Friday, April 29, 2022

UN 123: A simpler Night Network - Can we do it?

In September 2021 Night Network got a refresh. About half the special Night Bus routes (that few understood) got replaced with 24 hour weekend service on 21 regular bus routes (list here). This delivered benefits such as an easier to understand network and extended weekend operating hours  in suburbs that long only had buses that were late to rise and early to bed. Early morning and evening passengers who are neither late night revellers nor workers were among those who gained. 

Implementation cost was low due to some trade-offs including starting some services in the suburbs rather than the CBD and reducing some frequencies (in exchange for 24 hour service without multi-hour gaps). More on the revised network here

Completing the job

What if we wanted to finish the job? After all there remain thirteen special bus routes that are Night Network only (all numbered 940+). These non-standard routes sometimes feature loops designed to maximise coverage. And they only operate for about 4 hours overnight, causing long gaps in areas where regular buses finish around 9pm or don't start until about 8 or 9am on weekends. 

Doing what we did for the rest of the network could eliminate these special routes and extend 24 hour weekend service to more regular bus routes. That's a major advance on the road to making Melbourne's buses simpler and more useful. 

Where are the special routes to reform? Mostly in the outer west, north and east, they're shown in black on the PTV map below. Click here for a clearer map and further information on them. 

Trade-offs and complications

While desirable, reform isn't all plain sailing. 

The first issue is coverage. Loop routes (like many Night Network buses) are good at delivering wide coverage with a small number of buses. Given there's less traffic at night buses can go faster even if the routes they're on are less direct than during the day. So operationally and financially loops might be a reasonable trade-off. 

The customer's view is different. They don't see operational considerations. But they do know that their local bus finishes at 9pm. They might or might not know about the existence of a dedicated Night Network bus. And even if they do they might not know where it runs or the station they should board to catch it (as it might be different to that used by their regular bus). Points like the latter can be significant disincentives to use.

We saw that with the pre-2021 Night Bus Network. Unlike trains and trams, all bus routes that ran were special services that didn't run at other times. Operationally efficient they were but they didn't carry many passengers. A bit like Yes Minister's perfectly efficient hospital with no patients. 

Replacing one Night Network loop route with a regular route improves simplicity and may boost patronage. Early indications on reformed routes have been encouraging. So what was done last year was a good thing to do; it just need to be extended. 

The problem with replacing special routes with a regular route is that it can cut coverage unless multiple regular routes are upgraded. While desirable that adds extra operating costs. This is especially where current operating hours are short and 10pm - 1am and 5am - 8am trips need to be added for a full 24 hour weekend service. So even if you could do a 1:1 replacement you still need to fund about 12 hours per week more service per route. It's small in the scheme of things (you could do quite a bit with a few million per year) but getting even that for service is hard due to a bias towards capital projects.   

Even if you had funding it can be difficult to decided which regular routes are best replacements for a Night Network special in some areas. The job was easy in Werribee/Tarneit because the area recently  (2015) had a wholesale bus revamp including direct routes along main roads. Whereas other areas (eg Knox and Greater Dandenong) have 30+ year backlogs in bus network reform. That gives rise to cases where no logical regular route replacements exist for Night Network routes. As high as needs might be, you might leave Night Network arrangements as they are until regular routes are suitably reformed.     

Route length is another factor. Planners often base daytime bus timetables on a speed of about 22 km/h. With that speed a single bus can efficiently provide an hourly service on a 10km long route (20 km return trip) without having to sit long at termini. At night where there's less traffic and patronage buses can go faster - maybe 30 km/h or more. 

Thus a regular bus route that is efficient during the day may not be efficient at night. Whereas dedicated Night Network bus routes will be efficient as they've been scheduled for night traffic and loading conditions. The latter is operationally efficient but if it's a complex route that no one understands then it won't be an effective service as few will use it. 

Juggling operational efficiency versus running a legible service people will use is a core problem faced by Department of Transport planners. Unless the operational case against is compelling I will normally favour the legible option. However one still needs to be aware of operational considerations to know what can be done cheaply and quickly and to accurately assess costs. 

Night Network bus upgrades 

Some Night Network routes are easier to replace with regular routes than others. Those easiest to reform can be reasonably replaced with just one regular route without losing too much coverage. And its length should permit reasonable operational efficiency. In contrast harder to reform routes may require changes to the regular bus network so can't be done as quickly. I'll list them in order of night route number, with potential upgraded regular routes identified.   

941 >> upgraded / rerouted 419

The northern two-thirds of this route is quite similar to Route 419. The main difference is that Route 941 starts at Sunshine whereas the 419 starts at St Albans. 

The area around Sunshine got bus reform in 2014 with the Brimbank network. However Route 903 wasn't touched. Sunshine is being considered a major jobs centre and transport hub with airport rail to depart from there. And sooner than that so will Metro tunnel services allowing fast access to Melbourne University, hospitals and the St Kilda Rd corridor. Looking slightly wider a case exists for Route 903 to operate via Highpoint instead of overlapping much of the 465 in Essendon. 

All this raises the potential for reform of the regular bus network. If 903 is rerouted from Buckley St then there will need to be another route to replace it on McIntyre Rd in Sunshine North. That route could be the 419 run south to Sunshine, especially if its run to St Albans could be replaced by an extended Route 406. As discussed here there are many advantages in this including less network duplication, a Highpoint SmartBus, upgraded Braybrook services and a direct connection to Sunshine from Keilor Downs.

With a Route 419 now similar to the 941 alignment, scope exists to make it, like the mirror image 420 on the other side of Sunshine, a Night Network route. Because 419's route length is similar to its 941 anagram, the operating cost should be minimal. That is unless (as should happen) extra trips are added to make 419 a 24 hour weekend route. 

943 >> upgraded 460 (and others)

The first half of Route 941 is like most of the 460 from Watergardens Station. The second half of Route 941 is like the western half of Route 456. Melton needs its Night Network feeder bus to come from the Sunbury line as its unelectrified V/Line service does not operate Night Network. 

Potential may exist to remove Route 941 if adding complication to the 460 is considered acceptable (ie Night Network trips not serving Caroline Springs Station and operating to Melton as per 941 during Night Network times only). Ideally extra trips would run between Watergardens and at least Caroline Springs to deliver the latter area with a 24 hour weekend service on Route 460. 

