After a slow decade where Melbourne hardly improved its off-peak Metro train frequencies, the last year has seen some significant action. The February 2026 Metro Tunnel timetable cut maximum waits on the Sunbury - Pakenham/Cranbourne corridor while the August 2026 timetable cut maximum waits on the Craigieburn, Upfield, Mernda, Hurstbridge and Sandringham lines.
The infamous Sunday morning 40 minute gaps are now gone and many more lines have 20 rather than 30 minute evening service. Sunbury, Craigieburn, Upfield and Sandringham were the big winners - their maximum waits fell from 40 to 20 minutes. Pakenham and Cranbourne had maximum waits fall from 30 to 20 minutes.
30 minute gaps remain to Mernda, Eltham, Ringwood, Belgrave, Lilydale, Alamein and Glen Waverley. Some (Clifton Hill group weekend evenings and Burnley group weeknights) will be closed to 20 minutes after all the 2026 budget funded upgrades are implemented. This is likely next year.
After then the remaining unfunded 30 minute gaps include weeknights on the Clifton Hill group, weekend evenings and Sunday mornings on the Burnley group, and midday weekday for the outer part of Belgrave and Lilydale. Outer Hurstbridge will remain every 40 to 60 minutes.
Still we seem to be on a good trajectory where 20 minute maximum waits at all but a handful of stations on the network will be the rule rather than the exception.
Every 20 minutes and then what?
Every 20 minutes is a middling frequency for urban transit. It's not a super long wait but neither is it turn up and go. Passengers still need to use timetables and plan trips, especially if making connections. A cancelled train can mean 40 minutes between services.
Cities such as Sydney and Perth do much better than Melbourne with a 15 minute base frequency the rule. Whereas Melbourne's trains run to a complex mix of 10, 15, 20, 30 and 40 minute frequencies.
At the start of the 1990s the typical Melbourne train line ran every 20 min off-peak, 30 min evening and 40 min on Sunday. The Burnley group (the busiest and most established part of the network) was an exception with Glen Waverley and Ringwood having 15 minute midday weekday frequencies. This was joined by Sandringham in the early 1990s and Dandenong and Frankston in 1996. 1999 saw network wide train and tram upgrades after about 10am on Sundays to match Saturday times.
In the early 2000s, around the time of the Melbourne 2030 plan and public transport '20% of trips by 2020' target, there was interest in the bureaucracy to get the rail network to a Perth-style 15 minute all day frequency. Nothing came of it - government rail priorities were then more regionally based and picking up the pieces from the National Express walk-out. SmartBuses were being implemented with a weekday 15 minute headway but these then only served lines in the east and south-east that were already operating at matching frequencies.
The Bracks/Brumby government's reduced interest in suburban rail came back to bite it as patronage surged and reliability plummeted. Political interest and investment revived but the benefits came too late to save that government from defeat in 2010. Still for a while there was a bipartisan consensus in favour of timetable reform and improved rail service levels. Some got done, notably on the Frankston and Pakenham/Cranbourne lines in the south-east.
A product of this era was the Network Development Plan (Metropolitan Rail) of 2012. This proposed an ambitious service upgrade program, including new timetables that would deliver off-peak services every 10 minutes on main lines and 20 minutes on branch lines. This was meant to be in stages with key lines running every 10 min by 2016 and more by 2022. However interest slowed in favour of Big Build projects from 2015 and by 2026 not even all of the 2016 upgrades had been implemented.
This leaves us to where we are now. The Big Build does not have the social licence and political commitment it did in its peak around 2018. Its higher than expected costs have led to high interest bills and constrained budgets for service delivery. However political will to deliver improved train frequencies is higher than it has been for a decade, with the 2025 and 2026 state budgets demonstrating this.
And we are in an election campaign so there will be various promises (including hopefully on public transport service) made. Even though the spirit of the times favours maintenance and smaller projects rather than megaprojects. 'Cheap and cheerful' might be the new catch cry.
