What is Headway? Understanding Train Frequency

The headway debate in urban rail is increasingly not about signalling — CBTC has largely solved the signalling constraint at 90 seconds — but about dwell time and platform throughput.

November 26, 2025 7:37 am | Last Update: March 20, 2026 9:49 am

⚡ In Brief

Headway is the time interval between consecutive trains passing the same point — the inverse of frequency and the primary measure of railway line capacity.
The theoretical minimum headway of any rail line is determined by the longest of three constraints: signalling separation, dwell time at the busiest station, and turnback time at terminals.
The world’s minimum operational headway is approximately 85–90 seconds, achieved on CBTC-equipped fully automated metro lines such as Paris Métro Line 14 and Dubai Metro.
Halving headway doubles line capacity without building new infrastructure — making headway reduction the most cost-effective capacity investment available to congested urban rail systems.
Passenger throughput at a single station platform is the binding constraint on minimum headway for high-demand metro systems — not signalling — once CBTC is installed, meaning wider doors, longer trains, and platform screen doors often deliver more capacity than further signalling upgrades.

Tokyo’s Yamanote Line carries 3.8 million passengers per day in a loop around the city — more than the entire annual ridership of many national railways. At peak hour, trains run every 2 minutes 30 seconds, a headway so short that as one train pulls out of Shinjuku station, the next is already visible entering the other end of the platform. The entire operation depends on a precise orchestration of signalling, dwell time management, and driver behaviour — any deviation from the timetable ripples through the entire loop in minutes.

Headway is the heartbeat of a railway line. Everything about how a railway operates — its capacity, its punctuality resilience, its fleet requirement, its operating cost — flows from this single number. Understanding headway is understanding how railway capacity works.

What Is Headway?

Headway is the time interval between the front of one train and the front of the following train, measured as they pass the same reference point. If a train passes a station platform entry at 08:00:00 and the next train passes the same point at 08:02:30, the headway is 2 minutes 30 seconds.

Headway and frequency are reciprocals of each other:

Frequency (trains/hour) = 60 minutes ÷ Headway (minutes)
Headway (minutes) = 60 minutes ÷ Frequency (trains/hour)

A 2-minute headway gives 30 trains per hour. A 5-minute headway gives 12 trains per hour. A 90-second headway gives 40 trains per hour — the upper limit of what current signalling technology can achieve reliably on a high-capacity metro.

Headway vs Frequency: The Passenger Perspective

Headway	Trains per Hour	Average Wait (random arrival)	Passenger Experience
90 seconds	40	45 seconds	“Turn up and go” — no timetable needed
2 minutes	30	1 minute	Excellent — equivalent to driverless metro standard
5 minutes	12	2.5 minutes	Good — passengers begin checking timetables
10 minutes	6	5 minutes	Acceptable — timetable-dependent behaviour
20 minutes	3	10 minutes	Poor — significant modal shift to car at this level
60 minutes	1	30 minutes	Very poor — service only viable for captive riders

The Three Constraints on Minimum Headway

The minimum achievable headway on any railway line is determined by whichever of these three constraints is longest. All three must be satisfied simultaneously for safe, reliable operation:

1. Signalling Separation

The signalling system must ensure a safe separation between consecutive trains at all times. The safety distance is determined by the braking distance of the following train — if the lead train stops suddenly, the following train must be able to stop before reaching it.

In fixed-block signalling, trains cannot enter a block section until the previous train has cleared it entirely, plus a safety overlap. This creates a minimum headway of 3–5 minutes depending on block length and speed.

In CBTC moving block, the safety distance is calculated in real time based on the following train’s speed and braking capability. At low speed (near a station), the safety distance is short — perhaps 50–100 metres. The minimum signalling headway under CBTC is typically 60–75 seconds at normal metro speeds.

2. Dwell Time at the Critical Station

A train cannot enter a station until the previous train has departed. The minimum headway through a station therefore equals the dwell time (time doors are open + time to close doors and depart) plus the time for the next train to approach and stop.

On high-demand metro lines, dwell time is often the binding constraint — not signalling. If 2,000 passengers need to board and alight at a major interchange station and the doors are open for 45 seconds, the minimum headway through that station cannot fall below 45 seconds plus approach time, regardless of how good the signalling system is.

This is why increasing door width, deploying platform staff to manage passenger flow, and using wider trains (larger cross-section rolling stock) can increase line capacity more effectively than signalling upgrades on already-CBTC-equipped lines.