I don't see the above as being more than a stop gap measure. Ultimately all of 460 should be a Night Network route with no special deviations or extensions. The corridor between Deanside and Melton is rapidly growing and needs its own routes, especially given they have no Night Network Metro trains. These solutions may involve a  new Night Network (and preferably also regular) route along Taylors Rd from Keilor Plains/St Albans and potentially also an upgraded Route 456 from Sunshine. 

Bus coverage in these parts is difficult to provide as many subdivisions are well off a limited access freeway rather than grouped around stations or even an old-style highway. Also I would not recommend changes for Melton township until its town bus network is reformed and the Night Network route can be planned in harmony.

947 >> upgraded 411 

Most of Night Network special route 947 parallels the existing Route 411 from Footscray to Altona North. Then it runs east to Newport direct via Mason St. A simplified service could instead operate as Route 411 to replace the 947 (and part of 949 - see later). There would be some coverage loss in the central part of Mason St however Werribee trains at Newport station would likely provide a faster connection than changing to the bus at Footscray.  

Adding Night Network services to Route 411 would tie in well with a boost to its daytime service, including as a frequent SmartBus to replace the 903 on Millers Rd and better serve a more locally popular destination. Buses in the Altona North area haven't been reformed for years and are more complex than they need to be with a mixture of overserviced/underperforming routes and underserviced/well performing bus routes. More detail on that here.  

949 >> upgraded 411, 494, 495

949 is a complex and indirect loop route that runs from Williams Landing Station via Point Cook and Altona Meadows. Adding Route 411 to Night Network would make the latter coverage unnecessary. Thus the route could be tightened to focus on the Point Cook area only. 

A potential could be counterclockwise loop leaving Point Cook as Route 495 then returning via Route 494. There would be no dwell time at the southern terminus so Route 494 passengers would ride through it on the way home. The pattern would be similar to that which already operates between 10pm and midnight on Friday and Saturday evening, making usage easier than the current Night Network special route. 

This could be one of the higher priority simplifications, although it needs to be done in conjunction with the Route 411 upgrade to retain Altona Meadows coverage.

951 >> ?

Route starts from Brunswick Station then heads north, paralleling the 58 tram. It keeps going north paralleling the Route 527 bus. It runs to Glenroy station via a path similar but not identical to the 534 bus. 

A very minor simplification may be possible if it follows the 534's path. More detailed review should be done at the same time as a Hadfield area local bus network review. The latter is needed as routes have hardly changed in years and service levels on routes like the 536 lag the area's demographics and usage pattern. 

953 >> upgraded 541

Night Network route 953 is roughly like regular route 541. Both start at Broadmeadows and run north of Craigieburn. Route 541 is slightly less direct but has more residential area coverage away from train stations. Thus overnight trips 541 could completely replace 953 with no changes required to other routes. 

Route 953 is done with one bus. The timetabled Broadmeadows-Craigieburn time for Route 541 is 41 minutes (last Sunday night trip). Key to the feasibility of this reform is whether an hourly return trip can be done with one bus. Ways to allow this include running inbound trips via an express route (as done with Route 357 and 386) and/or not going the full route (as done with 788). 

This is a low income / high bus using catchment that would benefit from extended operating hours so a 541 upgrade is considered high priority.  

959 >> -

Much of this route parallels the Route 59 tram (which is not one of the six Night Network tram routes). From Airport West it follows the 902 SmartBus to Broadmeadows It is probably best kept as is unless the tram is upgraded. 

965 >> upgraded 683 & 685

Forming a big loop from Lilydale, the 965 is an unusual Night Network route. It's rare for an area to be so sparsely populated yet have Night Network service. If you were designing Night Network from the ground up other, more densely populated areas might get coverage before you'd consider a route here.

If you were going to do something it would be a 'bare bones' replacement, possibly parts of 683 and 685 operating to a two hour frequency. Like the 788, the 683 is a long route. It might just run between Lilydale station and Woori Yallock before turning back to form the next trip. 

Reform here could be in conjunction with the proposed revamped Healesville network which will simplify and add trips to Route 685. Some areas (including around Mt Evelyn) would lose coverage but they are relatively low population density and don't have particularly favourable demographics for buses. 

967 >> upgraded / rerouted 664 & 690?

967 starts at Glen Waverley Station, goes north to the 75 tram terminus, heads east through Knox and does a large clockwise loop, going as far north as Croydon. There is no single bus route that it approximates though it overlaps parts of several. The existence of 967 does however mean that a section of busy Scoresby Rd has more service at 2am on a Saturday or Sunday than at any other time. 

Like the Altona North and Glenroy areas, the City of Knox has a complex bus network with new routes laid over old routes. A revision of the local Night Network should ideally happen in concert with a reappraisal of regular routes. Two key regular network reforms could include: 

(i) Rerouting of 664 to operate via Scoresby Rd instead of Stud Rd. 

(ii) Extension of Route 690 from Boronia to Knox City (replacing portion of 753).

Both are relatively low cost and were discussed here

The thing about both of these reformed routes is that they would become approximations of the north-eastern part of Route 967. Thus they could supplant it. The main trade-off is some coverage loss, particularly near parts of Burwood Hwy. If a concern, that could be filled by running part of Route 732 as a Night Network route, possibly starting at the Vermont South tram terminus. 

978 / 979 >> ?

These routes serve a large area between Clayton and Dandenong. The area has excellent demographics for bus ridership but service doesn't reflect this. The regular bus network features many overlapping routes with few comprehensive network reviews being done. Almost all, except for some widely spaced SmartBuses, are only every 40 - 60 minutes, often with limited operating hours. 

No regular route is sufficiently like the Night Network 978 or 979 to adequately take their place. Highest priority is to review the regular network and consider which of the new reformed and simplified routes should take the place of 978 and 979 with 24 hour weekend service. 

Of all the areas in Melbourne, Greater Dandenong should be one of the highest priorities for local bus network reform on patronage and social equity grounds. The most recent attempt, involving the 813, has been a step forward but very half-hearted with many unreformed overlapping, infrequent and duplicative routes remaining. 

