UN 199: Should we keep quiet Night Network bus routes?

Ask yourself this. 

At what point do you say that a bus route has so few passengers that you stop running it? 

a. Usage of one-quarter the network average? 

b. Usage of one-tenth the network average (including many empty trips)? 

c. Or you keep it running on principle that buses are a public rather than a commercial service? 

If you answered c. would you change your answer if the bus hours you saved was put towards a 'greater good' improvement that carried 5 or 10 times as many passengers in the same area that the quiet route was cut from?   

The above is not an academic question, however much new service funding you wish would come.

We live under a state government that has invested heavily in major infrastructure but has presided over stagnant or declining service per capita on most metropolitan public transport modes.

Notwithstanding its privileged institutional position as adviser to government, DTP has a record of writing business cases that don't win funding. GAIC (ie limited term growth area bus funding issued under tight conditions) has been almost its only bright spots in 2023 and 2024.

DTP must get the hint. Firstly it needs to get better at writing submissions this government likes. Secondly it must more aggressively find ways to cut 'fat' to internally fund high priority service improvements so more happen during this period of relative service parsimony.

The government for its part needs to be less risk averse and more open to reform ideas from DTP, including more easily making available the (very small) one-off capital funds needed each time a route gets a new timetable or minor alignment change. Some comfort for the government could come if the department streamlined its slow internal processes so that if a genuine problem did arise with a new network then it could be attended to without too much political damage.   

A recent success

The very successful Manningham area bus timetable reforms of January 2025 can be held up as a recent model for what I was talking about above. It just needs to be done on a bigger scale in more areas.  

The 280/282 or 'Manningham Mover' was a route put in by the Brumby government when it was feeling the heat on not providing Doncaster Rail but at that stage was unwilling to invest seriously in more rail-like frequent and direct buses. It involved the removal of some other bus routes to fund, introduced wasteful overlaps of its own and was never a patronage success. 

Neither did it help the Brumby government's political problems regarding transport in the area. So a few months before the 2010 state election the government had another go with Manningham buses, this time introducing the direct and frequent DART routes 905 906 907 and 908.  

While the DART routes succeeded, usage on the still remaining Manningham Mover continued to languish at around 6 boardings per hour - well below average for a Melbourne bus route. And there were higher bus needs in the area. So in January 2025 its weekday timetable was cut from every 30 to every 60 minutes. Because 280/282 are long routes, that freed up a lot of service kilometres per week.

That got recycled to boost four busier routes with between 2 and 7 times the boardings per hour. The medium usage routes 284 and 285 gained new Sunday service. The high usage 905 joined the exclusive club of buses running every 15 minutes or better 7 days. And the even busier Route 907 got its existing 15 minute weekend service extended over  more hours. There were also some small consistency changes involving routes 207 and 309. Overall there were a lot of wins with few complaints about the reduced 280 / 282 timetable as its use was so low. 

More examples

The 280/282 aren't the only bus routes understood to be very quiet. There are some short weekday shopper style routes that also don't carry many. But their running costs are low and little is saved by cutting them. And they can be surprisingly resilient, like the (then) DoT's abandoned attempt to delete the 706 between Mordialloc and Chelsea.  

Another group of routes potentially worth looking at is our dedicated Night Network routes. That is special routes that only operate 1-5am Saturday and Sunday. They are typically longer than short shopper routes. And, because they operate on weekends at unsociable hours, their driver costs are likely higher, especially relative to passenger loadings.   

Tracing their history back to the NightRider buses from the late '90s, there used to be a lot more of these, especially when Night Network (involving trains and trams) started in 2016. However reforms in 2021 saw most replaced with extra trips on 21 regular routes. That made travel simpler for passengers and gave some bus served areas 24 hour weekend service. An added benefit for weekend morning travellers is, Good Friday and Christmas Day excepted, they resolve the late starts that some of these routes previously had. 

The remaining thirteen dedicated Night Network only routes are numbered 941 - 982. Most are listed on PTV's website here (967 is missing) and dotted on local area maps. These are the ones that I'll talk about today.  

Options for quiet Night Network bus routes

Three years ago I looked at whether those thirteen routes could be folded into improved services on existing regular routes. That would improve legibility but there would be a cost involved. Not least because you'd be wanting to add 9pm - midnight and 5am-7am service to deliver true 24 hour weekend service similar to the 21 regular routes already operating. 

Another approach (with a smaller budget) is to consider an area's existing service levels and consider whether Night Network services are the best use of bus resources. Especially if (i) their usage is very low and (ii) they ply streets that don't get a proper daytime service. 

The latter is more common than you might think. For example, over 70 routes don't run during the day on Sundays, with some not operating Saturday afternoons.  There are busy streets in Melbourne that have buses at 2am on a Sunday but not 2pm on a Sunday. Nor even 2pm on a Saturday (eg Jacksons Rd Noble Park North).  There's also parts of main roads where service at 2am on weekends is more frequent than at any time during the day (eg Scoresby Rd).

 

'Greater good' service reforms

It may be that if you have a neighbourhood with a combination of (i) high productivity but limited service regular routes and (ii) very quiet Night Network routes there may be an argument to shift resources from the latter to the former if you can demonstrate that (say) 90% of passengers will be better off. 

You might start by reviewing usage patterns of lower productivity routes. There may well be route reforms you can do. And/or timetable adjustments that release resources for 'greater good' service upgrades on higher productivity routes. One order of doing things might be to tackle the widest productivity gaps first as that maximises the ratio of those who gain versus those who might lose.

This is shown below, where routes are ordered from least to most productive, left to right (although proper analysis would look at day and preferably trip level boarding data). 

On the other hand you might opt for a lower (but still large) gap, especially if the resources freed from a timetable reduced route are not large and you might only have budget for a few trips on a long route.

