S6 Tournament | Matchmaker, Matchmaker, Make Me a Match

BattleBots matchmaking has always been a bit of a black box. We know the showrunner is responsible for the majority of the matchups and we believe has veto power on all matchups. During Fight Nights there are a lot of reasons to see a given matchup. Witch Doctor’s reliability was tested against Duck, Kraken’s ability to fight horizontals was tested twice, and HUGE saw two rookie opponents that could reach the body. There have been calls to make matchups random, and while that sounds appealing at first glance, there’s still the possibility that a robot exclusively faces its hard-counters. Teams have some ability to back out of a match, but the selection committee won’t like that. With how strained the schedule is to make matchups, it might be no different than forfeiting a fight.

With the move to Discovery in Season 3, the show’s format changed to have a pre- and post-season. The preseason “fight night” format gave the majority of teams unprecedented opportunities for airtime. In Seasons 3 and 4, most teams were in at least 4 episodes between Fight Nights and whiteboards. The schedule changed to a standard fight night schedule of 3 fights per team to accommodate the filming time set aside for Bounty Hunters in Season 5. Most teams were still able to appear in 3 episodes during the main season and almost everyone who didn’t make the round of 16 was able to participate in Bounty Hunters, netting them a 4th appearance. 

Counting appearances is important because each time a team appears in an episode, there’s a minimum amount of airtime they’re going to have. Airtime is a big chunk of what sponsors are interested in. Since the show is a fixed length, any airtime one team picks up is airtime another team can’t have. Main events come with a fair amount of interview time so that further increases the airtime of the teams. The short featurettes in the middle of the episodes vary, and the film has several more of them than they’ll use on top of hundreds of hours of “B-roll” that probably won’t get used. Landing in B-roll is always a nice treat, it's another opportunity for fans and sponsors to see a team’s brand. 

This season, they changed up the format to create space for the ending of the Bounty Hunters content. There’s a little more to that, but spoilers and NDA. There are only 9 filming days over the two weeks we’re there. Normally, each session corresponds to an episode. The goal is typically about 10 fights per session with about 8 fitting into an episode. With 2 sessions a day we get to 20 episodes. With the tournament getting bumped up a day, we needed to get through 10 episodes in 8 sessions. They set the goal of 12 fights per session so they could get through 3 episodes per day. We didn’t hit that goal. The first two days of fights always have terrible flow, but we also had a mix of power and technical issues that delayed the first session, which led to cascading delays later on. There were further technical issues during the fight night days too. Delays are expensive. We’ve been told it costs about $1,000 per minute to film at the arena. A lot of that is the sheer quantity of people required to make it happen, some of it is equipment costs, and finally the cost of having talent on set. Talent refers to Chris, Kenny, Faruq, the Judges, etc. People who are paid to be there. Builders are competitors, not talent. The distinction is a critical part of our status as a Game Show and BattleBots’  ability to dispense prize money. Something to keep in mind is BattleBots is a TV show about robot combat. Robot combat is a sport, but the show is not a sports broadcast. Maybe in a few years, it’ll be a sport, but we’re not there yet, not by miles.

The problem with missing the goal of an average of 12 fights per session (for a total of about 96 fights) is a bunch of robots didn’t get a third fight. Most notably, 8 robots who qualified for the tournament only had 2 fights. 21 of the 62 robots in Fight Nights didn’t get 3 fights. Sure there are a couple that just wouldn’t have made it to a third fight, but most of them would have. One question I’ve fielded a few times is whether this is fair. What we were told by the people in charge is they felt it was more important to give as many teams as possible a third chance to be on TV. The 8 teams who were accepted into the tournament with 2-0 records would get that minimum count of 3 fights in the Round of 32 episodes. Since the episodes only air with 7 matches, at least 10 fights would be downgraded to basement tapes to be revealed on YouTube. For better or worse, the decision was made to air those fights alongside Fight Nights airing rather than in batches after the season was over. Some of the fights have done very well on YouTube. Many are absolute duds and did not encourage people to pay money to watch the rest of the show. Dud fights happen. We all try our best to avoid creating them, but the fact of the matter is the robots can’t all be perfectly reliable all the time. We ended up with 80 fights filmed as part of Fight Nights, 70 on TV, 10 on YouTube. 