981 / 982 >> 828, 841, 888, 893?

A difficult area to reform. Night bus services comprise two routes between Dandenong and Cranbourne. Night Network route 982 is most like regular route 893 with the main exception being the former's coverage of Endeavour Hills. Route 981 is similar to Route 828 between Dandenong and Fountain Gate and other local routes south and east of there. Due to rapid population growth and suburban expansion more than two routes would be needed to retain reasonable coverage. 

Possible candidates for regular routes to be upgraded to Night Network include the following: 

* 893 - replacing 982 (may also desirable to operate a Dandenong - Endeavour Hills route 24 hr, possibly forming 841 listed later)

* 828 / 888 (Dandenong - Berwick as 828, then forms 888 at Berwick as through trip) - replacing 981   

* 841 (Fountain Gate - Cranbourne portion only) - replacing 981 

Route 893 and 828 are already main routes with a 20 minute interpeak weekday frequency. 888 and 841 are also main road routes but currently run at lower frequencies. 841 is quite direct but has areas with poor walking catchment to stops. Similar applies in parts of Endeavour Hills. 

A long overdue local bus network review may have scope to simplify local routes and create a main route hierarchy such as exists in Werribee/Tarneit where higher tier main road routes (eg 150, 180, 190) operate 24 hours on weekends as part of Night Network. Doing similar in Casey would make Night Network reform easier in this area.

Trains and trams too?

Not all after midnight services are part of Night Network. In about 2007 the operating hours of trains and trams were extended by an hour on Friday and Saturday nights. Thus instead of finishing around midnight services kept running until about 1am or later. These services were added on all Metro train lines (except Stony Point) and all tram lines. 

Night Network, introduced 2016, took a different tack. All train lines got Night Network services but only six tram routes did. 

There is a degree of confusion regarding public holidays. A Saturday timetable operates on public holidays except Good Friday and Christmas Day (where a Sunday timetable operates). Well it's not quite a Saturday timetable. It's a Saturday timetable without the after midnight trips added in 2007. Also Night Network does not operate unless the public holiday is on a Friday or Saturday (as Night Network ignores public holidays). 

There may be scope to tidy up these arrangements, with timetables identifying both sets of added services that don't run on non-Fri/non-Sat public holidays as 'Night Network'. This however introduces a differentiation for trams - with 6 routes having the full Night Network timetable, ie 24 hour weekend service. 

On the buses there's an added complication in that some bus routes that normally finish around 9 - 9:30pm have extra trips that operate on Friday and Saturday nights only. The latter do not operate on public holidays on most affected routes (eg 495) but there are exceptions like Route 460 where the holiday Saturday timetable is identical to the regular Saturday timetable. The result is that a public holiday like Labour Day has later trips running than it would if if it was a normal Monday - Thursday weekday. This contributes to the complexity of catching buses in Melbourne due to differing public holiday arrangements including the lack of universally applied timetable patterns. Ideally such routes (which tend to be amongst the most important in their areas) would have service until midnight on all days of the week with Night Network cutting in after midnight on many of these routes. 


Completing Night Network simplification by having more regular routes operate Night Network (instead of special routes) would be a worthy aim for bus service reform. This is because even if frequency is only hourly, having a 24 hour weekend service on more familiar routes is a major selling point on a network.  In many areas this would eliminate 3 or 4 hour gaps in service as currently exists between when the regular bus routes stop around 9pm and special Night Network buses start around 1am. 

Night Network services are generally quieter than services during the day. So you wouldn't spend a lot of money on this. But like fixing public holiday confusion and scrapping the last few remaining reduced Summer timetables, it's the sort of thing that a few million of dollars could pretty much solve. It also makes buses simpler and more saleable, especially if accompanied with local area network reforms and upgrades.   

As you can see from this run through some reforms are easier than others. In the harder cases local area network reform is needed to provide networks with high priority routes that you'd run 24 hours on weekends. Suggestions have been given where this applies. 

What are your thoughts? Are there other Night Network reforms not discussed here you'd like to see? Or maybe some here would be inefficient or cause excessive coverage gaps? Comments are invited and can be left below. 

Read other Useful Network items here

Friday, April 22, 2022

2022 federal election and public transport

Two Sundays ago Prime Minister Scott Morrison announced that the 2022 federal election will be held in just over four weeks, on Saturday May 21. 

With public transport largely a state matter, this election is less important for Melbourne public transport than the Victorian poll due in November. Especially for a blog that concentrates on service aspects. 

However federal elections still matter. Even for transport services, albeit indirectly. Federal governments collect most of our taxes. State governments do most of our spending. There are substantial federal - state grants for transport infrastructure projects that state governments couldn't fund on their own. 

Federal government of all stripes have enthusiastically funded roads. Their attitudes to urban public transport infrastructure has varied. The Howard and Abbott governments funded little if anything. Whereas the Keating, Rudd, Gillard, Turnbull and Morrison governments can point to funded public transport projects.

Federal funding of transport infrastructure can drive service improvements that would not have occurred otherwise. For instance the bus reform in Werribee and Geelong that followed Victoria's Regional Rail Link or on the Gold Coast following the tram's opening and extension. Bus service reform is possible without rail infrastructure but it wouldn't necessarily been as good or comprehensive. 

The role of infrastructure in stimulating service growth and reform is magnified where you've got a very infrastructure-focused state government such as Labor in Victoria. In such cases not having major infrastructure as an impetus can mean almost no service reform at all; if it was a gene bus reform would be as recessive as you can get. Even  here there's no guarantees, for instance the the new station at Southland or numerous suburban rail-road grade separations occurred without complementary bus network reform.  

Experiences from 2019 

Federal promises don't always get honoured. Projects that get federal funding need to be both feasible and supported by the state government. Otherwise they might not happen or be built where not needed. This has been the experience of some transport promises made before the 2019 federal election by the Liberal Party (which retained office despite previous unfavourable opinion polling). 

The most famous example has been the station car parks promised under the so-called Urban Congestion Fund. Promises have been made without suitable sites being identified, resulting in some projects being abandoned. The program was found by the Commonwealth Auditor-General to be a rort, with funding being funnelled to marginal seats rather than on a needs basis. And even if the latter was applied, park and ride is a very expensive way to boost train usage, costing tens of thousands of dollars per passenger gained. Whereas alternatives like bike paths and better buses are more cost-effective and deliver useful local access improvements that park & ride does not.  