That alternative is shown by dotted lines pointing to still popular but underserved routes in low income areas. These stack up because of (i) the significant social benefits derived from Sunday service and extended hours and (ii) the low cost due to the route's shortness (Brunswick's 503, Glenroy's 536, Thomastown's 559 and Dandenong's 844 are all good examples).  

You can do this with any set of quiet and busy routes in an area. But since Night Network routes are concentrated at the low usage end of the scale they may well be some of the first a planning agency like DTP might consider if they need to find underused service kilometres for desired service improvements.
 
The patronage uplift of upgraded routes would need to attract a high multiple of the Night Network route's patronage to justify the change and make it politically defensible in terms of increased benefit. That is the donor route and the recipient route are widely spaced on the diagram above. 

The Manningham network transferred resources from the quiet 280/282 to routes that were 2 to 7 times as productive. If you were planning to remove a route (rather than just cut its timetable) you might have a higher threshold, eg a minimum 3 times as productive before you would consider it. A high threshold focuses priorities, maximises early gains and improves return for the small capital costs involved (largely removing and installing stop timetables and adjusting flags). Both 2 and 3x benefit thresholds are also strong relative to big infrastructure build BCRs that can excite people even if as low as 1.1. 

Potential service reforms by Night Network route

Here's a quick run through each special Night Network route. If analysis confirms low usage for the Night Network route and there are high usage regular routes operate nearby then it might be worth investigating 'greater good' improvements for the latter. 

* Night Network 941 Sunshine Station – Watergardens Station

Investigate boosting frequency and/or hours on regular Route 419. Eg Saturday and preferably also Sunday service from every 60 to every 40 min.

* Night Network 943 Watergardens Station - Melton

Suggested retain as V/Line trains do not operate all night on weekends. 

* Night Network 947 Footscray -Newport Station

Potentially replace with later evening finishes or wider weekend span on Route 411 and/or 412 which have above average usage. 

* Night Network 949 Williams Landing Station - Altona Meadows

Potentially replace with longer span on Route 494 and/or 495 on more nights of the week (even if it was just a single trip departing Williams Landing at around 10-11pm on all weeknights). Some catchment would also gain from 411/412 improvement above. Both 494 and 495 have above average usage. 

* Night Network 951 Brunswick Station– Glenroy Station

Potentially replace with 7 day service on Route 503 along Albion St. Route 503 has above average patronage, serves dense housing development but has early evening finishes and no Sunday service. 

* Night Network 953 Broadmeadows Station - Craigieburn HIGH

Potentially replace with longer span on Route 541 which is main north-south bus between Broadmeadows and Craigieburn.

* Night Network 959 City – Broadmeadows Station

Suggested retain as may be busier than most other Night Network routes and is effectively a replacement for 59 tram (that is not a Night Network route). 

* Night Network 965 Lilydale - Healesville Loop HIGH

Potential to replace with longer hours on popular Route 683, especially extra early weekend morning trips. Potentially 685 also. 

* Night Network 967 Glen Waverley - Croydon 

Potential to reform in conjunction with Knox area bus network review that adds a route on Scoresby Rd. Currently Scoresby Rd gets a better service around 2-3am weekends than at any other time.  

* Night Network 978 Dandenong Station - Clayton Station HIGH
* Night Network 979 Clayton Station – Dandenong Station HIGH

Potentially replace with 7 day service and longer operating hours on 814. Route 814 covers most catchment of 978 and 979, has very strong patronage but very short operating hours (including a 1pm Saturday finish and no Sunday service). 

Route 814 is indirect with a weak western terminus. It disobeys all of the attributes of a well planned bus route. Yet, thanks to its strong demographic catchment (around Springvale/Dandenong) it has significantly above average usage of around 25 boardings per hour Monday to Saturday. Because there is an almost perfect correlation between Saturday and Sunday bus patronage, you can be confident that a 7 day Route 814 will be strongly used on Sunday with around 20 boardings per hour. Potential network reform could even involve the 814 being run to Noble Park station to reduce costs and provide a new local connection. 

* Night Network 981 Dandenong Station – Cranbourne

* Night Network 982 Dandenong Station – Cranbourne

Potentially replace with 7 day service on 844, later trips on one Endeavour Hills route and/or simplify Route 894. 

HIGH priorities for consideration are due to factors like (i) a regular route covering a large proportion of a Night Network route, thus minimising the catchment that gains nothing, or (ii) the regular route being productive but having low service levels, in some cases even lacking Saturday afternoon and Sunday service. 

Some upgrades would need funding supplemented from other sources (eg timetable reforms on regular routes). But the greater good framework identified above could remain valid. There may also be issues if Night Network routes are not operated by the same bus company that runs regular routes. 

Conclusion 

Identified above are cases where a review of bus resourcing for Night Network versus regular routes may result in some significant overall benefits for bus passengers.

These benefits are likely to be greatest in parts of Melbourne where the network is least developed, eg on routes that lack 7 day service and/or where productivity warrants improved service. 

Winning public support or at least acceptance is also made easier with widely agreed points that more people will be needing buses at 2pm on a Saturday or Sunday than 2am on those days. Especially if corroborating data can be produced showing the much lower use of Night Network routes versus regular routes that are proposed for upgrade.   

Index to other Useful Network items here

TT 202: Substitutes Skimping Sunday Service - How our trains didn't run for the kids

 

Replacement buses have become almost an expectation rather than an exception for Melbourne train travellers. Right now some lines go hardly a day where there isn't a replacement bus at some time of it.

 

Some shutdowns are due to necessary rail maintenance. Some is due to rail upgrades including extensive testing of new train and signal systems. Others are due to more road-based projects like level crossing removals and freeway construction like the North-East Link project.

Some rail old-timers tell me that calling in the buses is nearer to being the first rather than the last resort option these days. Although it is probably also true that much of their career was spent during times when patronage was stagnant and we weren't building much. Also Melbourne is not alone when it comes to extensive closures. Auckland's entire rail system is closed as we speak and will remain so for most of the rest of this month. And trains to Armadale in Perth won't be running for the better part of 18 months straight.

