• Home
  • Education
  • Sport
  • Aviation
  • Publications
  • CV
  • About me
  • Participate

What if the fastest way to board a plane is one airlines barely use?

Aviation
Eleven boarding strategies, three operating environments and 66,000 simulations — and the fastest method turns out to depend on how many doors the aircraft uses.
Published

August 24, 2026

66,000 final boarding simulations

What if the fastest way to board a plane is one airlines barely use?

Eleven boarding strategies. Three operating environments. And one surprisingly complicated queue inside a very narrow tube.
Antonin Kremenak · Analysts Diary · 24 August 2026
Why study boarding?

An aircraft earns money by flying — not by waiting at the gate.

Every commercial aircraft spends part of its day on the ground. Passengers need to leave, the cabin must be prepared and the next group has to board before the aircraft can depart again. Airlines therefore try to keep turnaround times predictable and as short as operationally possible.

A few minutes may not sound important on a single flight, but airlines repeat the same process hundreds or thousands of times across their networks. Time saved during boarding can contribute to faster turnarounds, better aircraft utilisation and more resilient schedules. Conversely, slow or unpredictable boarding can consume part of the available turnaround buffer before the aircraft has even pushed back.

The important point: boarding is not only a passenger-experience problem. It is also an operations problem. The faster and more predictably an aircraft can be prepared for departure, the sooner it can return to the activity that generates revenue: flying.
Final simulations
66,000
Boarding strategies
11
Operating scenarios
3
Three operating environments

The same aircraft, tested under three different conditions.

Single door

One front entrance, full aircraft, baseline luggage.

Passengers follow the assigned boarding order perfectly and no travel groups disrupt the sequence. This is the clean baseline for seeing what a strategy can do with one conventional front door.

Dual door

Front and rear entrances, full aircraft, baseline luggage.

The passengers are split between two real streams entering from opposite ends of the same aisle. Compliance remains perfect and there are no travel groups. The second door changes the geometry - and can change the winner.

Operational stress

One front entrance, more luggage, 75% compliance and travel groups.

This scenario adds the everyday friction that a neat boarding plan cannot fully control. One quarter of passengers travel in groups and only 75% keep their prescribed position in the boarding order.

Meet the contenders

Eleven ways to fill the same aircraft.

Some methods are beautifully simple. Others look almost absurd when written as a boarding queue. But each one is trying to solve the same problem: how do you stop passengers from getting in one another's way?

CRBF — Column Rotated Back-to-Front
Board one seat column at a time, from the rear of the aircraft toward the front.
CRBF stands for Column Rotated Back-to-Front. Instead of sending an entire rear section of passengers into the cabin at once, the aircraft is boarded column by column. In the 3-3 cabin used here, window, middle and aisle columns are called separately and each column proceeds from the rear toward the front.

The idea is to reduce seat interference while spreading passengers along the aisle. That allows several passengers to store luggage at the same time rather than creating one large queue around the same few rows.
Steffen method
Deliberately space passengers several rows apart so they can stow luggage in parallel.
The Steffen method uses a highly structured passenger sequence. Consecutive passengers in the boarding line are usually seated several rows apart, and window-seat passengers are handled before passengers closer to the aisle.

The unusual order has one main purpose: parallelism. Instead of one passenger blocking everyone behind while storing a bag, multiple passengers should be able to use different overhead bins simultaneously.
Steffen-Lug
Steffen boarding, but passengers with more luggage move earlier inside each boarding wave.
Steffen-Lug keeps the spatial logic of the Steffen method but adds information about carry-on luggage. Inside each Steffen boarding wave, passengers carrying more bags are placed earlier in the sequence.

The goal is to let luggage-heavy passengers use relatively empty overhead bins before those bins become crowded. In the simulation this modification became especially interesting when two aircraft doors were available.
WilMA — Window, Middle, Aisle
Window passengers first, middle seats second and aisle passengers last.
WilMA stands for Window, Middle and Aisle. Its logic is easy to understand: board passengers sitting next to the windows first, then passengers in middle seats, and finally passengers sitting beside the aisle.

This largely removes one common source of delay: a window passenger arriving after the middle and aisle passengers are already seated and forcing them to stand up again.
Reverse Pyramid
Combine seat position and distance from the door into diagonal boarding zones.
Reverse Pyramid does not simply divide the aircraft into front and rear blocks. Instead, boarding zones combine seat type and row position. Rear window passengers tend to board earlier, followed by overlapping combinations of windows, middle seats and eventually aisle seats.

