Hogan's Alley

What is Hogan's Alley?

Hogan's Alley is a reaction-time and marksmanship trainer modelled on the FBI Academy facility of the same name, running entirely in your browser. It is built around the two things a training tool is actually judged on: whether the aiming model behaves like a human holding a gun, and whether the numbers it reports are measurements rather than decoration. The graphics are deliberately plain — the effort went into the simulation.

Timing you can trust

Reaction times are easy to report and hard to report honestly. Targets are timestamped from the compositor — when the photons actually reach the screen — not when the code decided to draw them. Trigger timestamps come from the OS, ahead of any JavaScript scheduling delay, so a busy frame cannot inflate a result. Start delays are drawn from a memoryless distribution, because a uniform delay is timeable and shooters learn to time it. Responses under 100ms are logged as false starts and excluded rather than counted as brilliant ones.

Your display's refresh rate and the resulting quantisation floor are reported alongside your results. On a 60Hz panel you cannot resolve better than about 8ms, and the app says so rather than implying a precision it does not have.

A hold that behaves like a hold

The wobble zone is synthesised from the components of a real hold rather than a single noise function: mean-reverting postural drift, breathing you can suppress at a cost, physiological tremor, and a cardiac pulse tied to a modelled heart rate that rises with arousal. Recoil is an angular impulse into a spring–damper, so muzzle rise and settle time emerge from the physics — which is what puts a real floor under your split times. Scoring zones are defined in metres on the anatomy and projected at true distance, so a hard shot is hard for the same reason it is hard on a range.

Shoot / don't-shoot

Half of what this measures is discrimination: pistol or phone, badge or knife. The figures are photographed rather than abstract, because a grey rectangle in a stick figure's hands poses an easier question than a real hand does.

Appearance deliberately carries no information about threat. Every adult in the cast appears with the complete set of objects — the same person shows up holding a pistol and holding a wallet — so there is no correlation between how someone looks and whether they are armed. Learning such a correlation would be training a bias, which is the precise opposite of the point. Only the object changes between images of a figure, so when a target's hands change mid-exposure your reaction time measures the object changing, not the person being swapped. Youth figures are generated with benign objects only; the app never renders an armed child.

What it tells you afterwards

• Called shot vs actual impact — a scattered impact group with a tight call group is a trigger problem, a scattered call group is a hold problem, and they need opposite corrections
• Splits and transitions reported separately, so averaging cannot hide which one you are slow at
• d′ and criterion from signal detection theory — separating whether you can tell a gun from a phone from how much evidence you demand before firing
• Group statistics in centimetres at true distance: centroid offset, mean radius, extreme spread
• CSV export with the seed, refresh rate, and timing conditions in the header, so a run is reproducible

Drills

• Reaction Lab — simple reaction, and shoot / don't-shoot choice reaction
• Standards — Bill Drill, Mozambique, El Presidente, 25yd precision, Hostage, Transitions
• Judgment — Hogan's Alley, and a degraded low-light variant
• Qualification — an FBI-pattern qualification course

Every scenario is seeded and randomised — which figure is armed, where it appears, and when. A fixed layout is memorised after two runs and stops measuring anything. You can re-run a seed from the results screen to compare against yourself under identical conditions.

Shooter profiles from Novice to Expert scale sway, flinch and recoil recovery so an amateur has a simulated handicap to work against. An Instrumented profile zeroes the simulated shooter entirely, measuring your own reaction time and pointing accuracy with no handicap invented on top.

Requirements

Hogan's Alley needs a desktop or laptop with a mouse. Set your cm/360 and DPI in Settings before your first run — everything is scaled off real mouse travel, so a wrong sensitivity makes the pointing measurements meaningless. The app uses pointer lock, so you'll click once to capture the mouse and press Esc to release it. A high refresh rate display materially improves what the timing can resolve.

Privacy

Hogan's Alley runs in your browser with no account and no sign-up. Your settings and run history stay on your own device, and exported CSV files are saved wherever you choose. For full details, see our Privacy Policy.