Politically Melbourne's rolling rail closures is regarded as a necessary cost of the Big Build in which we have disruption now for a better transport future. The extent to which we'll get much of latter was debated last week with media reports saying that the Metro Tunnel timetable may not fully deliver envisaged frequencies. This is something to watch as not getting the full benefits from major projects may affect their 'social licence' including concern over the cost and disruptions they involve. 

Whether you think that the number of shutdowns is excessive relative to what was achieved under them or not, it is undeniable that their time burden is shouldered disproportionately. Rail travellers shoulder it more than car drivers. Non car-owning rail travellers on the weekends and evenings are hit hardest of all. This is because our train services are sparse at those times with us far trailing Sydney in this area.  That means longer waits especially for time-critical trips where passengers might build in 30 to 60 minutes contingency over and above regular travel times. Connections when substitute buses are running also become chancy with it being basically impossible to time one's arrival to meet a feeder bus (which will be infrequent at best or have stopped running for the night at worst).  

A much better future in terms of improved travel will be needed before Big Build benefits outweigh current delays for this group of passengers. Yet it's a future the government seems wont to postpone,  with significant 7 day service frequency increases receding like mirages, rarely following infrastructure builds even if possible all along. Even the Bus Plan appears dead, with DTP proposals rarely winning budget funding. 

A contempt for the time of evening and weekend travellers may pervade influential parts of the public service. From the dismissive tone of the reports they write it seems that weekday bureaucrats as senior as the state auditor-general regard evening and weekend travellers as expendable, despite those needing to work then typically lacking work from home options. This often diverse working class demographic is also the type that is becoming less 'rusted on', with the Werribee by-election having lessons for both major parties. 

Furthermore, functioning and frequent evening and weekend transport affects the experience around major events, including those that civic leaders rely on to bolster the 'world city' credentials of harbourless and often grey-skied Melbourne.  

That's also a problem with residents because travel to such events may be the only time that Joe or Jo Average has with the train system (if they don't work in the CBD). Not to mention the huge number of (not well paid) people who need to travel into the CBD on weekends to staff major events. 

Where there are new parallel freeways, disruption fatigue has contributed to rail patronage not yet recovering to pre-pandemic levels on often disrupted lines like Frankston. 

Maybe they are trying to shoehorn as much works as possible in before the Metro Tunnel opens, with some relief between that and the 2026 state election? We'll see. 

Let's put all this aside and assume that we're going to continue having lots of rail shutdowns. What can go wrong when you ask buses in traffic to do the job of higher capacity trains on their own rights of way? Here's a few things. 

Not enough road space

Replacement buses can get stuck in traffic. They also contribute to it, though they are much more space efficient than the average single-occupant car. The slower a replacement bus the less effective and more expensive the operation is, with more buses and drivers (which you might not have) required to move a given number of people per hour. Travel and waiting times blow out, there is more variability and connectivity reliability suffers.

Solutions might include clearways for replacement buses on major roads and revised traffic light timings to assist buses to leave and enter station precincts. There may also be scope to relieve pressure in some areas by adding tram services and/or operating tangential replacement buses on quieter streets from other railway lines. However the latter may need additional trains scheduled to provide needed improved capacity and frequency. We have done some of that already. Maybe there's scope for more. 

Not enough buses and drivers

This is particularly a problem for unplanned disruptions, especially at peak times. There is more latitude for planned disruptions due to works but even for those there can be pressure if many are happening at once. 

There may also be a bias against scheduling too many buses as each one costs money. But one 45 seat bus is not the equivalent of a 6 car train. Even at night trains do need to be met by more than one bus, at least on major lines. That also provides a contingency if traffic congestion means that buses can't keep to timetables and connect properly with trains.  This problem is exacerbated by low rail frequencies as if rail is infrequent replacement buses may be too. Or even if they are not then bus/rail connections may be poor.

Not enough trains scheduled

Simpler occupations

The simplest type of rail occupations keep the same train timetable but turn back the trains early, with buses operating the outer portion. For example Ringwood to Belgrave or Mordialloc to Frankston. Lack of trains won't be a problem for these as usage won't be higher (and could well be lower) than normal.

Under these circumstances the only reason you'd need to add more trains is to preserve span and frequency. If you skimp on that then you don't have a true rail substitute service. 

Firstly span. As replacement buses are slower than trains earlier outbound and later inbound trains should be scheduled so that operating spans from the termini are no shorter than now. That is if the first weekday morning train from the city to Frankston arrives there at 5:50am (having left Flinders Street at 4:46am) and the replacement bus from Mordialloc takes 30 minutes longer (including transfer time) then an additional train departing the city at 4:16am should be scheduled. The same is true for inbound trains so that those getting the last bus from Frankston are not stranded at Mordialloc. In other words spans need to be no less than when the full train service was running for both early morning arrivals and late evening departures. 

Secondly frequency. This is an issue during transition times, that is when frequency goes from low to high, or the reverse. It's particularly notable on weekend mornings (especially Sundays) when the timetable is transitioning from the hourly Night Network to something better. Again as buses are slower you need to have the replacement buses starting their higher frequency earlier than the trains. It is  also desirable to have one or two more inbound trains scheduled to avoid the risk of extreme bus>train waits.

More complex occupations

More complex occupations can happen on the city side of junction stations like Caulfield and North Melbourne. Normally these junctions have multiple lines with high frequencies at almost all times.

Trains tend to be fullest at these locations. However some closures can suspend services on one (or multiple) lines through these junction stations, reducing capacity and (importantly) frequency by 50% or more. A common example would be where all trains from the busy Pakenham and Cranbourne lines terminate at Caulfield with passengers expected to change (via a non-DDA compliant underpass) to trains on the generally quieter Frankston line. That can cause overcrowding if additional trains on the lines that are still running are not put on (which they often are in this case). 

