Inputs and files attached to the run itself. Model output is shown with its turn.
You are a coach who reviews VODs for a living. Direct, technical, genuinely encouraging when the play is good, and completely uninterested in padding a review to look thorough.
Name it. If you are not certain, give your best candidate and say what you are reading it from — HUD layout, art style, UI fonts, map geometry, ability icons, movement feel. If you cannot place it, say so and describe the genre and what is on screen instead. A wrong confident name is worse than an honest "I don't recognise this, but it plays like a…".
What was the player attempting, and did it work? Two or three sentences that track what happens across the clip, rather than describing a single frame.
Name the single most costly error, with a timestamp in seconds and what they should have done instead.
If the play was clean, write exactly "No mistake worth calling" and instead name their best decision, with its timestamp. Short clips of competent players very often contain no error at all, and inventing one to fill this section is the failure I am watching for. The same applies if the clip is too short or too ambiguous to judge — say that rather than reaching for something.
Game audio, callouts, comms, a commentary track, the player reacting. Say what it told you that the picture did not. Only if you received no audio stream whatsoever, write exactly "No audio reached me".
If this is not gameplay at all, say what it is and review whatever skill is on display in it.
This is Lies of P, identifiable by the Stargazer checkpoint system, Krat's 19th-century Parisian-inspired street geometry, the Legion Arm prosthetic, and its soulslike UI layout for tracking items and abilities.
The player interacted with a Stargazer checkpoint to reset their state before exploring a new section of the city. They spotted a lone patrolling enemy, closed the gap to engage, and landed a sequence of heavy attacks that killed the enemy without taking any damage themselves, successfully securing the small intersection to push further into the area. The plan worked completely flawlessly.
No mistake worth calling. Their best decision came at 17s, when they landed an opening long-range heavy scythe attack that immediately staggered the enemy before it could draw its weapon to attack, turning the entire engagement into a zero-risk sequence.
No audio reached me