TT #177: World class or world worst? Comparing Melbourne's evening train frequencies

Night Network, introduced in 2016, is the most recent major service initiative on Melbourne's trains. While most large cities have long used buses to run a skeleton after midnight network, Melbourne was one of the first to operate 24/7 weekend service on all its three modes of transport, including all  electrified rail lines. We were shortly followed by London which introduced overnight service on some Tube lines. Further improvements came in 2021 with 24/7 weekend service on regular bus routes replacing less known special routes in many areas. 

7pm to midnight service

Of much lower profile is the rail service that runs before Night Network starts - say 7 pm to midnight. These trips run 7 days per week and carry many times more passengers than Night Network does (or will). Many Night Network users, including workers as much as revellers, would be riding trains at these times to reach their destination before returning on a Night Network service. 

7pm to midnight transport services are also important for the night-time economy and various sporting, cultural and entertainment events. These vary in scale from major events like the Australian Open to a couple's night at the theatre. Without effective public transport operating at these times the CBD becomes strictly a daytime employment and business precinct with few lingering later than the pm 'rush hour'. 

Comparison with other cities

Melburnians see themselves as cosmopolitan and well-travelled. Most have at least visited and often lived in major cities either interstate or overseas. It's not surprising then that we like to compare ourselves and our cities with others. Not least our civic leaders, who quickly take credit if we top liveability rankings against other major cities. In years where the results are not as hoped we can at least go back to comparing ourselves with Sydney.  

In this spirit let's do this for our evening metropolitan train services. Is our mid-evening service as pacesetting as our Night Network? Or is it more that of a mid-tier city, like a Brisbane, Perth or Manchester?     

To find out I compared train timetables across the world. I chose well-known cities with which we might compare ourselves. To fit it on a simple graph I looked at service frequency around 10pm Saturday. That graph is below, with taller lines representing more trains per hour than shorter lines.  Not all cities run all lines the same frequency but what's shown is representative.    

The answer could hardly be clearer. Melbourne has about the longest waits for evening trains of any similar developed world city. Our typical 30 minute interval between evening (Metro) trains is double Sydney's 15 minutes and triple the 10 minute waits of London, Chicago and Stockholm. 

Canadian cities like Toronto and Vancouver have smaller (and more metro-style) networks but beat us even more soundly on frequency. Even much smaller Calgary runs a 16 minute evening service on its CTrain. Yet you'd be hard pressed to find cities more similar to one another on opposite sides of the world as those in Canada and those in Australia. 

V/Line rates even lower, despite stations like Ardeer and Deer Park almost qualifying as middle suburbs which would have a Metro service if in the east. 

History of evening train service levels & Sydney comparison 

Melbourne has approximately doubled in population since the 1970s. We were also a different society then with much more rigid working and shopping hours. Everyone aspired to be home in time for the 6 or 6:30pm TV news. 

Have its evening train frequencies kept up? At least on Saturday evenings the answer is no. Indeed they've gone backwards. Prior to the 1978 timetable cuts (whose hand continues to govern how long Melburnians wait for trains nearly 50 years later) typical Saturday evening train frequencies in built up areas was 20 minutes. The 1978 cuts reduced evening service to every 30 minutes at which it has remained ever since on most established area lines. Weeknights are slightly better but 30 minute service remains most common (2019 comparison here).  

The Melbourne/Sydney disparity became a gulf when Sydney added 1500 more weekly trips in its revamped 2017 timetable . It got even wider after its frequent Metro opened in 2019. The 2017 timetable boosted many evening services from every 30 to every 15 min while the Metro enjoys a 10 minute evening service. Consequently it is now fair to say that Sydney enjoys about double Melbourne's off-peak frequency, especially at night and Sunday morning.  

The widening gap is because in the last 5 or 10 years Sydney added both infrastructure and service  to its railway while Melbourne has been serious about infrastructure only. A sign that this might be starting to become more balanced occurred in 2021 when Werribee, Williamstown and Frankston received evening upgrades to every 20 minutes. This is high by our standards but low service by almost everyone else's standard (including Sydney). 2022 was basically a lost year for off-peak train timetable boosts and we don't yet know if 2023 will bring better news.   

The high costs of skimping on service

For all the electrification extensions, new trains and level crossing removals successive state governments can claim as achievements, one thing has remained pretty constant. On Saturday (and other) nights all but a few trains remain locked away in stabling despite proven demand for travel and long queues for taxis.

Meanwhile our system remains staffed with signallers, customer service staff and PSOs whose numbers are largely independent of whether we see one train per hour (Night Network), two trains per hour (typical evening service) or a frequent service as we (mostly) enjoy in the peaks but not, unlike more advanced cities, at other times on most lines. 

The combination of long operating hours (especially since Night Network started) and mostly low frequency has meant that our rail assets are underused, patronage is less than what it should be and operating costs are spread over fewer passengers. 

Railways that have been better at maintaining their patronage since the pandemic have been more broadly useful services rather than narrow commuter operations. To change our system into the former requires a funding and operating culture of 'all day frequent service'. 

'Getting to Twenty' A cheap partial fix

I discuss the most cost-effective moves towards better evening service here. The take-home message is how few extra trains need to be added to the weekly timetable to make an impact. Even a 1% uplift would be worthwhile with really significant gains possible with 3% more trips. 

For example boosting service from every 30 to every 20 minutes requires just two extra return trips per two hour period per line. This has a small marginal cost compared to many other things we have been willing to fund, including Night Network itself, but would likely induce similar if not more usage. 

20 minute maximum waits are still only Atlanta or Los Angeles standards but would at least harmonise with trams and narrow the gap with Sydney. But they could be a prelude to widespread 10 minute all-week service until midnight.

Ultimately this should be almost network-wide. A good start would be 18 hour/7 day 10 minute trains to busier line termini like Dandenong, Watergardens, Ringwood, Craigieburn, Werribee and Mernda.  Then you'd do Greensborough, Frankston, Wyndham Vale, Sandringham, Glen Waverley,  Melton and Upfield to basically complete the job. Then, and only then, would we have a genuinely world-class network.  

See other Timetable Tuesday items here 

UN140: Level crossings gone AND better service - having both in Narre Warren / Beaconsfield

Level crossing removals in Melbourne have typically been done as stand-alone infrastructure-only projects. This is even where new stations and bus interchanges were built as part of the project and level crossing free roads could have unlocked network gains like reformed more useful and efficient bus routes. The same has also applied for some recent station additions.