There are instances where unharmonised timetables can create connection complexity. In the above example, Sunday morning trains from Cranbourne/Pakenham are every 30 minutes versus every 20 minutes for trains from Frankston. That can create uneven connections with long waits unless extra trips are added. Variable connection times at an inaccessible station is a special concern for passengers with accessibility needs who might take longer to change platforms or even need assistance (petition to fix that here btw). 

Trains that terminate can at least be timetabled with connections planned. Replacement buses are less predictable, with more potential for these to arrive early or late. And on the longer lines buses may have different express stopping patterns to cut travel times. However if they arrive at a station that only has infrequent trains then those time savings may be lost for passengers travelling onwards. 

You would think that a station like North Melbourne wouldn't have these problems.

After all, on paper it looks highly served. No less than five Metro lines (Werribee, Williamstown, Sunbury, Craigieburn and Upfield) stop there.

You wouldn't think recourse to a timetable would be necessary to get a train from there into the city at any time trains are running. 

But you'd be wrong.

Keep reading! 


The weekend before last at North Melbourne

As happens on many Spring/Summer/Autumn weekends there were major events in the city. Often with early starts they attract thousands of participants, spectators and, most importantly, workers to make it all happen.

It's pretty much impossible to schedule a weekend rail closure without affecting some major event. Or the millions of little humdrum events in Melburnians' lives, some of which involve train travel.

The weekend before last closed the busiest lines from the west to perform what insiders call a 'heart transplant'. That included replacing the entire signalling interlocking system at Franklin St between North Melbourne and Southern Cross. 

That Sunday, April 6, also had Run for the Kids, which in 2024 attracted 25000 participants.

Extra trains

Hence extra trains were put on to supplement Melbourne's notoriously sparse Sunday morning timetable. Two trips per operating line were typically added. This was a flat number, regardless of whether the line already had a relatively frequent service (eg Frankston), an infrequent service (eg Mernda) or was quieter than average (eg Alamein).  

Replacement busing

Note I said operating line. Werribee and Sunbury lines, though busy, were conspicuously absent. That's because they had no trains. Instead it was replacement buses all weekend, as noted below. 

Travel from Werribee or Williamstown would involve a train to Newport (typically every 20 minutes), then a replacement bus to North Melbourne and then a train into the city. Sunbury line passengers had a similar arrangement with a change to a bus at Sunshine. Although their morning trains are only every 40 minutes, so connections would be chancier.

Neither the Sunbury, Werribee nor Williamstown lines had special Run for the Kids trains added on the sections that were still operating. 

If all that sounds complicated here's a schematic below. Click for a clearer view. 

Even if it meant some backtracking, savvy passengers would likely do what they could do to board a V/Line train at stations like Tarneit, Deer Park or Ardeer. But not Sunshine due to V/Line boarding restrictions. 

The North Melbourne problem

Let's give everyone the benefit of the doubt and assume perfect timed train > bus connections at Sunshine and Newport.

The complexity comes when buses arrive at the North Melbourne end. This is because with half the lines knocked out (including the more frequent Werribee/Williamstown service), there are big gaps between trips in to the city. The only trains running are from Upfield and Craigieburn, both of which are infrequent on Sunday mornings.

Consequently North Melbourne loses its normal turn-up-and-go service. Which poses problems for those arriving via replacement buses (which, if express and stopping, will be at varied times). 

Do the extra Craigieburn and Upfield line trains put on for Run for the Kids help? They do but there are still large gaps. Because they cheaped out by adding 2 and not 3 trains per line there remain 40 minute gaps on both. This false economy prevented a more consistent 20 minute service on each line (which could have run in a 2 hour block). 

Not only that but the additions gave North Melbourne 10 trains (unevenly spaced and with different CBD destinations) in a 2 hour block. That is less than the 15 trains per 2 hour block that run under the regular timetable (being Upfield, Craigieburn, Sunbury 3 each plus Werribee/Williamstown on 6). 

The consequence of this is a lumpy timetable at North Melbourne with gaps of up to 25 minutes (around 8am) and as long as 46 minutes around 6am. These are unsuitable for a replacement bus (whose timetable is constrained by infrequent arriving trains, notably at Sunshine) to reliably connect to. More details on that are below. 

 

CLICK ABOVE FOR BETTER VIEW

Even if the closure of the Sunbury and Werribee lines couldn't be helped, nothing stops other ways to make the connections at North Melbourne reliable. The long gaps could have been closed by running (say) a 10 minute service from Craigieburn and a 20 minute service from Upfield. That would have compensated North Melbourne for the loss of Sunbury and Werribee services. Thus maintaining the existing regular non-event 15 trips per 2 hours Sunday morning frequency from there. However this wasn't done, leading to the issues you've read about above. 

Swing seats vs safe seats - guess which got more trains! 

On the other side of Melbourne precedent does exist for running extra trips on operating lines when other lines through a junction are closed.

An example, also from April 6, was at Caulfield where extra Frankston line trains were run to and from Moorabbin to partly compensate for the Pakenham and Cranbourne line trains terminating there. Like North Melbourne, big numbers will be boarding trains to the city. The only difference is they're coming off trains that are terminating there, rather than buses. That timetable may still have had problems for early participants. But there was at least acknowledgement that you could add extra frequency on a running line to assist capacity from the junction station - something that Craigieburn and Upfield largely missed out on.   

Will next year be better? 

We'll have the Metro Tunnel running for a start. This should mean better service between at least Sunbury and Dandenong. So that's two lines. These improvements will help even when bus replacements are needed because the starting point when these are scheduled are that they at least match train frequencies. 

As for the others, that depends on whether the Metro Tunnel timetable has wider benefits for related lines like Sandringham, Craigieburn and Upfield. The Age online for 9/4/2025 (printed edition 10/4/2025) reported on post-Metro Tunnel services, including off-peak frequencies and the risk they might not get all envisaged  frequency improvements. 