Here at Team HyperShock, we place a lot of value on showmanship. We understand and have actualized the concept that it’s a TV show. BattleBots is other things too, but it is and must be a TV show first. Before anyone knee-jerks into saying “no it's a sport!” consider that professional sports sometimes take a break from play to wait for a commercial to finish. Sure there’s some weirdness to how live broadcasts work, but the commercialization and marketability of an event must be at the forefront of everyone’s minds. Professional sports fine players for neglecting interviews after events. It varies between sports and leagues, but some amount of the competitors are contractually obligated to spend a certain minimum amount of time talking to the press. At BattleBots, that’s all baked into the show format. Before each fight they’ll interview a team’s pit, then after the fight, Chris and Kenny will interview them at the arena. Both the winner and loser get interviewed pretty much every fight. For main events, there are extra interviews done in the pits. Technically a team can try to say no, but they’ll just circle back around and get those interviews after the fight if they have to. Refusing to cooperate with production is a really fast way to get erased from the show, if not banned from the event. BattleBots as an organization doesn’t have the capital to put on nearly the same quality of event without the millions of dollars of investment from cable television. It’s wouldn’t be remotely sustainable to do anything of the sort with GoPros and volunteers. They’d have to charge teams a sizable registration fee and the robot build quality would fall right off a cliff. The incentive to build interesting or innovative designs would be solely for personal fulfillment and the average build would shift to much more generic, competitive designs. More importantly, it wouldn’t be two weeks long. The only way to pack more fights in is if the robots can be turned around faster. Without TV, the sponsorship opportunities dwindle to nothing, so the solution for serviceability shifts from design intent to overall simplicity. The fights would be less interesting, the robots would be much flatter, and at least half the field would be unpainted or monochrome. 100% of the cost of competing would come out of teams’ pockets. The competition would be much less geographically diverse too as teams would be paying for their shipping on top of travel, hotels, and food. I just can’t see how the sport is supposed to have sustainable growth without the backing of massive investors. 

On the topic of growth, the competitor field size is massive. Before Covid dashed everyone’s plans, there were over 80 robots on the preliminary roster for Season 5. BattleBots was still going to be the same 16 days on-site with 9 filming days. Using this year as the yardstick, we’d be lucky for any team to get a 3rd fight before the tournament started. Season 4 had 68 robots (plus Nightmare for one exhibition fight). Most of the field managed to get 4 Fight Night fights in, but the tournament was 25 robots with the top 8 getting a bye round. It was only 23 matches. The round of 32 is 9 more matches, and then there are 2 play-in fights which had to happen the same day as the 16 ro32 fights. So the tournament is 11 fights longer than Season 4 and gets stretched over more days so Bounty Hunters can be going on at the same time. They can’t compress it too much because they don’t know in advance how the Bounty Bosses are going to perform in the tournament. They don’t want to make a team fight as a Bounty Boss while they’re still in the tournament. I’ll take 3 fight nights in exchange for Bounty Hunters. I didn’t like only having 3 fights last year, but once I saw how great the final product for BH was, I have been convinced of its value. Only 2 fights if you win both though? It’s not enough.

This first slide is a beat-map within the robots we more or less interacted with. I’ll admit I’m just not motivated to do this for the whole bracket. As comparison tools go, it’s pretty weak, but it does a great job of highlighting how narrow the scope is for analyzing our seeding. The robots shown with a blue highlight are the ones who qualified for the tournament. Each has its seed on the left side and every robot has its current-season record listed below. The arrow between Copperhead and Lock Jaw is yellow because it’s the only JD on this map. There are 11 matches shown here and only 1 didn’t end in a KO. I didn’t trace the maps for the robots who didn’t make the tournament, but in an interesting twist, Gruff’s one win is against Switchback. Kinda closes the loop on the wins shown. 