We effectively have two-speed planning in transport. Infrastructure-based authorities and projects get the resources and profile while the network planning, operations and service areas (typically core areas of the DoTP) sit mouldering in the shadows, with backlogs growing under their feet. 

Examples of service lagging infrastructure aren't hard to find. Cardinia Rd station near Pakenham didn't get the dead-end bus route 925 extended to it. Southland (opened 2017) missed out on local bus reform that would have finally connected several routes to the Frankston line. The now grade separated Centre Rd in Clayton East has a Night Network but not a regular day bus even though such a route would have provided a faster Clayton Station connection for Waverley Gardens. Timetable reform is also lagging with buses to the rebuilt Reservoir station remaining at every ~22 minutes trying (and mostly failing) to meet trains every 20 minutes. 

The result of our failure to integrate infrastructure with service planning is that many level crossing removals have not delivered all benefits they could have, especially for bus passengers. On this we lag cities like Perth which routinely reforms buses when it adds new stations. And we've done well before, with the high point reached in 2015 when Regional Rail Link also saw dramatically reformed bus networks in Geelong and Wyndham.  

The solution is to (i) fix the multi-decade backlog of bus reform, including around recent infrastructure projects, and (ii) ensure that new infrastructure projects, including level crossing removals and new station builds, get bus reform done in concert with them. 

Even where the area really needs a major network overhaul, opportunities for smaller scale improvements to address the most pressing service shortfalls around the project location should not be passed up. 

Beneficial examples could include timetable uplifts eg 7 day service (where lacking), longer operating hours, extra weekend trips and/or frequency harmonisation with trains. Minor route changes including short extensions to dead end routes that currently stop short of the nearest train station and network reforms for simpler routes or better coverage are other possibilities.

All would aid connectivity, improve the effectiveness of stations as transport hubs and broaden the benefits of level crossing removals if included within their project scope. The latter would be especially beneficial where projects have been somewhat controversial, such as those involving station consolidations or the construction of elevated rail.  

In 2021 the government announced 10 more level crossings to be removed by 2025. That would increase the number removed from 75 to 85 since Labor took office in 2014. Then, before the 2022 state election, Labor promised the removal of 25 more by 2030, taking the total to 110.

With the Andrews government re-elected this means that there is a lot of opportunity to integrate bus upgrades with level crossing removals and station rebuilds in the next few years. 

Webb St, Narre Warren

One of the removals on the agenda for 2025 is Webb St, Narre Warren. This project will involve elevating the rail and rebuilding the station. Narre Warren is a major station on the Pakenham line with many bus routes passing or terminating. It is also the nearest station to the large Fountain Gate Shopping Centre, the civic centre and Bunjil Place. Narre Warren Station abuts growth areas to the south and, unlike Cranbourne, hasn't had a full bus network review for many years. Thus if there was ever a place that deserved at least some bus service reform in conjunction with the level crossing removal, Narre Warren, the subject of today's Useful Network feature, would be it. 

Existing transport services at Narre Warren 

Trains at Narre Warren typically operate every 20 minutes during the day and 20 to 30 minutes at night. Peak frequency is higher. Waits between Sunday morning trains are as much as 60 minutes. This is amongst the lowest service on the network and one-third the 20 minutes enjoyed by lines like Frankston and Werribee which received upgrades in early 2021.  

Surrounding bus routes are best described as complex and infrequent. Part, though not all, of the complexity can be attributed to the indirect street networks of suburbs north and south. Past reforms have often layered new routes over existing routes without serious attempt at taking a wider network view. There are some coverage gaps north of Princes Fwy with some people having to backtrack to Berwick rather than be able to get a bus to Narre Warren Station. 

As for frequency, just one bus route in Narre Warren operates better than every 20 minutes and even that is just on weekdays during the day (Route 828). Other routes are typically only every 40 to 60 minutes at most times.  

Network issues are summarised on the map below:

Priority service upgrades for Narre Warren

The following service initiatives could address the above issues at Narre Warren:
 

Train

* Sunday morning train frequency boost from 60 minutes (either 40 min for combined 20 min service from Dandenong inward with Cranbourne line) or 30 min or 20 min 

* In conjunction with Dandenong line upgrade boost weekend evening service from 30 to 20 min

The above would have wider benefits than just Narre Warren with gains at multiple stations on Melbourne's busiest rail line. These boosts could be a prelude to further improvements to be delivered as part of the Metro Tunnel. We made a great start in early 2021 when we gave similar boosts for the Frankston, Williamstown and Werribee lines but seem to have dropped the ball since, despite these upgrades being cheap, high priority and involving relatively few added trips.  

Bus

* Upgrade Route 841 frequency to every 20 min 7 days (at least Cranbourne - Fountain Gate portion) with longer operating hours. 

Route 841 is the major north-south bus serving Narre Warren Station, linking it to major destinations and interchanges including Fountain Gate Shopping Centre and Cranbourne. Route 893, a similar north-south bus route through the rebuilt Hallam Station, already operates at the suggested 20 minute 7 day frequency.  More on Route 841. 

* Boost 828 weekend frequency from 40-60 min to every 20 min with longer operating hours. 

While it doesn't pull in to Narre Warren station, it passes nearby. Route 828 is the area's major east-west route serving locations such as Keysborough, Doveton, Fountain Gate and Berwick's main street away from the train line. Its weekday 20 minute frequency means that upgrades can be done by working the existing fleet harder for more of the week. I discussed the 828 corridor (along with other key routes) here.

* Extend Route 839 from currently weak terminus to Narre Warren Station to lessen backtracking. Investigate options for running via Ryelands Dr for improved coverage. 

Route 839 currently runs west from Berwick but ends at a local shopping centre rather than continuing onto key interchanges or destinations like Narre Warren Station or Fountain Gate. Weak termini reduce the patronage potential of a bus route and create unnecessary need to change. It may be possible for the extension to serve an area currently with poor coverage. 

* Buy or find an extra bus or two so that all Route 894 and 895 trips can go the full route. 

Currently both routes have to skip stops during peak times with 3 hour gaps, making catching buses chancy. Addressing this could be a prelude to subsequent network reforms as local routes are currently very complex, indirect and miss key destinations including jobs-rich Dandenong South and Fountain Gate from some areas. Information and background on this is here, here and here. Reform could complement that already budgeted in 2022 for other Narre Warren and Cranbourne West local routes.   

* Ensure interchange design at Narre Warren maximises efficient through running and local travel for buses with provision made for fast and direct tram-like frequent routes. 