Conclusion

The governments of Daniel Andrews and Jacinta Allan have been great at the concrete mixer but have  displayed weaknesses in service planning and operations. Even for travel to major events that their government seeks to promote and which cabinet ministers attend. To date the predominant interest has been in the building, not in the running or the maximising of the public benefit arising from potentially very worthwhile constructions.  

With expectations raised for the Metro Tunnel opening later this year, the government will need to get smarter at service across the week and across Melbourne if it is to deliver the benefits everyone hopes it will.  

Other Timetable Tuesday items are here

TT 201: How many hours of frequent service does each line get?

Every large city with substantial public transport usage has a core trunk network of high capacity lines that are fast, direct and have their own rights of way. Most Australian capitals have electrified suburban rail systems to do this job, with some having busways too.   

What about service levels? 6am to midnight is about the basic span, with Sydney and Melbourne having somewhat longer hours. Notably Sydney's early starts and Melbourne's all-mode Night Network on weekends.

The biggest differences are frequency. This typically varies between 10 and 60 minutes, with 15 to 30 minutes being most common. Some cities run lines that are only frequent in peak times while others operate all their lines frequently all week. This has a major bearing on whether the network is largely for CBD commuters or it can also suit diverse trips at more varied times. 

Melbourne's maximum waits

Where does Melbourne stand on suburban rail frequency? Last year I looked at maximum waits across the network. Two things were notable: (a) big service variations, with the better served lines having half the maximum waits of others, and (b) longer maximum waits than other cities, with 40 minute gaps being a Melbourne specialty versus the 15, 20 or 30 minute maximums more common elsewhere. 

Most would assume there's a rational reason for such service variations with lower frequencies provided because the lines were quieter. However this is not so, with busy stations on lines like Sunbury and Craigieburn having amongst the longest maximum waits. Instead it's more a historical/ political thing with lines in the south and east having shorter maximum waits. However Sunbury line stations should get some relief under the Metro Tunnel timetable starting later this year.  

Who gets the most hours of frequent service?

Maximum waits was one way to compare service between lines. Shortening them by using otherwise idle trains is a highly cost-effective way to massively speed end to end PT travel. Yes, I said speed. Cutting trip time variability by boosting frequency is the single most effective way to increase all-day speed on a suburban network like Melbourne's, especially for multimode trips. 

It reflects poorly on rail planning priorities in Melbourne that so little has been achieved in this area. However if the maximum wait is only experienced for a few hours of the week (eg Sunday mornings) it can be less representative than other measures of service quality. 

Counting the number of hours per day or week a line offers frequent service is one such better measure of rail service and thus usefulness. If it's only a few hours per day then we know the line favours peak commuters. On the other hand a line that is frequent for 15 to 20 hours per day favours more diverse trips. The latter also also offers better return on infrastructure and greater scope to reform tram and bus networks to feed rather than parallel trains (as happens in cities with less frequent rail like Brisbane and Adelaide). 

What should one define as a frequent service? It depends a lot on the trip but this exercise requires a single number. While arguably not quite turn-up-and-go, a 15 minute all-day frequency is a popular choice for key rail and bus routes in Australia, so I've chosen it here. 

You'll all want to see the map first, so without more words here it is. 

CLICK MAP FOR CLEARER VIEW

The four per cent of week classifications I use makes the data fall into convenient groupings; That is frequent service (i) Mon - Fri peak only, (ii) Mon - Fri day only, (iii) Mon-Sun day and (iv) almost all week day and night. 

Apart from some inner area stations (bordered by Newport, Footscray, North Melbourne and Clifton Hill), frequent suburban rail in the west and north is a peak only affair. As it is in the outer east and south-east (check the grey lines on the map - deliberately made less prominent). 

Whereas in the east and south there is frequent weekday and sometimes frequent weekend service during the day. Sydney and Perth are shown for comparison in the top right. Both beat Melbourne on weekly hours of frequent rail service whereas Brisbane and Adelaide lag us.  

The map shows a huge inequality across Melbourne in the provision of rail service, with this even more marked when we check the numbers per line. That's up next.  

Variations between lines and cities graphed

The map above was derived from a graph. That's the product of a spreadsheet that is in turn the result of checking PTV website timetables. 

Let's talk about the graph first. It shows weekly hours of frequent service by line. Lines are shown in clockwise order, starting with Werribee in the west. The last three columns are typical interstate lines for comparison. 

CLICK GRAPH FOR CLEARER VIEW

I usually analysed service at the last station on each line unless many trains terminate before it. These exceptions apply for Sunbury (Watergardens), Hurstbridge (Greensborough) and Lilydale (Croydon). Doing this prevents unfair undercounts of service on a line's busiest portion. Similarly junction stations like Ringwood and Dandenong get their own bars on the graph where these are at the end of a corridor with many busy stations. I left Newport and Clifton Hill off the graph but did have them on the map.  

Lines with around 20 to 40 hours of frequent service per week are only frequent during the weekday peaks. Examples include most major lines in the west and north including Werribee, Craigieburn and Mernda. Branches in the outer east and south-east such as east of Ringwood and Dandenong also only operate frequently during the peaks. 

Lines with around 60 or 70 hours enjoy 12 to 14 hours of frequent service each weekday. That includes  midday and both peaks. Evenings and weekends are not frequent. Examples include Glen Waverley and Sandringham, both in the east or south. 

Melbourne's best served lines have 90 to 100 hours of frequent service per week. This includes daytime every day of the week. However in Melbourne's case frequent service vanishes at night and doesn't start until late on weekend mornings. Trains to Ringwood, Dandenong and Frankston fall into this category. All are in the south or east, with their prevalence of frequent service being 3 to 6 times that of lines in the west or north.  

Williamstown and Alamein are both quiet lines. However Alamein, with frequent service between the peaks, is better served than Williamstown which has none and even Upfield which has very little. 

How do lines in other cities compare? 