Earmarked federal funding counts for nothing if Canberra only kicks in a fraction of what's needed and the state government isn't particularly interested. A recent example is Baxter rail electrification with potential benefits for a couple of federal seats then considered marginal. Canberra's $225 million is nowhere near enough for the state to jump at it. It remains to be seen whether this offer will be scrapped, reaffirmed or increased in 2022, noting likely increased skepticism given the lack of progress since last time. 

More optimistically we've seen revived interest in Melbourne Airport Rail which has long been wanted by frequent-flying business and opinion-writing elites for decades. While the rest of us might only occasionally use it, airport rail polls better than (say) Melton, Wyndham Vale or Clyde electrification as a good project to build. The latter lower profile projects may contribute more daily trips but suffer from their benefits being more geographically concentrated. Commentariat promoters of airport rail might also claim intangibles like 'being good for the state', 'fostering tourism' or similar. Unlike with Baxter, airport rail looks a goer; the feds' contribution ($5b) has been matched by the state with an announced start year of 2029.  

Victoria's battleground seats in 2022

See Antony Green's 2022 electoral pendulum here. According to this our only really marginal Liberal seat is Chisholm in Melbourne's middle-east (around Box Hill - Glen Waverley). This is on a 0.5% margin. The current member is Gladys Liu. 

A 5% anti-government swing would also see Casey and Deakin fall. Chisholm, Deakin and Casey form a contiguous finger stretching through Melbourne's east and beyond out to the Maroondah and Warburton highways. This is a bit like the state pattern where after the 2018 Victorian election the metropolitan marginal belt has swung to this corridor. A 6% swing would make La Trobe and Flinders vulnerable. These are again eastern and south-eastern Melbourne suburban and peri-urban fringe seats.

On the other side, Corangamite (Geelong fringe and rural districts) is Labor's weakest seat. A 1% swing would see them lose it. Labor's second weakest is the Frankston-based Dunkley with a 2.7% buffer. 

In practice swings are rarely uniform. This election is really 151 seat-by-seat elections with different dynamics in each seat. Seats that appear marginal may hardly swing while those not on the marginal list might return a new member with a double-digit swing. 

Notable in Melbourne are challenges in its wealthiest Liberal-held seats by so-called 'teal independents' campaigning on climate change and integrity issues. Seats that could come into play here include Kooyong (current member Josh Frydenberg) and Goldstein (current member Tim Wilson). Meanwhile, the incumbent independent Helen Haines will need to fend off a Liberal challenge in Indi, in north-east Victoria. 

More seat by seat analysis on The Tally Room

2022 public transport promises for Victoria

What will be promised in the next five weeks in the public transport sphere? The campaign is still young. And Victoria may not be seen to be as big a political battleground for marginal federal seats as other states such as NSW and Qld. However there may still be promises. 

Incumbent governments (whether state or federal) should be judged on their competent design, planning and management of existing programs (eg pork barrelled station car parks in dense suburbs some of which have been found to be unfeasible). Whereas non-government parties need to convince voters that their schemes are desirable, practical and affordable.

While concentrating on services, some of the seat-by-seat wish lists in my 7-part state election series do include rail infrastructure projects like Baxter and Clyde rail electrification. If these become part of federal campaigns and promises I'll note them below as details come to hand. 


Main policy page

* Melbourne Airport rail & faster Melbourne - Geelong rail (10 year infrastructure plan)


Main policy page


Main policy page

Electric car discount

* $2.2 bn for Suburban Rail Loop (only a small percentage of total cost)


Main policy page

Cheaper electric cars & phase out fossil fuel cars

* Public transport and high speed rail

(Palmer) United Australia Party

Main policy page

One Nation

Main policy page

Smaller parties and independents (including 'Voices for' and 'Climate 200' candidates)

See individual candidate websites and social media

Bicycle Network's list of promises by party

(more to come as further information becomes available)

Tuesday, April 19, 2022

TT #161: Network performance - how we're travelling

Let's take a break from looking at timetables to their adherence. After all whether your service is punctual or even shows up at all is an important part of passenger experience. 

For at least two decades passengers could check monthly performance for trains and trams. Indeed  there had to be posters on the system reporting it each month. This is tied to a contractual operational performance regime which fines operators and pays compensation to passengers (as travel credits) if delivery and punctuality fall short. 

Passenger compensation arrangements

The main game should be running a reliable service rather than the minutiae of compensation. Still it's  worth spending a moment to consider who can get it. The answer is that arrangements are imperfect and favour the private operator, eg passengers must formally apply rather than get an automatic myki credit. 

Also the archaic 28 day pass or longer rule makes few eligible, especially now with widespread working from home and less regular commuting. But even before the pandemic compensation  didn't go to those most affected. This is because compensation was available to higher income full time white collar CBD earners more likely to use periodical tickets (ie myki pass at least 28 days long). In contrast it was less available to lower income casual workers who are more likely to use myki 7 day pass or myki money, neither of which offer compensation entitlements. This is despite, due to their weaker position in the labour market, those in the latter group being more likely to lose pay or even their job if trains are consistently late or cancelled. 

No compensation arrangements apply for bus passengers. For years the only bus performance statistic appeared in the government's budget documents. This might show suspiciously high service delivery (like 99%) even though individual experiences with particular routes could vary greatly. In July 2021 the state government addressed this issue with per route bus performance data now being made available

What about when service gets really bad and disruptions are both network-wide and hit the headlines? Then the government may declare a free travel day open to everyone. It's been a while since we've had one in Victoria. However Sydney's recent major rail disruptions has led to them offering 12 days free travel as compensation in April 2022 (coinciding with school holidays).  

Where do you find performance data?

The short answer is that it's on the PTV website. The long answer is that it's more complicated. There's a lot of information but you need to be a bit of a data wonk to use it. And some PTV website links that should work have been broken for months. But once you find a working link you could be enjoying this data for hours. Let's dive in.

Data is presented via an interactive thing called PowerBI Dashboard. My first attempt to find it failed with the search ending in a broken link.  