Currently there's a planning attitude that sees stations as bus route termini where everyone alights so it's fine to have buses weave off the main road in to a side street near the station. This can work at end termini where there are few other destinations and those that exist are grouped tightly around the station. However Narre Warren's main destination (the area around Fountain Gate) is away from the station. Bus interchange design at Narre Warren Station needs to support a network topology where most (if not all) bus routes from the south continue north to Fountain Gate. And stop locations facilitate a simple frequent service to it from the station. Interchange designs should minimise (and preferably eliminate) bus turning movements for fastest run times. Hallam is a very good example (for routes like 893) while Clayton is a poor example (with routes like 703 and 733 being made indirect northbound). 

The above service upgrades would have wider benefits as well, involving catchments of tens to hundreds of thousands of people. They could be the start of a staged program of bus network reform in the area. This is particularly needed here as it's been a long time since the Narre Warren bus network was comprehensively reviewed (unlike Cranbourne which got a completely reformed network in 2016).  

Beaconsfield, too

Station St Beaconsfield is also getting its level crossing removed through the construction of a road bridge. The area has traditionally been less densely populated than Narre Warren/Berwick to the west. However there is substantial new suburban growth at Officer to the east. 

The abovementioned train service upgrades for Narre Warren would also benefit Beaconsfield. As for buses, currently only Route 836 directly serves the station though routes 837, 838 and 926 pass nearby on the highway. Route 846 goes more than half way to Beaconsfield from Berwick but then heads west after going south. Local routes in the area tend to be short. Notably 831 and 837. Both have dead end termini that are begging to be extended.  

Bus reform in the area needs to consider the growth of surrounding areas including Pakenham and Officer. A need not currently met is a frequency hierarchy with much higher frequencies on main roads. The most notable case is Princes Hwy whose Route 926 generally operates every 60 - 75 minutes rather than the 20 minute service available on some other parts of Princes Hwy. Reform for Pakenham North routes are discussed here and Pakenham South here.  

Until a comprehensive Cranbourne-style review is done the highest priority upgrade to local buses would be a frequency boost for Route 926, added coverage in the Officer area and extensions to some 'dead-end' routes. However I wouldn't necessarily have many (if any) bus routes terminate at Beaconsfield as it is unlikely to ever be as significant a bus hub as Officer, Berwick or Narre Warren.  

Conclusion

Sequencing train and bus network and service upgrades with level crossing removals allows the delivery of integrated transport upgrades to communities with benefits for everyone - not just car drivers. The latter is important as in some previous projects passengers have borne the brunt of disruptions but had the fewest direct benefits from completion. 

Today I discussed potential service upgrades for the Narre Warren and Beaconsfield areas. Let me know what you think in the comments below. I'll return next week with a similar item on another level crossing removal.  

More Building Melbourne's Useful Network items here

ChatGPT guest item: A budget submission for better train services

On Tuesday I used ChatGPT to write about how you would improve buses in Melbourne.

I thought it did a pretty good job. I later tried other uses like writing haiku, limericks, proposals for TV shows and ideas for community arts projects. Again quite good results. Where it tripped up on was detail - on that it made stuff up with factual errors. You would certainly not trust ChatGPT to write something without being reviewed by a human. 

Below is another job I tried with it. Given we have a state budget in May, I asked it to draft a plan for  boosting Metro train frequency in Melbourne. 

Q. Write a budget submission for Metro train frequency to be upgraded from a 40 min maximum wait to every 20 minutes or better across Melbourne.

A. Introduction:

This budget submission proposes the upgrade of metro train frequency from a maximum wait time of 40 minutes to a maximum wait time of 20 minutes or better across Melbourne. This upgrade is necessary to improve the overall public transport experience for commuters and to support the continued growth of the city.

Background:

Melbourne's population is projected to continue growing in the coming years, with estimates predicting an additional 1.5 million residents by 2040. This increase in population will put added pressure on the city's existing public transport infrastructure, and it is essential that measures are taken to improve the efficiency and reliability of the system to meet the needs of commuters.

Costs:

The cost of upgrading metro train frequency will include the purchase of additional trains and the necessary infrastructure to support them, including tracks, signals, and power systems. These costs will be offset by increased revenue from higher ridership and improved accessibility for commuters.

Implementation:

The upgrade will be implemented in phases, starting with the busiest lines and then expanding to other lines as resources become available. The project will be closely monitored to ensure that it is on schedule and within budget.

Conclusion:

Upgrading metro train frequency is a necessary step in ensuring that Melbourne's public transport system is able to meet the needs of its growing population. The investment in additional trains and infrastructure will pay for itself in increased revenue and improved accessibility for commuters. We recommend that this proposal be given serious consideration in the upcoming budget.

It's light on the detail but I think it did well. And a reminder that none of the above was written by or edited by me. The main thing wrong is it fell for the widespread misconception that you needed to buy more trains or add a lot of infrastructure to boost off-peak service. 

Next I tried to see if it could write an implementation plan. I gave some clues on some planning tasks required (it won't be exhaustive). 

Q. Write a staged implementation plan for improving Melbourne metro train services on key lines from every 30-40 to every 20 min. Start with cheapest upgrades like 7 - 10am Sunday mornings and early evenings 7 to 10pm. Tasks include rescheduling train maintenance, rerostering and recruiting drivers, producing fleet stabling plans, rewriting timetables and passenger information. Aim for the first round of upgrades to be in 2024 with subsequent stages every 6 months after that. 

A1. Implementation Plan:

Stage 1: (2023-2024)

Reschedule train maintenance to minimize disruptions during peak hours

Reroster and recruit additional train drivers to increase the number of available trains during peak hours

Produce fleet stabling plans to ensure that trains are in the right place at the right time

Rewrite timetables and passenger information to reflect the new schedule

Implement the first round of upgrades, starting with the cheapest options such as 7 - 10am Sunday mornings and early evenings 7 to 10pm

Stage 2: (2024-Q1)

Assess the effectiveness of the first round of upgrades and make any necessary adjustments

Begin the next round of upgrades, focusing on increasing the frequency of trains during peak hours

Continue to reroster and recruit additional train drivers as needed

Update fleet stabling plans to reflect the increased frequency of trains

Update passenger information to reflect the new schedule

Stage 3: (2024-Q2)

Continue to monitor the effectiveness of the upgrades and make any necessary adjustments

Implement additional upgrades as needed

Continue to reroster and recruit additional train drivers as needed

Update fleet stabling plans to reflect the increased frequency of trains

Update passenger information to reflect the new schedule

Stage 4: (2024-Q3)

Continue to monitor the effectiveness of the upgrades and make any necessary adjustments

Implement additional upgrades as needed

Continue to reroster and recruit additional train drivers as needed

Update fleet stabling plans to reflect the increased frequency of trains

Update passenger information to reflect the new schedule

Subsequent stages:

Continuously monitor and evaluate the effectiveness of the upgrades and make any necessary adjustments

Implement additional upgrades as needed

Continue to reroster and recruit additional train drivers as needed

Update fleet stabling plans to reflect the increased frequency of trains

Update passenger information to reflect the new schedule

Repeat the process every 6 months.