Perth's two lines reflect that city's new and heritage lines and thus the network generally. Both enjoy over 100 hours of frequent service each week, exceeding any line in Melbourne. 100 out of the week's 168 hours means frequent service 14 hours per day extending to approximately 9 pm. In Perth you can rock up at any station and, as long as it's daylight, have a train arrive within 15 minutes. Whereas that number in Melbourne is 20, 30, 40 or even 60 minutes due to our higher maximum waits. 

Sydney is however the gold standard with frequent service for 140 hours per week, or 20 hours per day for the Penrith example given. That ranges from before 5am to after midnight on any day of the week.

To put the differences in another way, Sydney's outer western station of Penrith (55km from CBD) gets as much frequent service in one day as Melbourne's outer western station of Werribee (30km from CBD) has in a week. Similarly Perth's Mandurah (70km from CBD) gets more than triple the hours of frequent service that Melbourne's Craigieburn (30km from CBD) receives. 

The raw numbers

Want the raw numbers? They're below. 

How did I get the numbers? I just looked at PTV timetables and counted the span of hours in which there was service every 15 minutes or better. Sometimes that involved three time blocks as on some lines there were longer than 15 minute gaps in a peak period. I did this for both inbound and outbound directions and averaged the two. Weekday hours were multiplied by five and then added to Saturday and Sunday hours of frequent service (if any) to get a weekly total. 

Weekdays versus weekends

So far we've only looked at the amount of frequent service per week. That obscures differences between days, eg weekdays versus weekends. The graph below has hours of frequent service split out per day.  

CLICK GRAPH FOR CLEARER VIEW

Notable features include: 

* The hours of frequent weekday service at Melbourne's three best served corridors (Ringwood, Dandenong, Frankston and Sandringham) are about the same as what Perth's Midland line gets on weekdays (their Mandurah line isn't shown but is similar).  

* Sydney's Penrith has more hours of frequent service on a Sunday than any Perth or Melbourne line gets on any day of the week. In fact Penrith's frequent service is basically uniform all week with 20 hours per day provided.   

* Perth's rail network has approximately 15 hours of frequent service on weekdays. That steps down to 13 hours on weekends. 13 hours is midway between the wide span of frequent weekend service in Sydney and the much shorter spans in Melbourne.   

* Frequent weekend train service does not exist beyond inner suburbs in the west and north (ie beyond Newport or Clifton Hill) whereas it extends further in the east and south (to Ringwood, Dandenong and Frankston).

* Even on its best served lines, Melbourne is alone in having a much shorter span of frequent service on weekends than on weekdays. That is 8-9 hours (or ~10am - 7pm) versus 15 hours on weekdays. This means that unless your trip is short the chances are that at least one and possibly both legs of your trip will be travelled when service is only half-hourly or worse (especially on Sundays). To be fair, Melbourne passengers on its five frequent weekend corridors get a 10 rather than a 15 minute service as they do in Sydney. But it is probably also true that the Sydney approach of a long span 15 minute service is more attractive than Melbourne's lumpy 10-40 minute gaps while costing similar to run.  

Conclusions

The findings for access to frequent service are even starker than they are for maximum waits.

Whereas maximum waits had about a 2:1 difference between worst and best served major lines in Melbourne, the difference in relation to hours of frequent service per week are more like 3:1. Perth and Sydney lack this huge service inequality between lines. 

The same lines that were short-changed on maximum waits were also short-changed with the fewest  hours of frequent service per week. Examples of lines that are underserved relative to their patronage include Craigieburn, Mernda, Werribee, Upfield and Hurstbridge, roughly in that order. Alamein, in contrast, was generously served, with more hours of frequent service than main lines in the north and west. Although Alamein's frequent service is largely interpeak, not peak. 

  

Even Melbourne's busiest lines had fewer weekly hours of frequent service than Perth lines, with dramatic differences on weekends. Still influential austerity scheduling has given Melbourne's quirky  train timetables with long waits at stations and at times people would not expect them. There's no link between service and development intensity either. A lower rise Moonee Ponds had many more hours of frequent train service in 1939 than high-rise Moonee Ponds does now.  

The thing that really stands out though is Sydney. Their 20 hours of frequent service per day at most stations sets the standard that we should aspire to. Melbourne did the planning work required with the Network Development Plan (Metropolitan Rail) released in 2012. Finally delivering the service uplifts in that would likely have bigger patronage benefits than almost any conceivable rail infrastructure project in the next decade.  

** UPDATE: Thanks to Lachlan Abbott from The Age for covering this analysis in detail in an article published April 10. Read it here. A Herald Sun article the previous day focusing on less than envisaged peak service levels under the Metro Tunnel is here.  

Index to other Timetable Tuesday items

UN 198: How fast is Melbourne really upgrading bus services?

The other day I found an interesting Age article from September 2010 discussing the pace at which the government was upgrading buses. It's so old that you don't need a subscription to read. 

The article is on what was considered the slow pace of bus reform. Advocates wanted faster bus network reform while the government was defending its record of adding service. 

What we have now but lacked then was the vantage point of history. Both the advocates and the government spokespeople (who then allowed their names to be published) turned out to be right.

Yes, more could have been done. But it's equally true that the Brumby government was achieving a lot with buses with a plan that not only existed but also got significant funding. It was arguably even a golden age for bus service additions if not for bus network reform (that was to come a little later).    

Feast or famine, one constant is the imperative for ministers and departments to defend their achievements. Let's look at the past quarter century to see how this government's Bus Plan record stacks up compared to its predecessors so we can more fairly evaluate progress.  

Bracks/Brumby government policy approach

First network reform. The Bracks/Brumby government had commissioned 16 local area bus network reviews (which you can read here) but implemented only a minority of recommendations. The reviews were basically the third (and least successful) prong of that government's bus reform agenda. This agenda was assembled under a plan called Meeting our Transport Challenges (or MOTC) that had been put together to respond to building (but then still underestimated) stresses on the transport system.  