This bug has existed for months (at least). Its longevity may indicate that DoT people (especially contract managers) don't use their website's performance reporting tools as much as they should as otherwise they'd see such errors. Or they do but there aren't good internal systems or cultures for identifying, reporting and correcting errors quickly. Either way the DoT could take a tip from the IT industry on the product quality benefits of dogfooding their public-facing work.

To properly find Power BI do the following to get a working link.  

Click here for a short-cut to Step 5. Success looks like the following entry page. 

Then click on the mode you want and drill down from there. You can look at various time periods, lines and routes. 

Trends by mode

I'll present graph by mode for the last 30 months. This time period allows you to compare performance now with pre-COVID times when peak patronage (and crowding) was higher. Later though rising case numbers and more people in isolation increased work absenteeism. This would have affected driver availability and thus service delivery. 

Metropolitan train

COVID saw a large improvement in punctuality (blue line) with greatly reduced patronage. For quite a while service delivery (dotted line) was somewhat better at around 99%. However this fell off a cliff in October 2021 and hasn't yet recovered. Metro's delivery has hovered around 97%, meaning about twice the number of trains were cancelled compared to the 98.5% performance target. The result is that passenger compensation has sometimes been payable but, as noted before, almost no one qualifies. 

Metro punctuality is a touch off COVID lockdown highs and remains better than early 2020 (pre-pandemic). But don't read too much into the graph above as early 2020 wasn't particularly good. This can be seen if you set a wider date range (easy to do with PowerBI) you will see a slow deterioration since about 2014. Current punctuality is about at 2013 levels (though with less patronage still).  

Metropolitan tram

Yarra Trams show a different pattern to Metro Trains. Their service delivery was poor in early 2020, dropping as low as 96.2%. However it improved during the pandemic and since. While February and March 2022 have exceeded the 98.5% delivery target some previous months did not. 

Because so much of our tram network is on-road, punctuality (solid line) is greatly affected by car traffic volumes. After reaching historic highs during our lockdown periods, tram punctuality is now the same as pre-pandemic. 

V/Line train

V/Line has gone through a rough trot lately. If a Metro train is cancelled the next is normally within 20 or 30 minutes. Whereas V/Line waits are double or more that. They're only required to run 96% of their trains (versus 98.5% for Metro) but they routinely fail even that. 

In September 2021 V/Line had a major COVID-related driver availability crisis explaining the big drop in that month. However the regional operator's woes started well before that. You can see that if you set the months window to 240 so that you see 20 years of data. Around 2003 it was not uncommon for V/Line to run 99.5% of its trains. That is just one in 200 services were cancelled. 

The gap between 100% and actual delivery has been widening ever since. It's true that there have been significant growth pressures with V/Line becoming largely an outer suburban rail operator with the rise of stations like Tarneit (which didn't exist before 2015). But even as recently as 2017 V/Line could meet its (not particularly tight) 96% delivery target in most months. Since then it's been rare for it to have a good month. 

V/Line's punctuality target is 92%. This was routinely exceeded before everything fell off a cliff in 2003 (Connex had similar problems then). Lockdowns and local travel restrictions allowed V/Line to reach 92% for the first time in years during the pandemic. Unlike service delivery V/Line's punctuality remains relatively high today with 93.1% recorded as late as February 2022 before dropping to 90.6% in March. 

As I wrote last week, performance reform in V/Line is sorely needed if we are to successfully deliver transport for the largely regionally-based Commonwealth Games in 2026. 

Metropolitan bus

Unlike the other modes there are three lines. The lines near the top are described as journeys allocated and journeys monitored. They average 99.9 and 99% respectively. The solid line shows punctuality over the last 30 months. Buses like trams are subject to delays due to road traffic volumes. You can see the broad trend where punctuality was highest when there were few cars on the road while we were locked down. 

More than other modes, the data presented is an aggregate of hundreds of different stories. Whether they be issues with driver availability, road works or schools being closed. With a bit of background knowledge you can answer specific questions like whether there were significant changes when Transdev lost to Kinetic, roads were closed in a particular area or a bus operator had a COVID issue at their depot. Hence I recommend spending time looking at individual routes to check for patterns. I looked at some bus stats last year in this thread


It's great that we now have performance stats for all modes. I strongly recommend delving into them. If you find interesting patterns in the data then please share them in the comments below. 

See more Timetable Tuesday items here

Thursday, April 14, 2022

UN 122: Six winning moves for 2026 Commonwealth Games transport

On Tuesday it was announced Victoria will be hosting the 2026 Commonwealth Games. The event will bring competitors, organisers and spectators to Melbourne and Victoria. To be held 20 years after our last hosting, we now have four years to prepare. 

The most successful cities use major events to build improvements that improve peoples lives long after the closing ceremony. Nowhere is this more obvious than in transport and housing. Yes, many of the improvements should have happened even if we didn't have major events. However there is nothing like an occasion with a deadline and international attention to create an urgency and enable funding that might not otherwise occur. 

It may or may not have been a coincidence, but our 2006 experience was instructive. In 2005 Melbourne had a ramshackle bus network that hardly ran weekends or after 7pm. It was badly wounded by big cuts under Cain/Kirner and left to atrophy under Kennett. The first five years of Bracks saw only a minor revival with outer suburbs getting some small improvements in 2002.  

To their credit those organising March 2006's Games knew that our existing public transport only worked for daytime CBD workers, shoppers and schoolchildren. With many events operating at nights and weekends much better was needed. So they boosted evening trains to run every 20 minutes and added night service to many bus routes. For at least a little while Melbourne had half-decent transport services. 

Two months after 2006's Melbourne Games came the Bracks government's Meeting our Transport Challenges. Not everything in it happened. But it did usher in four years of rising bus service levels including new SmartBus routes and numerous local buses upgraded to run at least hourly until 9pm seven days per week. While still not a particularly attractive frequency it transformed enough peoples' lives to greatly boost bus usage. 

Unfortunately for it, the (by then) Brumby government got little political credit as the bus upgrades were overshadowed by rail network failures for reasons discussed here. Much later some train lines got 20 minute evening frequencies, partly reversing large cuts made in 1978 (that we've never fully recovered from).

