Robot fighting is finally becoming inconvenient in the best possible way. There is enough of it now that “robots punched each other” is not an article. It is a calendar entry.
That changes the job.
Track the league, not just the clip.
A useful watch desk records event date, organization, robot model, rules, control mode, operator involvement, bout length, stoppages, falls, recoveries and source video. If the same platform appears again, the previous match becomes baseline data instead of forgotten entertainment.
Separate teleoperation from autonomy.
Human skill inside the loop is not cheating. It is a different system. Teleoperation tests communications, interface design, latency tolerance and the operator-machine control partnership. Autonomous combat tests perception, policy, prediction and control without the same live human command path. Both are worth studying. Mixing the labels destroys the comparison.
Watch what happens after contact.
Strikes are visually obvious. Recovery is more informative. Does the robot remain inside its support region? How much does the torso pitch? Does the trailing foot recover? Can the controller convert a bad collision into a step rather than a fall? How long until useful offense or defense returns?
The winner of a machine fight may eventually be the system that wastes the least time becoming a machine again after physics interrupts the plan.
The repair bay belongs in the statistics.
Between-round intervention is part of platform performance. Damaged covers, loose fasteners, overheated actuators, communications faults, battery swaps and calibration problems all expose design decisions. A spectacular robot that requires forty minutes of surgery after every three-minute bout is telling you something about maintainability.
Film everything.
The public technical value of robot sports depends on footage. Wide shots preserve footwork and ring geometry. Close shots preserve contact and damage. Replays expose timing. Ideally organizers will eventually publish telemetry and event logs, but stable video archives are already enough to build a comparative record if the source and control mode are preserved.
Robot fighting deserves a weekly watch because repetition turns novelty into evidence. The sport becomes technically interesting when we can ask not merely who won, but what changed between appearances and which failure modes are disappearing.