in terms of combat.
I hope that these robots can considerably improve very quickly. I won’t hold my breath for a decade or two, but I do think in time this could be really amazing. Eventually, the robots won’t be damaging themselves, and their heads won’t come off, simply by moving.
But what do you think? Did you love it more than I did? Do you think it's peak or do you agree it could be improved? Let me know in the guestbook or on social media. Thanks!