This second slide shows the high-level stats for us and the two robots we beat. It's worth noting that Craig Danby and Donald Hutson have both been competing since the 90s. Danby has competed in at least one season of almost every major televised event and Donald is one of the most acclaimed competitors in the sport. When looking at a bot’s legacy performance like overall wins, it’s important to consider how many fights they may have missed out on or how many fights were unaired or highlighted. Collecting that data without going back through old episodes is difficult. Some of it is captured by the fan-maintained Wiki. The reason I think those details need to be considered is that the selection committee is tasked with putting together the most interesting and exciting tournament they can. It’s not about carefully selecting rankings based on win-loss records and competitive stats, the goal is to create entertainment. The builder community has more or less accepted that the rankings are at worst by quadrant other than maybe the first two seeds, so it's less about being ranked tenth than it is about being in the second group of 8. I would essentially argue that there’s no strong reason to truthfully say the 12th seed is worse than the 9th seed. (I’m picking seed numbers at random here, not picking on any individual robots). 

This last slide shows the two other robots who beat Lock Jaw. The reason I’m focusing on Lock Jaw is it’s the larger prize and harder-won fight of our two fights. Lock Jaw was 2-1 going into the bracket last year and swept the Beta Bounty, beating the 2 seed, Bloodsport, in the process. Lock Jaw going 0-3 was massively surprising to everyone. The other detail I’d like to point out is Blip’s opponents (Lock Jaw, Rusty, and Overhaul) have seven loses and no wins between them. Of the three of us, I’d argue Copperhead had the best strength of schedule. They’re the only ones who have an opponent with a win, and Fusion’s win is over cobalt, a robot that otherwise went 2-1. 

Looking through the records of the top 32, with one exception (Sawblaze 2-1), the top 13 are undefeated (3-0, 2-0). Through 30, there are all 2-1. The 31 and 32 seed play-ins are all 1-2 except Defender (who only fought robots starting with the letter R). So for once the bracket is pretty well aligned with W/L records. The 2-1s are a bit of a scramble, but I don’t think any of them are egregiously out of place. How they got there is a different story, which I kinda go into at the bottom of this blog. Now, how do we deal with the schedule length disparity? As the lowest-ranked, undefeated bot, I can’t argue with our seed, not really. I mean, I can, I can argue about anything, but the seeds are fairly quantitatively justified. Where it all falls apart fast is many of the 2-0s would not have stayed undefeated after a third fight, especially if they had all fought each other. Many more robots would have ended up 2-1, but at the cost of a lot of 0-2s and 1-1s not getting a third fight. I can accept and understand their “share the wealth” explanation, but I still struggle with whether it was worth it. 2 fights aren’t enough to plot a trajectory. In S5, we only squeaked into the bracket because the drive issues we had in our first fight were resolved in our third. I know that because we were told as much. While we lost our bracket fight badly, the robot didn’t stop working until the last flip before Hydra send us into the toaster slot. Our first fight we'd been winning and generally doing well, until the drive chains fell off. Stupid oversight, easy fix. Our second fight lasted all of 20 seconds. Nothing could be gleaned from it, just another tally in the wrong column. Our third fight, everything worked just about perfectly. Well enough that only we'd notice the faults upon returning to the pits. The comparison of that first fight and the third fight is how we earned our bracket seed. We showed up with a mostly working robot and fixed the only glaring error we encountered during fight nights. I think a fourth fight would have gone well for us, especially if we'd rounded out our diversity with a horizontal.

Skip ahead to the appendix if you don’t want to read a short essay on where I think the show should go.

We were short on fights by about 10, that's a whole session’s worth of fights that didn’t happen. There were also a lot of dud fights this year, like most years. I believe that the way forward in the short term is to cut the field down to about 40 robots, ensure everyone gets 4 fights and focus on more engaging content. A more curated field should lead to fewer unwatchable fights. More importantly, by focusing the non-fight content on fewer builders, really picking out the storylines, and taking the interviews much more seriously. There’s currently an incredible amount of content that just isn’t engaging. I’d push for cutting everything between the tunnel entrances and the button pushes, then editing the gap between the buttons and the starting light tree to like 5 seconds. The vague, non-committal “strategy” interviews from each team before fights are mostly just an exercise in how many ways we can say “gonna get around them” or “going to avoid that big weapon”. It's not interesting, neither strategically, nor personably. Maybe some of the Faruq intros are long. I think most of them are hilarious, so my bias is to keep them. I think there’s enough time currently being wasted on low-value content that we could replace with more in-depth segments. The most important thing is getting a pit reporter back. 