I have no way of telling but it is possible that the 2006 Games' success and the experience with evening buses might have encouraged some service upgrades to be made permanent. Will 2026's games leave a similar legacy?

A regional games

A major difference in 2026 will be that events will be held outside Melbourne at four hubs - Geelong, Ballarat, Bendigo and Gippsland. The four games villages built will be repurposed for social and affordable housing after the event. This makes proximity to good transport important both short and long term. In addition some events will be held at Shepparton. 

Notwithstanding the regional focus there will also be increased activity in Melbourne, especially the CBD area. That's not just because the opening ceremony will be held at the MCG. The CBD will be a useful accommodation base for those visiting multiple events. As well conference organisers may take advantage of the games period to schedule their events here (again likely in the CBD area). 

Melbourne CBD and suburban interchanges such as Deer Park, Sunshine and Footscray will also function as transport hubs, not least because of the network's centralised focus with limited cross-country transport options. Similar comments may also apply for access from Melbourne Airport. 

2006 Games venues had no spectator parking, with high reliance on public transport (which proved successful). This approach is a major way for events to demonstrate their 'green credentials'. If this is to happen at 2026's regional venues, currently with less public transport access, significant thought will need to be given to providing quality options to enable this. 

Transport in 2026

What transport projects currently under construction will be open in 2026? Metro Tunnel certainly will be. This will improve inner area connectivity. We'll also have more grade separations done in Melbourne suburbs. 

Melbourne Airport rail though won't be done. Hence we'll be relying on buses and coaches for public transport from Tullamarine for the Games. Avalon will want to to attract its share of Games traffic, especially for those going to Geelong.  Avalon has proposed improved connectivity to the Geelong line via a new station. That could happen by 2026 if approvals and financing is sorted quickly. 

Very important, given the dispersed nature of events, will be projects under the Regional Rail Revival umbrella. And others. Because if they get a wriggle on now there's a chance that other small and medium sized regional rail benefiting infrastructure projects could be brought forward to be finished by the Games' start. 

The six moves

We know which sports will be played in which cities. Or at least for Geelong, Ballarat and Bendigo. It's less clear where in Gippsland. This is a large region including the La Trobe Valley which itself contains several substantial towns, most of which are strung along the Bairnsdale line. Some venues will be near a railway but for others 'last mile' connectivity will be key. 

Here's six moves, to benefit transport across the state, that could be key to making this aspect of the 2026 Commonwealth Games a success.

Move 1: Strengthen Melbourne - regional Victoria links 

This is basically about the V/Line rail network, most notably lines to Geelong, Ballarat, Bendigo, Gippsland and to a lesser extent Shepparton. Services need to be faster, more frequent and more reliable. Infrastructure, rolling stock and operational constraints such as sharing with metropolitan services, single line sections and not enough trains need to be overcome to deliver the service quality required. 

More immediately timetables need to be simplified and the existing fleet worked harder to deliver higher frequencies all week, not just on weekdays. This is best done progressively in the next year or two so that any hitches can be ironed out well before the games. 

Network-strengthening initiatives include: 

- Higher train frequencies on existing lines all day / all week to reduce waiting times. Ideally every 20 min or better. As a first step weekend daytime frequencies should be upgraded to what applies during the week. This would assist much non-games travel, particularly from suburban stations on the Ballarat and Geelong lines. Work is currently under way to improve Shepparton services. Thought should also be given on how best to serve regional spectators coming from centres like as Warrnambool, Horsham, Ararat, Mildura and Wodonga to make these truly a state-wide games. The often forgotten Overland from Adelaide may also have a role but needs improved frequency. 

- Faster speeds. This can make rail travel competitive with driving including on trips that involve a change (see later). Some improvements may be possible through a tiered type service model that separates Wyndham Vale/Tarneit passengers from Geelong services and Melton/Caroline Springs from Ballarat services. Timetables could be more consistent with carefully timed expressing and fewer unique stopping patterns that just confuse people. 

- Internal change to deliver better performance from V/Line. V/Line has been dogged by scandal including a revolving door of CEOs who have departed under probity clouds. The regional rail operator has been the subject of two recent IBAC investigations (Operations Lansdown & Esperance).  And it has frequently missed performance targets (including in ten of the last twelve months). This is despite having looser service delivery targets than either Metro Trains or Yarra Trams. There is no room for leniency here; V/Line needs world-standard performance to play its part in hosting a world-standard games. 

- Targeted infrastructure upgrades to supported a desired improved capacity, speed, frequency and reliability. Some relief and better connections could be enabled by extending the electrified Werribee line to Wyndham Vale with a connectivity improvement for Geelong (a key Games venue). Duplicating the Bendigo line could assist service on that corridor. Possibly the most difficult to speed up is the Gippsland line due to its long overlap with suburban trains to Pakenham (and eventually Pakenham East). However a solution is needed as current V/Line rail frequencies and speeds are don't compare well with (say) a direct drive from Melbourne's south-eastern suburbs to Games venues. 

- Potential supporting V/Line coach service upgrades to support increased tourism in areas such as the Great Ocean Road, Phillip Island and Daylesford. The adequacy of bus services on the Bellarine and Mornington Peninsulas might also need to be looked at. Bicycle-linked regional tourism is not something that V/Line has encouraged but perhaps should as a means of extending its coverage and usefulness. 

Other ideas, concentrating on rail projects, are in this very good Twitter thread

Move 2: Better inter-regional links

Just like public transport investment and service provision in metropolitan Melbourne has suffered from a heavy emphasis on getting people to and from the CBD rather than between suburbs, the same holds true for the entire state. We have no V/Line train service that does not at some point run to the CBD. Travel between regional centres not on a radial train line is restricted to coaches and these are often infrequent. With a regionally-based multi-centric Games the transport network will need to be much more than this for it to be useful.  

Geelong - Ballarat

If you're driving from Geelong to Ballarat you don't go via Melbourne. But if you're taking public transport you might need to. That can waste kilometres and extend travel time, especially if services aren't particularly fast or frequent to start off with. There is a faster coach but due to its low frequency (three times on weekdays and not on weekends) you have to be lucky for it to suit your travel plans. Direct rail exists between the two centres but is not in regular passenger use. See the case study below for a comparison between transport options and possible improvements (click for a clearer view). 