3. Turnback Time at Terminals

At the end of a metro or commuter line, trains must reverse direction. The time required for this — unloading passengers, moving the driver to the other cab, loading new passengers, and departing — sets a minimum headway for the entire line. If turnback takes 4 minutes and trains run at 2-minute headway, the terminal needs two platforms (sidings) to absorb trains while the previous train is still in the process of turning back.

Complex terminus track layouts — with multiple platform roads and crossovers — allow shorter effective turnback times by overlapping the movements of consecutive trains. Simplifying terminus track layouts (to save construction cost) often constrains minimum headway more than the signalling system does.

Typical Headways by System Type

System Type	Peak Headway	Off-Peak Headway	Limiting Factor	Example
Fully automated metro (GoA4)	85–100 sec	3–5 min	Dwell time	Paris Line 14, Dubai Metro
CBTC metro (GoA2)	90–120 sec	3–5 min	Dwell + signalling	London Victoria, Hong Kong MTR
Fixed-block metro	2.5–4 min	5–10 min	Signalling block length	Legacy metro systems
Tram / LRT (street-running)	3–6 min	8–15 min	Traffic signals, junction priority	Amsterdam, Vienna trams
Suburban / S-Bahn	5–10 min	15–30 min	Shared track, terminal capacity	Munich S-Bahn, Berlin S-Bahn
Mainline intercity	15–30 min	60–120 min	Journey time, demand, path availability	London–Birmingham, Paris–Lyon
High-speed rail (HSR)	3–8 min	15–60 min	Safety distance at speed, demand	Tokyo–Osaka Shinkansen, LGV

Headway and Line Capacity: The Calculation

A line’s maximum passenger throughput (capacity) is the product of train frequency and train passenger capacity:

Capacity (pax/hour) = (3,600 seconds ÷ Headway in seconds) × Train capacity (passengers)

For a metro with 6-car trains carrying 1,200 passengers, running at different headways:

Headway	Trains/hour	Capacity (pax/hour/direction)	Equivalent to
5 minutes	12	14,400	~7 motorway lanes of car traffic
2.5 minutes	24	28,800	~14 motorway lanes
90 seconds	40	48,000	~24 motorway lanes — more than any urban motorway

This comparison illustrates why metro systems are so space-efficient in cities: a single metro line at 90-second headways can carry as many people per hour as a 24-lane motorway, in a tunnel cross-section of perhaps 30 square metres.

Dwell Time: The Hidden Headway Bottleneck

On heavily loaded metro systems, dwell time — not signalling — is typically the constraint on minimum headway. Dwell time has several components:

Component	Typical Duration	Management Strategy
Door open time (passenger flow)	25–60 seconds (peak)	Wider doors, more doors per car, platform staff
Door closing time	3–5 seconds	Faster door mechanisms, obstacle detection
Safety check / departure signal	2–5 seconds	Platform screen doors, automated door-clear detection
Approach time (next train)	15–30 seconds	CBTC, precise train positioning, ATO

Research on the London Underground, Tokyo Metro, and Paris Métro consistently shows that the fastest feasible dwell times on heavily loaded platforms are 30–45 seconds — setting a practical minimum headway of 75–90 seconds even with perfect signalling. This is why the world’s minimum operational headway has remained around 85–90 seconds for the past decade, despite continued improvements in CBTC technology.

Headway Instability: Why Trains Bunch

One of the fundamental challenges of high-frequency service is headway instability — the tendency of short headways to degrade into irregular “bunching,” where trains run close together in groups with large gaps between groups.

The mechanism is self-reinforcing: if a train runs slightly late, it arrives at a station with more passengers waiting than expected (because they had more time to accumulate). More passengers means longer dwell time. Longer dwell time makes the train later. Meanwhile, the train behind runs slightly ahead of schedule — arriving with fewer passengers, dwelling for less time, and catching up to the late train. The result is a pair of trains running close together, with the train behind them rapidly catching up to form a trio.

Strategies to combat bunching include:

ATS-based regulation: The Automatic Train Supervision system detects developing bunching and instructs trains ahead to hold at stations briefly, restoring even spacing.
Holding points: Designated locations where trains can be held for 30–60 seconds without disrupting overall journey time — typically at intermediate stations with good passenger exchange.
Speed profile optimisation: ATO systems can adjust running speed between stations to maintain even spacing without stopping.
Skip-stop operation: On severely disrupted services, trains may skip lightly-used intermediate stations to restore headway regularity — typically undesirable but effective in recovery.