My medium-term outlook on the show is we need to circle the wagons and focus on the things people watch the show for. As much as Reddit wants it to be 15 fights per episode, I don’t think that would be watchable. There’s no space left for relatability. I hate to call it reality content, but that's really what it is and what it could be. With a smaller field and more concentrated storylines, there’s room to focus on the most engaging content. Between improved interviews and no less than 4 fights per robot, there’s so much room for team brand growth. Since BattleBots can’t carry all of the social media content on their own, the more engaged fans are with teams, the more everyone benefits. Teams can sell more merch and more fans become advocates. 

Fight Night Funniness 

Let me just climb into a jumbo-sized tin foil hat before continuing. Is it possible they gamed the Fight Nights too hard? (Cupcake is the term production uses to describe a matchup that one team should be able to win handily AND make a show out of it. Some robots are perpetual cupcakes.)

Throwing some weird ones out there:

• Witch Doctor (2-1, 24 seed) 2019 runner up, RE:MARS winner
• Duck (0-3) - great test of Witch Doctor’s reliability, something they’ve struggled with.
• End Game (3-0, 1 seed) [loss] - as the 2019 runner up, throw them at the most recent champion, see if they’re still championship material.
• Rusty (0-2) - Did Witch Doctor need a second win? Did Rusty need an opponent who could ensure an entertaining fight? Could be both of these things, definite cupcake.
• Did production see an opportunity to slip a great robot into a low seed, spreading the competitiveness out across the bracket?

• MaDCatTer (2-1, 15 seed) interesting parallel path to Witch Doctor
• Yeti (2-1, 19 seed) - possible mismatch, good test of MC’s reliability, compare to WD vs Duck!
• Sawblaze (2-1, 4 seed) [loss] - hard fight, compare to WD vs EG, both 1-1
• Rampage (0-2) - cupcake, compare to WD vs Rusty
• Valkyrie (2-1, 26 seed) Season 5 most destructive
• P1 (2-1, 20 seed) [loss] - Valkyrie has always struggled with wedges and P1 has ~70lbs of wedge. They were also sharing an AirBnB so it's just kinda funny they fought.
• Triple Crown (0-1) - Experimental robot that seems to have been assessed as too risky to fight anything that won’t tear it to pieces.
• PMF (0-2) - Rookie team with some awful logistics luck and technical difficulties. Seems to also have been assessed as a high risk for dud fights, thus fighting a big kinetic. Cupcake. 
• Seems to be filling the role of industrial shredder

• Ribbot (3-0, 2 seed)
• Defender (2-1, 31 seed play in) - Whiplash off-shoot, new robot, severely mismatched for a first fight
• Overhaul (0-2) - lifter from a veteran team with a 2-year gap.
• P1 (2-1, 20 seed) - veteran lifter, both 2-0
• How the heck does a returning quarter-finalist get 3 non-kinetics?

• HUGE (2-1, 28 seed)
• Riptide (2-1, 21 seed) [loss] - rookie, massive spinner
• Retrograde (1-2) - rookie/Bloodsport off-shoot, lifter can reach the body, undercutter could mess up the wheels
• Switchback (1-2) - rookie, massive drum can reach the body, both 1-1
• A member of the selection committee once explained they like to force robots to prove they can beat the things they used to struggle with. I think they see HUGE as having a particular weakness for robots that are also well outside the meta.

Two examples of robots with good schedules, both ended up with middling seeds.

• Bloodsport (2-1, 17 seed)
• Whiplash (2-1, 3 seed) [loss] - last year’s 2 seed vs last year’s runner up
• SubZero (0-3) - seems horribly mismatched, debateable cupcake
• Claw Viper (1-2) - both 1-1, big wedge
• 1 hard/counter fight, 1 easy fight, 1 matched fight, seems good

• Cobalt (2-1, 14 seed)
• Fusion (1-2) [loss] - fair matchup
• Ghost Raptor (0-2) - mismatch, cupcake fight
• Gruff (1-2) - durability test, both 1-1
• Overall seems to be an ok schedule for a mixed-veteran team/bot