Let's go to the other end of the state. It's very differently placed with regards to cross-regional transport. Unlike in western and northern Victoria, both geometry and topography mean that there is no reasonable connectivity to the other Games hubs except via Melbourne. Its main issue is travel speed with Gippsland trains being held up by the frequent suburban services with which they share tracks. There may be some things that could be done but dedicated tracks would be the gold-standard solution for fast and frequent service.  

Bendigo - Ballarat

These are north-north-west and north-west of Melbourne respectively. Their rails into Melbourne form an acute angle. In conjunction with operational practices (which do not stop Bendigo trains at Sunshine) this geometry means a lot of backtracking and often waiting (map below). 

Direct coaches can be faster but are infrequent. Travel times vary depending on route. For example the example above does Bendigo - Ballarat in under two hours. In contrast the once-daily coach that runs through to Geelong takes a little over three hours. Departures before 6am and three hour coach trips are not the sort of thing that is likely to entice day trips from anyone but the most enthusiastic visitors.  

What about rail options for these cross-country trips? There has been interest in running direct cross-trains from Geelong to Ballarat via the existing line. A sub 1 hour travel time would be ideal and allow (say) an hourly frequency with two trains. If you are going to run circumferential trains between two large regional cities, then the Geelong - Ballarat pair would be the front-runner.  

Ballarat - Bendigo rail is a weaker prospect with a lower combined population, longer distance and less directness (if going via Maryborough). Still concepts like these have got as far as Spring Street. For example a government investigation ten years ago under premier Baillieu following advocacy by the Rail Revival Alliance. Neither party has progressed it since and the report has gathered dust. Not even cheap to run coaches have seen significant service improvements.

To summarise, cross-regional travel between Victoria's largest regional centres has received little practical attention. Except possibly for Geelong - Ballarat, the best prospect for direct improvements is likely through upgraded (potentially express) coaches. However indirect improvements with wider benefits should be possible with rail infrastructure, speed, frequency and connectivity upgrades on existing lines. Ideas here: 

- Much higher 7 day frequency and profile for existing orbital coach services. These might even directly serve Games venues to reduce interchanging. Coach service upgrades to consider include Geelong - Ballarat, Ballarat - Bendigo and Bendigo - Shepparton. Castlemaine - Ballarat via Daylesford might be a possibility if it times well with trains to and from Bendigo. The Moe - Traralgon Route 1 bus parallels the radial railway but serves several intermediate destinations so might also need boosting with regards to frequency and operating hours. 

- Better interchanges and consistent stopping patterns that allow connections without backtracking. Sunshine will be the most important hub with it being a potential union between Geelong, Ballarat and Bendigo lines. Other interchanges such as Deer Park and Caulfield could assume a stronger network role to relieve stress on Sunshine and Southern Cross Station respectively.

- Improved radial V/Line links as per Move 1 above. Where possible timetables should consider key connections between regional lines likely to have high transfer volumes. Faster and more frequent trains to destinations like Geelong, Ballarat and Bendigo could slash rail travel times as shown in the example below.   

Move 3: Stronger Airport connections

Airport rail won't yet be running so there'll need to be some quality bus-based options for the Games. Direct coaches from Tullamarine to Geelong, Ballarat and Bendigo could work but they involve a lot of service kilometres to provide more than an occasional service. The most cost-effective airport connections likely involve fast shuttle buses to upgraded rail hubs serving Games destinations. For example:  

- Fast and frequent limited stop bus from Sunshine to Tullamarine. This can deliver large travel time savings over a wide area including parts of regional Victoria especially if introduced in conjunction with upgraded rail services serving a Sunshine 'superhub'. Main regional destinations to benefit include Geelong and Ballarat with Bendigo a possibility if its trains stopped at Sunshine. Sunshine will even be relatively accessible from Gippsland via a change to Metro Tunnel trains at Caulfield. To maximise benefits Sunshine should get all day / all week Metro Tunnel trains at least every 10 minutes to Caufield rather than suffer from having half the trains stopping short as mooted. 

- Avalon airport connectivity. No doubt this airport will seek to capitalise on the games, especially for Geelong based events where it is much closer than Tullamarine. Existing shuttle buses run but they are expensive and low frequency. The most cost-effective means of improving connectivity is a short link to Lara on the Geelong line. However as for previous moves this will need reliability and frequency upgrades on the V/Line rail network for it to be fully effective. Top priority (to happen almost immediately) should be an upgrade of Geelong weekend trains from every 40 to every 20 minutes. 

Move 4: Good local and 'last mile' connectivity 

Some venues may be near stations but not all will. Mars Stadium in Ballarat has a railway passing but no station. People may be willing to walk (say) 15 minutes but beyond that thought needs to be given to connections. As well as visitors coming off the train other travel needs include travel between the Games Village and the event site and spectators (who may be locals or visitors staying in the area). 

The best solutions will be site dependent but could include expanded bicycle and walking paths, service upgrades on existing bus routes or dedicated Games shuttles if there are no suitable regular services. A structural issue with most regional city bus services is their early evening finish time and limited weekend frequency. This may need upgrading during and (preferably) permanently after the Games. A precedent is White Night in Ballarat which saw extended hours bus services. 

Move 5: Major public and active transport network accessibility upgrades

A large component of Commonwealth (and Olympic) events is a 'para' component. As well as being discriminatory it would be highly embarrassing for a state hosting such events not to have accessible transport.

Accessibility challenges exist on all public transport modes but there is a particularly large backlog with our trams. We really need to accelerate stop upgrades and the roll-out of accessible trams. Such improvements would have many wider and lasting community benefits after the games.  More on tram accessibility in the context of a recent critical auditor-general's report here.  

Often ignored but no less necessary is on-street accessibility. It is no good having low floor trams, buses with accessible stops if you can't quickly, easily and safely reach your destination due to traffic volumes, the lack of pedestrian crossings, bad kerbing or no footpaths. A walkability audit followed by a program of perhaps 1000 small works in hosting centres would be highly desirable to ensure access is up to scratch. The same applies for short 'missing links' on the cycling network. Again this will catch up on a backlog and deliver lasting post-Games benefits. 