While it had implemented some service upgrades in 2002, it was not until its seventh year, in 2006, that it got really serious about buses. To understand this take yourself back to the mid-noughties when premier Bracks was enjoying a big majority from 2002 and the Liberals were busy churning leaders. 

Buses were cheaper than and served more suburbs than rail. Rail franchising had failed in its first iteration with a big bail-out, funded by tolling EastLink, needed. V/Line to marginal regional seats got big upgrades but metropolitan train and tram basically stagnated under multiple rebrandings, with 1999 election promises of extensions substantially broken. The Melbourne 2030 plan raised and then dashed expectations with no accompanying funded public transport program to complement the intended higher population densities around stations. 

Much of this austerity was due to Labor's wish to be fiscally prudent to avoid comparisons with their embarrassing last time in office a decade prior. Borrowings for big projects were off the menu. Hence the off-books PPP arrangements for Southern Cross Station's construction and operation. Which proved a mistake in retrospect as it ceded control of a key station to faceless financiers. Although to be fair, Eastlink, another PPP at the time, was well regarded as a project.

Improved financial conditions did however permit some increase in direct government recurrent spending, even if not yet on megaprojects. Instead a good proportion in transport went on improved bus services, encouraged by advocacy from people like John Stanley in BusVic. This is opposite to this government's recent policy settings but more on that later. 

Returning to bus reform, the other two prongs mentioned in the 2006 Meeting Our Transport Challenges Plan were (i) minimum service standards for buses (as most routes then finished at 7pm and didn't run Sundays or even necessarily Saturday afternoons) and (ii) a network of premium service SmartBus routes (notably the orbitals but also routes that substituted for expected rail extensions to Doncaster and Rowville). 

Although a significant minority of bus routes remained without 7 day service and the SmartBus orbital program was not fully delivered, you could reasonably say that about 60-70% was by late 2010. That's a pretty good promise/delivery success rate with bus patronage rises closely tracking the service increases.

In contrast only 153 out of 711 bus review recommendations were delivered according to the 2010 Age article. This made the bus reviews the weakest of the three prongs of bus reform. That can be attributed to factors like complexity, perceived political risk and cost as implementation was largely unfunded.

Bus network reform done right can be more cost-effective than just adding trips on existing indirect routes. The reformed 2014 Brimbank and 2022 Yarra Valley networks are prime examples. While it was largely timetable rather than route based, the 2025 Manningham area service redistribution (which upgraded weekend service on four routes) also deserves high credit. In contrast some of the reform proposals in the 2006-2010 era reviews were expensive and retained some inefficient overlaps. Then new SmartBus orbital alignments (sections of which suffer low productivity and/or duplicate other routes) were also unchallenged.   

Another factor weakening government interest in buses was changing transport politics as rail meltdowns increasingly dominated the headlines. MOTC's weaknesses on rail made its period as the main transport plan for Melbourne short. Hence the Victorian Transport Plan of 2008, fed by work from Sir Rod Eddington. However that plan's increased investment in rail came too late for the Brumby government which was defeated in 2010 (as the reliability turnaround happened in 2012-2013).        

Baillieu/Napthine government policy approach

The subsequent, fiscally tighter, Baillieu/Napthine Coalition governments had other priorities, including restoring rail reliability, staffing stations with PSOs and making a descoped myki work. All three could be considered successful. However there was also scrimping through ill-advised decisions to reduce the scope of the Regional Rail Link (then under construction) and accept a cheap and nasty offer from Transdev to take over the franchise for Melbourne's longest and busiest bus routes.   

The Liberals (and the Andrews Labor government that succeeded it) added no new SmartBus routes. However the (effectively then bipartisan) approach to improved metropolitan rail frequency continued. The pace of bus network reform also picked up under new PTV CEO Ian Dobbs (back for his second sojourn after leading the PTC in the '90s).

An Englishman like many Melbourne transport leaders, the high-handed Dobbs would readily wield a felt pen to shorten or extend a route his staff had planned. He was also a radical, much less sensitive than Labor ministers (especially) to the political risks of network reform. The highlight of his (and PTV's) planning was reached on July 27 2014 which saw large reforms on the franchised Transdev bus network, reformed Brimbank bus routes and turn-up-and-go off-peak trains to Dandenong. This included an extra 3260 bus, 470 tram and 200 train services per week. While it had some shortcomings, no bigger integrated network and service reform has happened across Melbourne in the decade since. 

Transport minister Terry Mulder generally accepted bold service reform while his successor Jacinta Allan mostly recoiled, seeing more risk than reward. Dobbs was gone within six months of the new Labor government taking office. Proposed metropolitan rail and bus service reform was abandoned a few months later, with a decade of big infrastructure replacing the previous decade of service.   

Bracks/Brumby government record on bus service

How much bus service were we adding when political interest in it was at its peak? One answer was in the final paragraph of the 2010 Age article. That cited 13000 new services per week added in the four years since 2006. Or about 3250 weekly trips added per year. 

Note that these services could be anything from new Sunday trips on a 10km long local bus route to trips on an entirely new orbital route (which could be up to four hours long) as were added up to 2010. Or they could be just a ten minute ride on a short university shuttle (such as have been added since 2016).

Even though it's the preferred measurement metric in the political playground, counting trips is a terrible measure for anyone wishing to understand the real magnitude of a service increase. For that purpose it's better to look up annual service kilometres in the budget papers, like I did here.

By any measure the Brumby government added a lot of bus service. That grew patronage, with this continuing after that article was published. However certain network planning efficiency gains that could have been achieved then were not. However some were tackled under the following Baillieu/Napthine government, notably in the big 2014 reforms mentioned above.   

Andrews government record on bus service

Sticking to trips added, what have been more recent claims?

The government, which did largely honour its 2014 promises on bus services, told the Metropolitan Transport Forum that it has added "20 000 new bus services in 9 years" to 2023. That's about 2200 weekly trips added per year. Or about 1000 less than the 3200 per year average between 2006 and 2010. 