Note: This plan is just a general guide, actual implementation will depend on various factors and may require additional steps or actions.

I added underlining of stages for clarity but otherwise above is verbatim. If you didn't like it here's an alternative. 

A2. Implementation Plan for Improving Melbourne Metro Train Services:

Phase 1 (2024):

Reschedule train maintenance to minimize disruptions during peak hours

Reroster and recruit drivers to increase the number of available personnel

Produce fleet stabling plans to optimize the use of available trains

Rewrite timetables and passenger information to reflect the new schedule

Focus on the cheapest upgrades first, such as increasing frequency on key lines during 7 - 10am Sunday mornings and early evenings 7 to 10pm.

Phase 2 (2024 - 6 months after Phase 1):

Expand the frequency upgrades to other off-peak hours, such as mid-day and late evenings

Continue rerostering and recruiting drivers to ensure a sufficient number of personnel

Continue to optimize the use of available trains through fleet stabling plans

Update passenger information to reflect the new schedule

Phase 3 (2025 - 6 months after Phase 2):

Further increase the frequency of trains during peak hours

Continue to optimize the use of available personnel and trains

Update passenger information to reflect the new schedule

Phase 4 (2025 - 6 months after Phase 3):

Finalize the schedule and make any necessary adjustments

Continue to monitor and improve the service to ensure that train frequency is maintained at every 20 minutes during peak hours.

Note: This is a general plan, it's important to conduct a detailed analysis of the current infrastructure, maintenance, staff and budget constraints to create a more accurate plan.

Next I kept the above request but gave some hints as to a longer term vision these upgrades would be a step towards. Here's the first answer it came up with. 

Q. Write a staged implementation plan for improving Melbourne metro train services on key lines from every 30-40 to every 20 min. Start with cheapest upgrades like 7 - 10am Sunday mornings and early evenings 7 to 10pm. Tasks include rescheduling train maintenance, rerostering and recruiting drivers, producing fleet stabling plans, rewriting timetables and passenger information. Aim for the first round of upgrades to be in 2024 with subsequent stages every 6 months after that. Bear in mind that this is a prelude to a longer term frequent network vision which is all day trains, trams and buses on main lines to be every 10 min or better from about 2026.

Implementation Plan for Improving Melbourne Metro Train Services:

Phase 1 (2024):

Conduct a detailed analysis of the current infrastructure, maintenance, staff, and budget constraints to create a more accurate plan.

Reschedule train maintenance to minimize disruptions during peak hours and off-peak hours

Reroster and recruit drivers to increase the number of available personnel

Produce fleet stabling plans to optimize the use of available trains

Rewrite timetables and passenger information to reflect the new schedule

Focus on the cheapest upgrades first, such as increasing frequency on key lines during 7 - 10am Sunday mornings and early evenings 7 to 10pm.

Phase 2 (2024 - 6 months after Phase 1):

Expand the frequency upgrades to other off-peak hours, such as mid-day and late evenings

Continue rerostering and recruiting drivers to ensure a sufficient number of personnel

Continue to optimize the use of available trains through fleet stabling plans

Update passenger information to reflect the new schedule

Phase 3 (2025 - 6 months after Phase 2):

Further increase the frequency of trains during peak hours

Develop plans for the integration of all-day services, such as trams and buses on main lines, to be every 10 min or better by 2026.

Continue to optimize the use of available personnel and trains

Update passenger information to reflect the new schedule

Phase 4 (2025 - 6 months after Phase 3):

Finalize the schedule and make any necessary adjustments

Continue to monitor and improve the service to ensure that train frequency is maintained at every 20 minutes during peak hours.

Begin implementation of the all-day services, such as trams and buses on main lines, to be every 10 min or better by 2026.

Note: The overall goal is to build towards a frequent network vision of all day, trams and buses on main lines to be every 10 min or better by 2026. This plan is a prelude to the long-term vision and will be adjusted accordingly as more information and resources become available.

Anyway that's it from ChatGPT. As I said lots on generalities but it needs help with specifics. Back to my own writing in future weeks. 

How to improve bus services in Melbourne, Australia (Guest writer)

Q. How to improve bus services in Melbourne, Australia?

A. There are several ways to improve bus services in Melbourne, Australia. Some potential strategies include:

1. Increasing the frequency of bus routes: This can make it more convenient for people to use the bus system, as they don't have to wait as long for the next bus to arrive.

2. Improving bus route coverage: Expanding the reach of bus routes to more areas of the city can make it more accessible for people to use the bus system.

3. Investing in new and upgraded buses: Newer and more comfortable buses can attract more riders.

4. Incorporating technology to improve the customer experience: For example, real-time bus tracking, mobile ticketing, and better information systems can make it easier for people to plan their trips and find the bus they need.

5. Encourage the use of public transportation by providing discounts or other incentives.

6. Investing in bus-priority infrastructure: This can include dedicated bus lanes, traffic signal priority, and bus-only ramps and shoulders, which can help to reduce travel times and increase reliability.

7. Improving the integration of the bus system with other modes of transportation, such as trains and trams, can make it more convenient for people to make connections and complete their trips.

8. Improving the bus stop infrastructure, which includes shelter, seating, lighting, and real-time information displays.

9. Reviewing and optimizing bus routes, which can help to reduce travel times and increase service reliability.

10. Improving the communication between the transport authority and the public, which can help to build trust and encourage more people to use the bus system.

If you thought I wrote that you'd be wrong. 

Instead it is the unedited result of posing the question in ChatGPT an online app using artificial intelligence to give answers. I liked the above answer. It was the easiest and quickest blog post I've ever 'done'! 

But if you didn't like that answer then you can hit the 'regenerate' button for alternatives. Here's one. 

A. There are several ways to improve bus services in Melbourne, Australia, including:

1. Increasing frequency of bus services: By running buses more often, passengers are less likely to have to wait long periods of time for a bus to arrive.

2. Improving bus route planning: By optimizing bus routes to better serve high-traffic areas and popular destinations, buses can be more efficient and reduce travel times for passengers.