Considerations when choosing a base frequency for rail
To me the core decision when reforming a rail timetable beyond the widespread 20 minute off-peak frequency on many lines (particularly in the north and west) is whether to go for a 10 or 15 minute base frequency.
There are connectivity benefits in having a single harmonised frequency 'family' for the whole network but above a certain point it becomes less important as a higher proportion of passengers treat it as turn up and go.
Especially if groups are operated independently with no interaction there may be no technical reason why some can't be (say) every 7.5 minutes and others every 10 minutes. The length, run times and infrastructure constraints of each group may point to a particular solution. There may also be trade-offs for instance short-workings to allow higher frequencies at busier closer in stations. And/or all day express running - something that Melbourne generally doesn't currently do - may on some lines be better accommodated with a 15 rather than a 10 minute frequency.
Outer suburban residents may actually prefer a 15 minute service that has express running versus a slow 10 minute that does not. But when service levels are lower, eg every 30 minutes, then frequency might assume higher importance. Also inner suburban locations with short trips and high transfer rates might prefer a 5 or 7.5 minute frequency to the 10 minute service that currently operates on the (in my view underserved) Metro Tunnel core.
Another factor is multimodal coordination. Trams are a mix of 6-8, 10 and 12 minutes. Buses are a mix of 10, 15, 20, 30, 40 and 60 minute frequencies, though some routes have oddities like 22-24 minutes. A rail network with a 10 minute base permits a hierarchy of 10, 20, 30, 40 and 60 minute frequencies as opposed to a base 15 minute network that offers a somewhat reduced choice of 15, 30, 45 or 60 minute headway for buses.
Public presentation and passenger information is important. A major part of marketing the network is to highlight which parts of it have frequent service. A 20 minute frequency is substantially less marketable than a 10 or even 15 minute frequency. Cities that have marketed the multimodal frequent parts of their network typically choose a 15 minute or (less commonly) 10 minute cut off. Every 20 minutes doesn't really qualify for frequent, especially in a big city where many trips would involve connections. These customer-facing considerations should be a part of the decision mix just as rail engineering considerations are.
Finally there is driver hours and costs. This is substantial given (a) the length of time it takes to train a train driver and (b) cost, especially where there are budget constraints.
Below is a table showing the number of trains per hour added (or subtracted) when adjusting service levels.
The X axis is existing frequency and the Y axis is proposed frequency. As an example, moving from a 20 to a 10 minute service requires an extra 3 trains per hour. Whereas moving from every 20 to every 15 minutes adds just 1 train per hour.
In other words assuming equal line lengths you can improve service on three times as many lines by going to every 15 minutes than going to every 10 minutes. Provided the infrastructure allows a reliable timetable, there may be cases that going to every 15 minutes may be a reasonable decision now, even if the ultimate goal remains a 10 minute headway.
Or in tighter times there may be opportunities when budget is available to upgrade one line to a 15 minute service but none to a 10 minute service. Should perfect be the enemy of the good in this case? I would suggest 'get what you can when you can'.
Any costing needs to factor in return trips and the hours per week that increased frequency is desirable. Some examples are in the table below.
Careful service additions to create a frequent network over consistent hours all week (eg 7am - 9pm 7 days) can assist passenger legibility and messaging in ways that the current mess, with widely varying frequencies by hour and day type cannot.
No answers, just many questions and trade-offs
(click for better view)
Something like every 10 or 15 minutes should be considered for suburbia with 5 or 7.5 minute frequencies more suited for dense core areas. This would bring our service levels up to what other cities like Sydney and Perth run on their systems.
However deciding between the two, as well as other considerations like suburban expresses and inner short-workings, have trade-offs, only some of which I've covered.
Once settled service considerations should drive infrastructure priorities, and not the reverse as was common during the big spending Big Build era (which often resulted in achievable service upgrades not happening or at best being scaled back or deferred).
What do you think is the most suitable frequency and operating patterns for each group of lines? And how important is it that it be the same base frequency network-wide?
Your comments would be welcome and can be left below.
See Timetable Tuesday items here