Headway and Fleet Size

The headway directly determines how many trains a line requires. For a line of given end-to-end journey time, the fleet requirement is:

Fleet required = (Round trip time ÷ Headway) + spare trains

For a metro line with a 40-minute end-to-end journey time (80-minute round trip including turnback):

At 5-minute headway: 80 ÷ 5 = 16 trains in service + spares
At 2-minute headway: 80 ÷ 2 = 40 trains in service + spares
At 90-second headway: 80 ÷ 1.5 = 53 trains in service + spares

Halving the headway more than doubles the fleet requirement because the same number of trains need to cover the route more frequently. This is why high-frequency metro operations require large fleets and represent very significant capital investment — each additional train in service typically costs €3–10 million.

Editor’s Analysis

The headway debate in urban rail is increasingly not about signalling — CBTC has largely solved the signalling constraint at 90 seconds — but about dwell time and platform throughput. The busiest metro lines in the world are approaching the physical limit of how fast passengers can board and alight through a fixed number of doors at fixed platform locations. The responses being explored are interesting: wider rolling stock (the Paris Grand Paris Express RER NG uses 2.8m wide trains, wider than standard metro), additional doors per car length, real-time passenger flow management using cameras and AI to optimise door opening timing, and platform design that directs passengers to less-crowded door positions. Some networks are exploring the elimination of seats on peak-hour trains entirely — increasing capacity by 30% on the same rolling stock. The deeper question is whether the 90-second headway barrier is permanent or whether a new generation of train control technology could push practical minimums to 60 seconds. The mathematics say yes — 60-second headway is technically achievable with moving block. The dwell time constraint says no — you cannot physically move 2,000 passengers through six doors in under 60 seconds. The bottleneck has moved from the signalling room to the platform edge, and that is a harder problem to solve with technology alone. — Railway News Editorial

Frequently Asked Questions

Q: What is the difference between headway and interval?: In common usage, headway and interval are often used interchangeably. Technically, headway is the time between the fronts of consecutive trains passing a fixed point, while interval can refer to the scheduled time between departures from a specific station. On a line with consistent journey times and no intermediate branches, the two are equivalent. On complex networks where trains have different stopping patterns or destinations, the scheduled departure interval at a station may differ from the actual headway on the track between stations.
Q: Why do some metro lines run shorter trains at off-peak times instead of longer headways?: This is a capacity management trade-off. Maintaining short headways (high frequency) but with shorter trains keeps wait times low for passengers — a key factor in attracting discretionary riders. Extending headways to save operational cost (fewer trains per hour) increases passenger wait times disproportionately and can trigger modal shift to cars. Many operators find it preferable to run shorter trains more frequently than longer trains less frequently, even if total seat capacity is similar, because frequency is the primary driver of passenger satisfaction and ridership. On fully automated lines (GoA4), running shorter trains more frequently has essentially zero additional operating cost, making it the clear preference.
Q: What is the “crush capacity” headway and how does it relate to design headway?: The design headway is the minimum headway the line is engineered to achieve reliably under normal operating conditions. The crush capacity headway is the minimum headway achievable when every constraint is pushed to its absolute limit — maximum permitted dwell times, minimum signalling separations, no margin for recovery from incidents. The crush capacity headway is typically 10–20% shorter than the design headway. Operating at crush capacity is inherently unstable — any perturbation causes headway instability and bunching — and most operators target 80–90% of crush capacity as their peak service level, preserving some recovery margin.
Q: How does headway affect punctuality?: Very short headways are inherently more vulnerable to punctuality disruption than longer headways, for two reasons. First, the recovery capacity of a short-headway system is limited — if a train is delayed by 2 minutes on a 2-minute headway service, every train behind it is immediately affected. On a 15-minute headway service, a 2-minute delay is easily absorbed. Second, short-headway systems have less slack in their timetable — the time buffers that allow small delays to be absorbed before propagating. This is why automated metro systems (GoA3/4) have a significant punctuality advantage over manually-driven high-frequency services: automation eliminates driver-induced variability in dwell time and departure behaviour, which is the primary source of small delays that compound into larger disruption.
Q: What headway do the world’s busiest rail lines achieve?: The Tokyo Yamanote Line achieves 2-minute 30-second headways during peak hour on a 35 km loop, carrying 3.8 million passengers per day. The Paris Métro Line 13 runs at 85-second headways during peak hour using CBTC. The Dubai Metro Red Line (fully automated, GoA4) operates at 90-second headways on a 52 km line with no drivers on board. In terms of train frequency rather than passenger volume, the Paris Line 14 and Dubai Metro are among the highest-frequency heavy rail operations in the world. On light rail, some systems in Germany and Netherlands achieve 2.5-minute headways with tram-train operations at major interchanges.