3. Enhancing bus stop amenities: By providing bus stops with seating, shelters, and real-time information displays, the wait for a bus can be more comfortable and convenient for passengers.

4. Investing in new technology: By incorporating technology such as GPS tracking and real-time information systems, bus services can be made more efficient and reliable.

5. Encouraging use of public transportation: By making bus services more affordable, convenient and reliable, more people will choose to use public transportation, reducing traffic congestion and air pollution.

6. Improving public transport integration: By coordinating bus services with other forms of public transportation, such as trains and trams, passengers can more easily make connections and reach their destinations.

7. Investing in bus priority measures: By implementing bus priority measures such as bus lanes, traffic signal priority and queue jumping, buses can move more efficiently through traffic.

8. Improving customer service: By providing friendly and helpful customer service, passengers will be more likely to use the bus services and recommend them to others.

And another:

A. There are several ways to improve bus services in Melbourne, Australia:

1. Increase frequency of buses on popular routes: This will reduce wait times for passengers and make the service more convenient.

2. Expand the bus network: Adding new routes in areas that are currently underserved will make the service more accessible to more people.

3. Improve bus stop infrastructure: Providing better shelters, seating, and lighting at bus stops will make the service more comfortable for passengers.

4. Invest in technology: Implementing real-time tracking of buses and mobile ticketing can make the service more efficient and convenient for passengers.

5. Improve coordination with other forms of transportation: By coordinating bus schedules with trains and trams, the service can be made more convenient for passengers making connections.

6. Encourage use of public transportation: Government could provide incentives or subsidies for people to use public transportation and discourage the use of private cars.

7. Improve customer service: Regular feedback and complaint handling, training of staff to handle different situations and better communication with the public can enhance the overall experience of the service.

Also pretty good. 

What about more specific questions? 

Q. Are flexible or fixed route buses better?

A. The choice between flexible or fixed route buses depends on the specific transportation needs of an area. Both have their advantages and disadvantages.

Fixed route buses operate on a set schedule, following a predetermined route and stopping at designated bus stops. They are generally considered more cost-effective than flexible route buses because they can be scheduled to run at the most efficient times and follow the most direct routes. They are also easier for passengers to navigate, as they follow a predictable pattern.

Flexible route buses, also known as demand-responsive transport (DRT) or dial-a-ride, provide service to areas not served by fixed-route buses. They can be scheduled to pick up passengers at specific locations and drop them off at specific locations, rather than at designated bus stops. They are more suitable for less densely populated areas, where the demand for bus service is low and the routes would not be viable for fixed-route buses. They are also more expensive to operate than fixed-route buses, as they require more resources to coordinate rides and navigate the service area.

In conclusion, both flexible and fixed-route buses have their advantages and disadvantages. The choice between them depends on the specific transportation needs of an area, such as population density, demand for service and budget availability.

Not a bad answer. Points for highlighting their generally lower productivity. Another question:

Q. Are electric or hydrogen buses best?

A. Electric and hydrogen fuel cell buses are both considered zero-emission alternatives to diesel buses, but they have some different advantages and disadvantages.

Electric buses are powered by electricity stored in batteries, which can be recharged at charging stations. Electric buses are considered to be more mature technology, with a proven track record in many cities around the world. They are also cheaper to operate than hydrogen fuel cell buses, as electricity is cheaper than hydrogen fuel. They also have less maintenance costs.

Hydrogen fuel cell buses, on the other hand, use hydrogen fuel stored in tanks to generate electricity, which powers the vehicle. Hydrogen fuel cell buses are considered to be more energy efficient than electric buses, and they can also travel further on a single tank of hydrogen than electric buses can travel on a single charge. This can make them more suitable for long-distance routes or routes with low charging infrastructure.

Both types of zero-emission buses still have some challenges to overcome. The cost of purchasing electric buses is typically higher than that of diesel buses, and the cost of hydrogen fuel cell buses is even higher. Additionally, the infrastructure required to support electric buses, such as charging stations and hydrogen fueling stations, can also be expensive to install.

In conclusion, both electric and hydrogen fuel cell buses are considered to be zero-emission alternatives to diesel buses and have their own advantages and disadvantages. The decision on which technology to use will depend on the specific needs of a community, such as the availability of charging and fueling infrastructure, the cost of the technology and the distance of the routes.

 Let's ask it about network planning approaches. I'll use a question similar to that DoT asked the public to answer in consultation for newly launched bus network reviews in Melbourne's north and north-east. 

Q. Should bus networks have fewer routes that are more frequent or more routes that are closer to peoples homes but are less frequent?

A. The decision on whether to have fewer routes that are more frequent or more routes that are closer to people's homes but are less frequent depends on the specific transportation needs of a community.

Having fewer routes that are more frequent can provide a higher level of service for passengers and can make public transportation more convenient and accessible. This can also help reduce wait times and improve reliability. This approach can be beneficial for people who need to travel to major destinations such as city centers, universities, or commercial areas.

On the other hand, having more routes that are closer to people's homes but are less frequent can provide better access to public transportation for people who live in less densely populated areas or in areas not served by existing bus routes. This approach allows for more direct service to people's homes, reducing the need for transfers and making it easier for people to access public transportation.

A compromise solution would be to have a mix of both, where the bus network includes a core of high-frequency routes connecting major destinations, and additional, less frequent routes that provide service to more residential areas. This would provide a good level of service to both major destinations and residential areas.

Ultimately, the decision on the bus network design will depend on the specific transportation needs of a community, as well as the availability of resources and funding. It's important to consider factors such as population density, demand for service, traffic conditions, and budget availability when developing a bus service plan.

I'll try another question.

Q. What is a smartbus?

A. SmartBus is a bus service in Melbourne, Australia, that provides high-frequency and high-capacity services on major corridors. These buses are designed to improve the reliability and speed of bus services, and to provide better connectivity with other forms of public transport, such as trains and trams. SmartBus services typically run at least every 15 minutes, seven days a week, during the day and evening, and have fewer stops than regular bus services, which allows for faster and more efficient journeys.

Some of the features of SmartBus services include:

Real-time passenger information: This includes digital displays at bus stops and on board the buses, which provide passengers with up-to-date information on the location of buses and estimated arrival times.

Priority at traffic lights: SmartBuses have equipment on board that communicate with the traffic lights, allowing them to be given priority and spend less time at traffic lights.

Bus Rapid Transit (BRT) infrastructure : This includes dedicated bus lanes, which are separated from other traffic, allowing buses to bypass congestion and run more efficiently.