The original published cabin configuration differs from the 26-row all-economy aircraft used in this project, so the zone boundaries were proportionally adapted. That is why the strategy is labelled Reverse Pyramid (adapted).
Random boarding
No spatial boarding logic — passengers enter in random order.
Random boarding simply shuffles the passenger order. It does not deliberately prioritise rows or seat positions.

That sounds inefficient, but randomisation naturally spreads people around the cabin. It is therefore a useful benchmark and can sometimes outperform boarding methods that create very concentrated congestion.
Back-to-Front
Call passengers in rear zones first and gradually move toward the front.
Back-to-Front is probably the most intuitive boarding method. Passengers in the rear section are called first, followed by successively more forward zones.

The problem is that passengers from the same boarding group often need the same part of the aisle at the same time. The queue may look organised at the gate while producing a dense moving bottleneck inside the aircraft.
Lufthansa-like
Priority boarding combined with more explicit seat-position logic.
The Lufthansa-like method first handles selected priority groups and then applies seat-position logic to the remaining economy passengers, with window passengers preceding middle and aisle passengers.

Among the simplified airline-style approaches, this creates more spatial organisation inside the cabin than a pure priority-first system.
Ryanair-like
A simplified priority-first boarding structure inspired by low-cost airline operations.
The Ryanair-like model separates priority passengers from the standard boarding population. Priority passengers enter first and the remaining passengers follow.

It is intentionally a simplified structural approximation. It is not intended to reproduce every detail of Ryanair's actual gate procedures, airport infrastructure or commercial priority product.
easyJet-like
Families and priority passengers board ahead of the remaining passengers.
The easyJet-like structure gives early boarding access to families and priority passengers before the standard passenger population enters.

Again, the purpose is not to recreate an airline's full real operation. It provides a simplified priority-based comparator for the strongly spatial academic strategies.
US Network-like
Several successive priority groups followed by standard boarding groups.
The US Network-like model represents the multi-group boarding structure common in large network-airline environments. Families and higher-priority passengers are processed before several standard boarding groups.

The model deliberately simplifies the much more complicated real systems involving cabin class, loyalty status, credit cards, accessibility and ticket products.
See the difference

Watch the cabin fill.

Choose a boarding strategy and operating scenario. The animation shows one representative run from the same simulation engine used for the final experiment.

Elapsed
0 min 00 sec
Seated
0 / 156
Walking
0
Stowing bags
0
Representative run
—
Walking Stowing Taking seat Seated
The scoreboard

There was no universal winner.

The fastest method depends on the environment. Change the door configuration or passenger behaviour and the ranking can change dramatically.

The bigger picture

What 66,000 final simulations revealed.

The interactive ranking above shows the headline result. The charts below come from the technical version of the study and show how the pattern changes across scenarios — and why.

Single-door boarding

Single-door boarding

Mean boarding times from 2,000 Monte Carlo runs per strategy. Lower is better.

Opening the rear door changes the ranking

Opening the rear door changes the ranking

The fastest single-door strategy does not remain the fastest when the aircraft is boarded from both ends.

What happens when operations get messy?

What happens when operations get messy?

Higher luggage volume, imperfect compliance and travel groups slow every method, but not by the same amount.

The winner depends on the scenario

The winner depends on the scenario

The same boarding strategy can move substantially up or down the ranking when infrastructure and passenger behaviour change.

Why aisle blocking matters

Why aisle blocking matters

Strategies that create more blocking also tend to produce longer total boarding times.

Academic optimisation vs airline-like boarding

Academic optimisation vs airline-like boarding

The comparison isolates boarding-flow efficiency. Real airlines must optimise many other commercial and operational objectives.

How the model evolved

This was an iterative model, not a one-script result.

The audit found 18 unique executable versions across 19 named simulation files. The two V3.0 filenames are byte-identical aliases, so they count as one version.

Unique versions
18
Simulation rows across development
523,460
Final experiment
66,000

The model moved from baseline aisle mechanics to experimental replication, movement and luggage sensitivity, behavioural compliance, travel groups, dual-door infrastructure and finally the 11-policy comparison. The total is based on the actual All Runs rows saved in the result workbooks, including the final V3.1.2 smoke test.

