The cost of a perfect boarding queue

This is a research toy, not an operational tool. The aircraft is built on field-validated published work. Everything about the gate, and every frustration number, is an assumption of the model. Below, the two are kept apart; in the repository they are kept apart parameter by parameter.

What the slider changes

One number: how long a gate agent takes to call a single passenger forward. Strict Steffen needs an exact 180-place queue, so in this model it is the only procedure called passenger by passenger. Random is one announcement. Back-to-front is six zone calls, twenty seconds apart. Nothing else about the flight moves — same seed, same manifest, same aircraft.

At four seconds a passenger, calling 180 people cannot finish before eleven minutes fifty-six, and that single figure is most of why Strict Steffen loses the whole journey. Somewhere between two and two and a half seconds it starts winning instead. That crossover is the honest headline of this project, and it is why the slider exists.

Where the numbers come from

Measured · field-validated

The aircraft

The cabin is Michael Schultz's stochastic boarding model: a 0.4 m cell grid with one passenger per aisle cell, 0.8 m/s maximum aisle speed, Weibull hand-luggage stowing (shape 1.7, scale 16 s), seat-interference resolution and a 3.7 s aircraft-door headway. Those values were measured during real A320 and 737 turnarounds, and his model reproduced his measurements to within 5%.

Schultz, M. (2018). Field Trial Measurements to Validate a Stochastic Aircraft Boarding Model. Aerospace 5(1), 27. doi:10.3390/aerospace5010027 — reproducing his reference conditions inside Boarding Lab is Layer 2 of its validation plan and has not been done. The 5% is his result for his model, not a claim about this one.

Assumed · not observed

The gate, and every frustration number

The call rate. No published source gives a real gate-agent rate, because no airline calls passengers individually. Four seconds is a choice, which is exactly why you can change it.
Zone calls twenty seconds apart. Also unobserved.
Being separated from your companions adds 0.25 to your stress load. Strict Steffen splits up 54 of 180 passengers by design. Setting this to zero moves its total burden from 8.90 to 7.47 F·min, so it is real but secondary — the thirteen minutes of queueing is what drives the result.
Every frustration coefficient. Frustration is measured in F·min, frustration-minutes per passenger: half-frustrated for two minutes accumulates 1.0. It is a model prediction under stated assumptions, never a measurement.

Evidence pointing the other way

The one experiment that measured boarding speed and passenger satisfaction together found the fastest method was the least liked. MythBusters boarded a 173-seat mock cabin five ways: free-for-all with no assigned seats was fastest at 14:07 and scored lowest on satisfaction at −5, while Reverse Pyramid took 15:10 and scored highest at 113. This model has no term at all for a process feeling organised, so its preference for loose, unsorted boarding may well be overstated.

MythBusters, "Plane Boarding", first aired 21 August 2014.

There is also a real counter-example to the premise of an expensive call. Southwest Airlines boarded roughly 175 passengers into an exact numbered order for decades — position printed on the boarding pass, numbered stanchions at the gate, no individual calls at all — and retired it only in January 2026 when it moved to assigned seating. Released as grouped calls rather than one at a time, Strict Steffen wins every run in this model.

The rest of the reading

Steffen, J. H., & Hotchkiss, J. (2012). Experimental test of airplane boarding methods. Journal of Air Transport Management 18(1), 64–67. doi:10.1016/j.jairtraman.2011.10.003 — the method, and the times that made it famous. Their clock started when boarding started.
Bachmat, E., Erland, S., Jaehn, F., & Neumann, S. (2023). Air Passenger Preferences: An International Comparison Affects Boarding Theory. Operations Research 71(3), 798–820. doi:10.1287/opre.2021.2148 — 1,500 passengers surveyed: how long you wait predicts satisfaction better than how long the aircraft takes.
Jiang, H., & Ren, X. (2019). Model of passenger behavior choice under flight delay based on dynamic reference point. Journal of Air Transport Management 75, 51–60. doi:10.1016/j.jairtraman.2018.11.008 — why the same delay does not feel the same to everyone.
Schultz, M. (2018). Fast Aircraft Turnaround Enabled by Reliable Passenger Boarding. Aerospace 5(1), 8. doi:10.3390/aerospace5010008
Shen, M., et al. (2026). Group boarding for airplanes. arXiv:2607.21512 — why companions are first-class objects in this model.

What would change my mind

A gate observation showing an ordered queue can be assembled far faster than four seconds a passenger. Move the slider below two and a half seconds and the conclusion inverts on its own.
Passenger self-reports contradicting the frustration ordering. The MythBusters result already points this way.
A measured companion-separation response near zero, which would close about a fifth of the frustration gap.

All of it — the simulator, the 100-run dataset, the parameter provenance registry and the full version of this document — is open source under MIT at github.com/kefalasdion/boarding-lab. If you work at a gate and know how long it really takes to call a boarding zone, that is the number this project is missing.