Enhanced stations: SmartBus stations are typically larger and more comfortable than regular bus stops, and often include amenities such as seating, shelters, and real-time information displays.

Connectivity with other forms of transport: SmartBuses are designed to provide better connections with other forms of public transport, such as trains and trams, to make it easier for commuters to transfer between different modes of transport.

Overall, SmartBus service is designed to provide a faster, more reliable and more convenient public transport option for commuters, and to improve the overall efficiency of the public transport system in Melbourne.

Above describes what we'd like SmartBus to be rather than what it is. Eg evening and weekend service often drops to every 30 min, and not the 15 min cited. And connections with trains are over-stated as frequencies are often unharmonised.

Here's a specific question. It's asking for an opinion which may be harder for AI. It also requires information unlikely to be obtained from published official sources. 

Q. What is the most complicated bus route in Melbourne?

A. It is difficult to say which bus route in Melbourne is the most complicated, as it can depend on various factors such as the number of stops, the number of turns and the level of traffic. However, some bus routes in Melbourne are known for their complexity, such as the route 250, which operates between the Melbourne CBD and Box Hill. This route covers a large area and has many stops, making it a long and winding route. It also covers a diverse range of neighborhoods and areas, from the busy city center to suburban residential areas.

Another route that could be considered complex is the route 546, which operates between the Melbourne CBD and Eltham. This route also covers a large area and has many stops, making it a long and winding route. It also covers a diverse range of neighborhoods and areas, from the busy city center to suburban residential areas.

It's important to note that the level of complexity of a bus route can change depending on the time of day and the traffic conditions. During rush hour or on weekends, some routes might be more complicated than usual due to the high volume of people and cars on the road.

It's also worth mentioning that many of the bus routes in Melbourne are interconnected and cross each other, making it more complex for the passenger to navigate through the system, especially for the first timers.

This one is full of errors, some factual. For instance the termini for both routes 250 and 546 are wrong. And they are not particularly complicated. The answer has a lot of general fluff which doesn't say very much. 

Just to prove the above wasn't a fluke here's another question. 

Q. What is the 788 bus?

A. The 788 bus is a bus route in Melbourne, Australia. It operates between the suburb of Mordialloc and the city of Frankston, via the suburb of Chelsea. It serves several key destinations along the route, such as Mordialloc railway station, Chelsea railway station, and Frankston railway station, connecting them with the surrounding areas.

The 788 bus is operated by the state-owned bus company, CDC Victoria, and it's part of the SmartBus network, which provides high-frequency and high-capacity bus services on major corridors. As such, it operates at least every 15 minutes, seven days a week, during the day and evening, and has fewer stops than regular bus services, which allows for faster and more efficient journeys.

SmartBus services typically include real-time passenger information, priority at traffic lights, Bus Rapid Transit (BRT) infrastructure, enhanced stations, and connectivity with other forms of transport, such as trains and trams, making it a convenient and efficient way to travel between Mordialloc, Chelsea and Frankston.

It's important to note that the route, schedule, and stops of the 788 bus are subject to change, and it's always best to check the Public Transport Victoria (PTV) website or app for the most up-to-date information.

Just about all the above is wrong. Route 788 does not serve Mordialloc and Chelsea, is not run by CDC (who are not state-owned) and is not a SmartBus with frequent service.  

The above shows the limitations of the ChatGPT 'artificial intelligence' when giving answers. They're a bit like a greenhorn senior executive recruited from a non-transport area on their first day. They might have a 'high level' general knowledge and be able to express themselves well. But knowledge of specific matters remains an undeveloped area. The same goes for recent events due to the 2021 cut-off. That's acknowledged here:

Q. How qualified are you to answer detailed questions about buses in Melbourne?

A. As a language model, I have been trained on a wide range of information, including public transport in Melbourne, including buses. However, my knowledge is based on the data that was available at my training cut off date, which is 2021 and my knowledge of the specific routes and schedules may be out of date.

In addition, my responses are generated based on patterns in the data used to train me and may not reflect the exact situation on the ground. In cases where specific, detailed and/or up-to-date information is needed, I would recommend consulting the official website of Public Transport Victoria (PTV) or speaking with a representative from PTV for the most accurate and current information.

So as much as I might sometimes appreciate a break, they won't be replacing me in writing items for this blog any time soon. But if you get ChatGPT to write a haiku, limerick, TV show idea, arts grants proposal or even a budget funding submission on a fairly general topic you might be surprised at the quality of some results. 

Try it today and let us know how you go in the comments. 

A role for developer-led bus services in new estates?

What happens when new estates develop and the government does not put on bus services when the people move in?

Sometimes the answer is nothing. Car ownership, cost of living and valuable real estate devoted to parking can be higher than they need to be in these areas. There remain neighbourhoods developed 30 or 40 years ago that still have no bus or a skeleton service of a few trips per day. 

Examples include areas around Ringwood East, Croydon South, Scoresby, Knoxfield and parts of the Mornington Peninsula. Hence the multi-decade longevity of one-trip routes like the 745 or confusing deviations like on Route 737. And the embarrassment of Epping North/Wollert which had been developed with the expectations of a train but was limping around for years on a single infrequent bus (Route 575) until extra routes got added in 2016. In our own time Route 511 in Beveridge has remained a token service with just a handful of peak period trips. 

Newer subdivisions, provided they are large scale and not middle of nowhere planning errors like Eynesbury, have generally got route buses a lot sooner than 30 years after development. Indeed that's pretty much been the main budgetary emphasis for buses in the last several years as established area network reform and service improvements took a back seat. Furthermore, the routes added typically run to the minimum 'safety net' standards established in 2006. This includes 7 day service until approximately 9pm. The minimum frequency is hourly but 40 minute frequencies are common. While harder to remember, buses every 40 minutes harmonise with every second train in most outer areas. Peak frequencies vary from the same as off-peak to every 20 - 30 minutes on better served routes. This is hardly the stuff of major modal shift but does constitute a basic service that gets surprisingly good usage in some areas. 

It can take too long for bus services to start after people move in. This is due to the absence of a reliable funding pipeline for growth area routes (GAIC has limitations) and sclerotic DoT internal processes for implementing them. Unlike industrial-scale level crossing removal and road construction projects, bus planning is more like a bespoke cottage industry in which it is not considered extraordinary that adding a bus route can take longer than large infrastructure builds. Other barriers include multiple land owners, the sequencing of estate development and the construction of houses before direct roads good for efficient bus routes are completed.