The Disability Resource Centre is currently campaigning for accessible transport ahead of this year's state election.

Move 6: Melbourne metropolitan service uplift

While most Games events will be in regional Victoria, there will still be a significant increase of visitor activity in Melbourne. This activity is likely to be both during and after regular business hours including weekends. Transport needs to suit the city we want to be and present to the world.

This includes no more 30 minute waits for trains and trams at night or the risk of getting trapped in the confusing vortex that is the midday reversing City Loop. Buses also need simpler, more direct and more frequent routes. This includes in the suburbs where limited bus services are restricting community sport participation. The latter is important if 2026 will be truly a 'peoples games'.  

Priorities for upgrades could include: 

- Greenfield Metro train timetables. To deliver consistent 7 day City Loop service, fewer stopping patterns, 20 min maximum waits and widespread 10 minute daytime frequencies. As much as possible to be done in 2023 - 2024 with the final stages to coincide with Metro Tunnel's opening (projected 2025). 

- Tram network and timetable revamp. Including maximum 10 minute waits on as many routes as possible for as much of the day as possible. Waits never more than 20 min on Sunday mornings and evenings. All routes go the same way 7 days. Subsequent reforms in conjunction with accessibility upgrades and Metro Tunnel commencement (most notably straighter routes, better western CBD / Southern Cross connectivity and fewer but individually more frequent routes on Swanston St). 

- Bus network reform for simpler more direct, faster and more frequent routes. Notably routes that would feed the Metro Tunnel, establish travel patterns for Melbourne Airport rail and deliver Suburban Rail Loop SmartBus. By the time the 2026 Games roll around it will have been over 15 years since Melbourne last added a SmartBus. However careful reform could deliver at least 10 new cost-effective SmartBus corridors for Melbourne between now and then. 

- Local suburban bus upgrades. Including route, operating hours and frequency upgrades that could assist travel for local sports participation (and many other purposes). Because community sports typically happens after school/work and on weekends a completion of minimum service standards (ie 7 day buses until at least 9pm) would be both desirable and economical without new bus purchases required. These upgrades should be skewed to outer areas with (i) low and middle income families currently feeling the pinch due to high petrol prices (eg outer west, north and south-east) and (ii) established area low income middle aged and senior people that often get passed over for bus service improvements (eg Glenroy, Thomastown, Mornington Peninsula). 


The 2026 Commonwealth Games could provide an impetus for improved public transport in Victoria. Not just for its own sake but for an enduring community benefits. I've outlined six major moves that could together transform how we get around. It is not exhaustive and there are likely to be further opportunities from rail infrastructure to ferries to regional tourist trams not covered here. 

Comments on these and other possibilities are welcome and can be left below. 

More Useful Network items here

Tuesday, April 12, 2022

TT #160: Bus 841 - Casey's vital link

Running from Cranbourne to Narre Warren North via Fountain Gate, the 841 bus is one of the City of Casey’s major bus routes. It runs almost exactly north-south with small deviations to serve stations and shopping centres. 

Route 841’s catchment is about 40km south-east of Melbourne. It is predominantly 1980s – 2000s detached housing built on curvy streets within a mile grid of north-south major roads built for the convenience of cars but the inconvenience of walkers. The 841, along with the 893 to the west and the 888 to the east, feeds catchments off these roads to train stations on the Pakenham and (in some cases) the Cranbourne train lines.   

You can see how the 841 fits in with the broader network below. It is fairly simple and straight unlike most other nearby routes like 894, 895 and the circular 834/835 pair.  

Route 841’s strongest destinations include Cranbourne Shopping Centre, Cranbourne Station, Narre Warren Station and Fountain Gate. Its northern terminus is weak with a residential area terminus. To be fair though it would need to be extended through a low density area before it could be extended to a possible stronger northern terminus such as Endeavour Hills Shopping Centre.  

The 841 bus serves several state electorates including Narre Warren North (Luke Donnellan - Labor - not recontesting), Narre Warren South (Gary Maas - Labor) and Cranbourne (Pauline Richards - Labor). Some boundaries will change for this year's state election. For more information see my 7 part seat by seat 2022 state election special. Route 841 is mentioned here as a prospect for an electorally significant service upgrade. 


Route 841 is your standard Melbourne suburban bus route operating seven days per week to approximately 9pm. Weekday frequencies are approximately every 30 minutes peak and 40 minutes off-peak. Weekend frequencies are roughly 40 minutes Saturday day, 60 minutes Saturday evening and 60 – 90 minutes Sunday. The existence of >60 minute headways means that the 841 does not meet hourly minimum service standards on Sundays.    

Compared with routes either side the 841 has much inferior service to the 893 (every 20 minutes all week) but usually better service than the always hourly 888. The 841 though has the by far the biggest mid-route trip generator of all three routes with it servicing the busy Fountain Gate Shopping Centre and the nearby civic centre. 


Route 841 is an above average patronage performer with boardings per km about 30-40% above that for the average bus route in Melbourne. More precisely it attracts 31 boardings per hour on weekdays and 22 boardings per hour on both Saturday and Sunday. School holiday weekdays are only very slightly less than school days, indicating the 841 enjoys strong usage from other sources. 

Route 841’s relatively good patronage is despite the roads it runs on which often offer poor pedestrian connectivity from residential streets near it and too few crossing opportunities. Its usage shows that many people will walk a bit to direct bus services. This is particularly if they serve a major destination and run more frequently than average. 


The northern half of the current 841 can trace its history back to 1987 in a shorter form. The 1992 Melbourne network map shows its southern portion operating as a route 840. Later these were amalgamated into a longer 841. Still later it gained 7 day service and the portion south of Narre Warren station straightened as additional routes were added to the area. 


Route 841 already has quite good usage but is there scope for it to perform better and be given a stronger northern terminus? For example should it continue to Endeavour Hills or even Dandenong to replace the 843 and 861? Does it deserve a 7-day 20 minute frequency (like 893) given its potential usefulness as a Narre Warren and Fountain Gate feeder? Would a southern expansion to Frankston to incorporate 791 to form an outer orbital SmartBus make more sense or would this reduce reliability? What do you think? Comments are invited and can be left below.