The disparity grows when you factor in Melbourne having a smaller population then and the longer average trip lengths due to the SmartBus roll-out up to 2010 versus the concentration on short university and Port Melbourne services more recently (eg 202, 235, 237 & 301 every 10 min).  

Measuring annual service kilometres added fixes that 'apples and oranges' problem to give a truer portrayal of bus service increases. For example around 8 million annual bus service kilometres were being added in 2007, 2008 and 2009 as opposed to around 2 or 3 million per year more recently according to budget papers. 

It also needs to be understood that service tended to be introduced at a faster pace in the first term of the Andrews government than later. For instance new bus networks in Geelong, Wyndham, Cranbourne and Plenty Valley plus university shuttles (which have a lot of short trips) were all running by late 2016. As you'll see in a moment the pace of bus service addition declined after then.    

Has the Bus Plan made a difference?

What about even more recently, in the era of Victoria's Bus Plan, ie since June 2021? In a written response to PAEC received 28 November 2024, then DTP Secretary Paul Younis said that 4600 bus services were added in that 3.5 year period. That's a rate of 1300 weekly trips added per year, even counting school services (see below). It also looks low compared to the 3260 weekly bus trips that one day's reform added in 2014. 

The evidence is that even if you have a plan it can't achieve much without a funded implementation strategy. As a result Victoria's Bus Plan has yet to deliver the step change in bus network service levels, directness and connectivity that was intended for it 1390 days ago.  

Conclusion

Victoria's Bus Plan notwithstanding, the decade-old Andrews/Allan government has yet to reach anywhere near the pace of bus service uplift achieved under the Brumby government. Or even its own 2015 - 2016 period when big new reformed bus networks (some planned under the previous government) commenced service. What five years ago I called the 'second stupor' for Melbourne buses has yet to be decisively broken. 

The pace of the sort of local tweak bus network reform that is possible without much money has also been slow. That presents many still untapped opportunities to cost-effectively get the most from our buses. 

If this government is serious that the time for buses is now it will need to get moving!  


See more Building Melbourne's Useful Network items here

Transport and Planning done differently: Introducing the Green Light People

Just  because little has been publicly heard from new DTP Secretary Jeroen Weimar since he took over two months ago does not mean he has been idle. Indeed it's been the contrary with much private work, the pre-requisite to public success, being completed. 

This work has included a strategic review, addressing questions such as how best to restore DTP's central role in the transport policy framework, reviving its public standing and reposition it to become, in premier Jacinta Allan's words, a builder not a blocker.

It's nothing short of a complete 'go revolution' for a department that's long needed something like it. 

A 'dep sec diet'? 

Having a staff of thousands to write business cases that too often failed to win state budget backing or draft strategies no one reads was the first bit of junk for Jeroen to jettison. 

Taking ideas from the Yanks but lingo from the Brits, Source Savings or Sod Off is said to be the survival priority for the department's deputy secretaries as the state struggles to find offsets for North East Link's $10 billion budget blowout.

Thus, to avoid potential embarrassment, the organisation chart listing dep secs was quietly deleted from the DTP website sometime between January 8 and February 25 2025 so the nosy public has to wait for the annual report (or lurk Linked-In) to see whose gone.

Each dep sec cut can buy about $400k in extra program funding from their employment costs saved. That excludes even more freed from supernumerary send-offs and time savings accruing from processes that have to be, by necessity, leaner. While those who remain may gain broader responsibilities this is partly mitigated by them having exponentially fewer meetings, as illustrated below. 

Sources close to the Secretary said that with 'razor gang' Silver Review staffing cuts looming, it was better to be on the front foot rather than effectively being under Treasury administration (as happened in the 1990s with the Transport Reform Unit). 

Enter 'The Green Light People'

The above may seem pretty dire if you were a targeted executive. But DTP leadership is very alive to the need to maintain morale internally and community standing externally. Full transparency requires a message that works for both internal self-talk and external engagement. 

That's best articulated by DTP becoming The Green Light People 

The new DTP will facilitate, not fumble.
The new DTP will catalyse, not clog-up.
The new DTP will revitalise, not retard.
Above all the new DTP will build and not block. 

Green Light will replace Red Tape in everything the Department does. 

Green Light's three key directions in the DTP portfolio include: 

Housing and Planning: The Green Light principle is most obviously topical for tacking Victoria's housing affordability crisis. At its core the new Plan for Victoria is about saying yes to more.  

Transport: the narrative here is about keeping people moving. First emphasis will be on cutting waiting at intersections (through shorter traffic light cycles), faster end-to-end public transport (through higher frequency partly funded by service reform) and improving railway reliability (by fixing unreliable recurring track and signal faults). A smaller but stronger DTP will also veto planned rail shutdowns that disrupt too many passengers. 

Cost of Living: An overarching theme across all of government for next month's state budget, with implications for DTP explained here. 

Branding to convey 'Vision Go'

For maximum transparency there will be no difference between internal and public-facing branding. After 20 years of obsessive branding and rebranding between 1998 and 2018, public transport (especially) went through a period of 'brandlessness' with little or no branding appearing. This opposite extreme led to a loss of network identity and confidence.

Green Light People branding will restore transport network pride and the sense that there's someone competent looking after the system and helping you go. You'll recognise this by the green light triangle progressively appearing on drivers licences, the public transport network (example below) and completed housing projects.  

QR codes will be installed at intersections to allow walkers to report waits of more than 60 seconds and drivers more than 120 seconds. The volume of requests thus received will dictate funding for "Vision Go" speed-up treatments. This will be done in a similar mechanism to the community-driven 'Pick My Project' scheme trialled before the 2018 state election. 

Summary

Delightful in both simplicity and profoundness, The Green Light People gives a new focus that the Department of Transport and Planning has lacked and needed for years. 

To mark the launch it is understood that DTP staff will enjoy a long morning tea, with this strictly concluding at noon today.