If there's no bus when people move in to an area, households will stretch their budgets to buy and run more cars than what could be the case if buses operated. Once this has been done it is harder to change habits and encourage bus use when service does get added. That risks harming productivity measures like passenger boardings per bus service hour and the farebox recovery ratio. In contrast a bus service that's there when people move in means people see it as part of the neighbourhood and use it from Day 1. 

Developers providing buses

Sometimes there's a local need for improved transport in an area where existing options are sparse or do not exist. Given funding and implementation difficulties on the government side, what are alternatives if you want public transport introduced early in an estate's development? Or even if the estate is old and the government never got around to introducing a service. 

Local organisations and other tiers of government can take matters into their own hands by establishing a bus service for a particular need in cases where the state government has historically been unresponsive. The Mornington Peninsula has been a prime example of this. That area has seen a confusing mix of bespoke bus services including weekly seniors shopper services (run by council), a student / jobseeker bus (initially set up with federal government grant money as PenBus) and some late night hotel services. More recent government investment has seen some (like the university bus) incorporated into an upgraded regular network (which however remains undeveloped eg 3 or 4pm finishes). This is possibly because Mornington Peninsula is neither a designated growth area nor treated as metropolitan for matters such as minimum coverage and service standards for metropolitan buses. 

New estates are very different. They're faster growing with a far younger and more diverse demographic. Parts are still developing. Yet they share some similar issues to the Mornington Peninsula regarding bus coverage and frequency. 

To partly mitigate this some developers have introduced their own bus, often running to the nearest station. This is either a voluntary measure to increase their estate's saleability relative to others or imposed as a condition of developing. Services are typically very limited (eg there may be no off-peak or weekend service) but may still be useful for some people. 

The role of developer-funded buses is discussed in an RMIT paper prepared for BusVic (the bus operators' industry body). Developer-led buses – a solution for growth area transport needs was released in December 2021 but has only just come to my attention. Authors are Annette Kroen, Steve Pemberton, Robin Goodman & Chris De Gruyter. You can read it here.

The paper has many facts and figures useful for anyone wanting to know about the state of growth area bus planning and provision. The research found that new estates typically got their first bus route four years after development though there was wide variation. Routes typically run every 40 - 60 minutes though sometimes have better peak services. Avenues for funding are explored with case studies given of developer funded buses in specific new estates in Melbourne.

I recommend you read the whole thing (it's about 30 pages). Points that jumped out for me include:  

* The difficult in finding recurrent funding such as is needed to introduce regular bus services to growth areas. Growth Area Infrastructure Contribution is another funding source but can only be used for five years. The service must either be funded from general revenue or it ceases. 

* A survey was done of 150 new estates and when they got buses. In some cases buses were already there when residents moved in. The extreme example was a 14 year lag. Average wait time for those without service is 4 years. 

* Ten developer-funded bus services in Australia were identified. Seven of those were in Victoria. They largely ran during peak periods only on fixed routes (which were seen to be more reliable than flexible routes especially given the limited fleet size). All were voluntary except Eynesbury's which was a requirement. Operation was done either by commissioning a bus company or through a non-profit body. The former has more experience while the latter may be cheaper. 

* Developers were more likely to commence an estate bus service if their development was large in scale and they expected that they would not need to run it for very long (eg the government would introduce regular routes). The latter was also a reason for developers not introducing a bus - they would not want to be in a position of cancelling the service if no route bus replacement had been introduced. 

* There is a possibility that early developer funded buses generate an appetite for bus usage that continues after a state funded bus route is introduced with these having higher usage than similar routes in areas whose estates did not have a developer funded bus. 

* Main limitations with developer funded buses include their limited service (eg no off-peaks or weekends) and the volatility of service (as developers can withdraw the service). However they can be introduced early (especially given the lead-times for a government-organised route). However the latter provide an overall better service due to better hours, frequencies and destinations (including not being artificially confined to particular estates). 

* Some operators had issues with booking systems and on-demand services due to unreliability. Hence services tended towards fixed route/fixed schedule services.

* There were two main areas of government that developers liaised with when considering whether to put on a bus. Experiences with DoT/PTV were negative - they were unable to get straight answers on when their area would get a regular route. However relationships with local MPs was critical and effective. Local governments were not seen as important as buses are a state responsibility. 

* Greater certainty on when an area would get a regular service would give developers an incentive to initiate a temporary service. This is because there is assurance a. that there will be a regular route service introduced and b. developers know that they will only need to fund the service for (say) 2 years rather than an indefinite period. 

Scope for improvement

The study concludes that developer funded buses are a good thing to provide a service early in an estate's development. They build habits of bus usage that can carry after a regular route is introduced. However to encourage them government needs to be more timely with regards to introducing services to new estates with definite and funded plans for same. This improves certainty and makes it more likely developers will fund temporary services. 

Priorities for government include having an overall transport plan with details and time-lines for new bus routes, better sequencing of development and improved communication between parts of government. Doing these would help ensure that people in growth areas get the better transport they need sooner.

RMIT and BusVic have produced a useful study which shines a light on the extent of growth area bus services and the role of government and developers. I recommend you read it!

Summer read: Make it free and make it worse?

If you make something free (or more formally remove user charges) does that mean it's considered so important that it should be a universal service? Or it's so marginal, unimportant, inconvenient or run-down that no one expects people to pay for it?

It's a question that's important for us to answer. This is because 'free' public transport has become a political cause in some cities, fashionable with some 'progressive' or 'radical' urbanists and environmentalists overseas. In Melbourne weaker strands of it are discernible when people seek to defend or extend the 'free' tram zone or within Infrastructure Victoria when arguing for lower bus fares. Such people, though well-meaning, are often ignorant of the true factors that drive major sustained patronage uplift and mode-shift, ie convenience and service quality, especially relative to car driving.  

Public telephones and automatic teller machines are two examples of services being made free as their number and coverage retreats. Are there parallels with public transport? Keep reading. 

ATMs

Let's look at automatic teller machines, which were rolled out from the '80s. If you saw something you wanted in the Trading Post or the newspaper classifieds and didn't have the cash laying around at home, you had to wait until a weekday to withdraw it from a bank branch. These had even more restricted opening hours than now so if you couldn't get there on a weekday between about 10 am and 4pm then you had no hope. Unless the seller accepted cheques, the concept which probably needs explaining to anyone much under 50. 

Initially the main benefit of ATMs was that they were operating when bank branches weren't. Then banks used the cover of the '90s recession to start closing branches. But they'd often leave an ATM. So every significant main street had one or more ATMs from each of the main banks.