Why?

The cabin rewards parallel work.

One door: spread passengers out.

With a single front entrance, CRBF performs extremely well because passengers are distributed through the cabin and several people can store luggage at the same time.

Two doors change the geometry.

When half the cabin enters from the rear, the flow pattern changes. Steffen-Lug becomes the fastest method in the final simulation.

Luggage is not a small detail.

A passenger placing a bag in an overhead bin may temporarily become a moving roadblock. Spread those passengers apart and the cabin can process several bags simultaneously.

Back-to-Front has an intuitive trap.

It organises passengers neatly outside the aircraft but then sends many of them into the same cabin area. The result can be a long, slow-moving bottleneck.

Turn minutes into money

What could faster boarding be worth?

Change the assumptions yourself. Choose any Strategy A and any Strategy B, the number of daily flights and an illustrative value for one minute of gate time. All 11 × 11 directed combinations are available, including A = B.

Time saved by Strategy A per flight
—
Illustrative value per flight
—
Illustrative annual value
—
Illustrative operating value, not profit. This is a gate-time extrapolation, not a prediction of real airline profit. It does not include schedule redesign, crew costs, network recovery, missed connections, commercial boarding products or aircraft utilisation effects outside the boarding model.
A fair question

If these methods are faster, why don't airlines just use them?

Because airlines are solving a much bigger problem than this simulation.

A mathematically efficient boarding sequence may require exact passenger ordering, stronger gate control or separating people who want to board together. Airlines also need to accommodate families, premium products, accessibility, loyalty status and passenger expectations.

So the simulation should not be read as “airlines are doing boarding wrong.” It asks a narrower question: if we isolate passenger flow inside the cabin, what boarding structures reduce congestion most effectively?

Behind the dots

A little more detail, if you want it.

What aircraft was simulated?

A 26-row narrow-body aircraft with a 3-3 seating configuration: 156 passengers at full load.

How many times was each method tested?

Every strategy was simulated 2,000 times in each of three operating scenarios. 11 strategies × 3 scenarios × 2,000 trials = 66,000 final simulations.

Why simulate the same thing thousands of times?

Because passengers are not identical. Walking speeds, luggage and other stochastic elements vary. Monte Carlo simulation lets the study measure not only the average result but also how reliable each method is.

Is the aircraft animation one of the 2,000-trial averages?

No. An average cannot literally walk down an aisle. The animation shows one representative deterministic run so the mechanism is visible. The numbers reported in the charts come from the full Monte Carlo experiment.

Are Ryanair-like and Lufthansa-like exact airline procedures?

No. They are simplified structural approximations used for comparison. Real airline boarding procedures contain many additional operational and commercial rules.

Where are the statistical tests and validation?

They are deliberately kept out of this general-audience story. The technical study contains paired comparisons, confidence intervals, effect sizes, P95 analysis, validation and detailed assumptions.

One important caveat

This is a model, not a departure board.

The simulation isolates passenger boarding inside a representative narrow-body cabin. It does not model every part of airport operations, the complete turnaround process or an airline network.

The results are therefore best understood as a controlled comparison of boarding mechanisms rather than a prediction that every real flight would achieve exactly the same time.

Learn more

The research behind the model.

These are selected sources from the supplied literature folder. They cover the Steffen method, experimental tests, Reverse Pyramid, robustness, luggage handling and recent Airbus A320 simulation work.

Steffen (2008) — Optimal boarding method for airline passengers

Steffen & Hotchkiss (2012) — Experimental test of airplane boarding methods

van den Briel et al. (2005) — America West Airlines develops efficient boarding strategies

Ferrari & Nagel (2005) — Robustness of efficient passenger boarding strategies

Qiang, Jia & Huang (2017) — Surrogate experimental test

Coppens et al. (2018) — Review, field study and luggage-stowing experiment

Moreira et al. (2023) — Airbus A320 discrete-event simulation

Want the technical version?

The methodology, validation and statistics are all there.

The professional study contains the full simulation development, assumptions, paired statistical analysis, confidence intervals, P95 results, economic methodology and limitations.

Read the technical study →
© 2026 Antonin Kremenak
Analysts Diary

© 2026 Analyst’s diary

Spotify · Apple Podcasts · Instagram · LinkedIn

Built with Quarto