Early on there were different systems, so that if you had a plastic card from one bank you couldn't use it in another bank's ATM if it was from a different system. That got resolved but you often had to pay a fee to use a different bank's ATM. So the thrifty had to hunt around for 'their' bank's ATM. 

Retailers got into electronic payment through EFTPOS and the larger ones (especially) offered cash out with purchases. So ATMs had competition for withdrawals. There were deals between financial institutions so you could withdraw cash from another institution's ATM for free. This was wonderful for user convenience as you could just use the nearest ATM. 

Bank branch closures seemed to abate for a while. Then they stepped up. This time though branches were removing all of their presence, including the ATM. But with 'free' ATMs and EFTPOS they could argue that opportunities remain for their customers to access physical cash banking facilities. 

Then banks started pulling out their ATMs. As long as there was an ATM from another bank they could claim that their customers still had access to cash withdrawal (and in some cases deposit) services. From the banks point of view fewer ATMs is more efficient. And it was cheaper for them to maintain less (and not charge) than to maintain more and charge people. So we ended up with no fees and less coverage. Having said that some still see opportunities in cash as some locations where there were once (free) bank ATMs now have non-bank ATMs which do charge for withdrawals. 

The banks are private profit-making businesses. They do not (and purists would argue should not) do anything not in their shareholder's profit interests. Making a service 'free' could just be an intermediate stop on the road to substantially downsizing or even ultimately removing a service. Though they'd argue that they're just responding to consumers' preferences with less use of cash.  

Public phones

Public phones are operated by Telstra. This area is more regulated than banking with Telstra subject to a Community Service Obligation in relation to the provision of services including public phones. They used to be coin only with call rates higher than from a home phone. Later they accepted various card systems for payment and even got the ability to send text messages. 

Notwithstanding CSOs there are way less public phones than say 30 or 40 years ago. While the CSOs may protect public phones in communities already with them they do not seem to be very effective in putting them into new estates. Hence walkability to a public phone along with that of other services like letter boxes tends to be much less in newer areas.  

The big change of course is mobile phone ownership and use. Up to the mid-late '90s they were expensive and rare. You did not use them for non-essential calls. But now almost everyone has one and calls are cheap. Not only that but they've evolved to be more like a personal electronic assistant with the ability to take photos, send data and post cat pictures on social media. 

Hence public phone usage has dropped. Telstra has still had to retain them under the CSO so it has looked at ways to reduce upkeep. Emptying cash containers was one of those costs. So it took the step of making calls from them free (within Australia). 

Free payphone calls was great for the unphoned and/or unhomed. But otherwise it didn't make a lot of difference given that most people had an affordable communications option right in their pocket. The lesson is that if there's a choice between (a) free and less accessible and (b) affordable at your fingertips then people with a choice will go with (b) almost every time.

With better (private) alternatives available to most people payphone usage has thus gone from a mainstream to a marginal residual experience since the '90s. Its survival depends on a political willingness to maintain a community service obligation rather than market demand. 

Parallels with public transport

Of the two above examples the phone analogy is the one nearest to public transport. This is because many people have access to fast convenient alternatives (eg driving) that are either at your door or in your pocket. Also, like telephone CSOs, our public transport has significant regulation in that fares apply across all modes and operators cannot unilaterally cut timetabled services without government approval. The history of public transport becoming a marginal transport mode in many cities (notably in the US) is not dissimilar to the move from public to private (mobile) phones. It's also notable that US transit fares often appear low by our standards. 

If you do make public transport free but don't improve the service then you are competing with walking and cycling more than you are with driving. Whereas if you use the money to make public transport better (faster on average door to door, achieved through a combination of higher frequency, better speed, priority over cars and reformed routes) then the relativities between mechanised modes change. Better service induces higher usage, at least some of which will have come from driving. Fare revenue and the cost-recovery ratio will likely also be higher, especially if improvements mostly boost off-peak services on routes with high patronage potential. That creates a virtuous cycle. 

People don't make phone calls as much as they did in the past. But when we do we think less about the charges. Whereas in the past we did for anything more than a local call because calls were timed. Value-conscious or less time-sensitive people took advantage of lower off-peak call rates. There are some parallels with transport where the system offers cheaper off-peak fares. We tend not to do that in Melbourne except for some weekend trips, cases where all your travel is after 6pm and before 7:15am Early Bird weekday train travel (which is free).

The case for cheaper off-peak travel is generally stronger than across the board fare cuts and much stronger than removing fares. This is for multiple reasons including higher travel/fare elasticity, the low and often almost zero marginal cost of carrying more passengers and non-transport social benefits. It's also politically attractive due to the low amount of revenue forgone. While we don't have the very high train fares that the UK has, you could argue that there are a few cases where our fares are uncompetitively high, eg some short trips, family V/Line travel and discontinuities in the fare scale such as between Lara and Geelong. Reform in cases like these have merit, though even here there are service-related obstacles like low weekend train frequencies. 

One could argue that any increased usage generated from a fare cut can build momentum for service increases which leads to another virtuous cycle. One could argue that the 2007 Zone 3 removal did just that for Melbourne train commuters, though a lot of people got a lot of bad service for a long time. And much of the growth came from other factors such as CBD employment and gentrifying inner areas (who didn't get a fare cut) so this point is debatable. It has been argued that Early Bird free fares helped redistribute train loadings away from the peak by providing an incentive for those who could start early to do so (at the cost of filling station parking earlier as buses were not included).  

Our tram network has a relatively flat timetable with off-peak frequencies not much lower than peak frequencies, at least during the day. But trains and buses are often different and taking a 'service first' approach to off-peak travel is preferable if I had to choose between one or the other.

Summary

Making public transport free risks 'residualise' a service and start a downward spiral since there are no (at least partly) offsetting financial gains from trying to make the service better and boosting usage. There may also be lower expectations from users that cause bad service to be more accepted ("Do you expect anything better - you didn't pay to use it?"). When service is so bad that usage drops then the constituency of users willing to advocate for improvements falls. And those users left (and left behind) have been of a demographic considered to be taken for granted voters in historically safe seats (though the 2022 federal and state elections resulted in some interesting shifts).  

It's also a bit of a one-trick pony. Making it 'free' is a one-off whereas boosting service offers a pathway to continuous improvement that make public transport a driving alternative for more trips for more people for more of the day. For example starting with off-peak frequency and bus network reform then graduating up to bus lanes, busways, light and heavy rail etc. 

In a world of multiple transport choices, public transport's role and contribution is most effective where it is good and affordable rather than bad and cheap (or free). You want it to rival driving rather than active transport. That means making it good before making it cheap and instead of making it 'free'.