Delta Meltdown 

While flying to Sydney, I got caught in the great Delta Meltdown of April 2017. While not as much of a kerfuffle as the United beating incident, it affected far more travellers. While my tale is seemingly a boring anecdote of “how was your trip?” that people ask as a way of small-talk, without really wanting to know the answer, it provides a window into the operation of a large transportation organization.

Jon Ostrower writes the official history of this breakdown of Atlanta-based Delta for Atlanta-based CNN.

After years of profitability and reliable service, Delta Air Lines struggled mightily last week with two basic functions of its business — flying airplanes and accommodating passengers.

Severe weather that pounded Atlanta in the middle of spring break caused a five-day meltdown across Delta’s flight network, leaving passengers fuming and its own crews waiting for instructions.

The weather is, of course, out of Delta’s control. But the chaos was amplified by the phenomenal complexity of running a modern-day mega-airline, according to interviews with Delta pilots, flight attendants and other staffers.

The episode was out of character for America’s second-biggest carrier. Its chief operating officer, Gil West, apologized and said the recovery from the storm had “not been ideal.”

Delta canceled more than 3,500 flights from Wednesday through Sunday. By Monday, the airline was getting back to normal, a spokesman said, but the effects of the storms still lingered.

The Atlanta Journal-Constitution also has a similar story.

I suspect the far-too sympathetic reporters underestimate the incompetence and overestimate the “bad luck” nature of this. Better contingency planning is an obvious need. In my mind, from watching the behaviour of actual crew members and ground staff at the airport, I imagine the crew-management computers were out (because of lack of backup systems, a lack of redundant control centers,  and a hard crash with a power outage), and schedulers were using note cards on a large board like in the old days. The pilot’s forum hints at this. However since the schedulers weren’t old-timers, they weren’t very good at it. I also suspect Delta has a very vulnerable network, and would be better off with fewer inter-dependencies between routes and crews. At the other extreme is a carrier that runs a lot of point-to-point flights, any of which can be disabled without taking out the whole network.


In my case, the flight from MSP to LAX (neither of which were affected by weather) was delayed 5 times (and the gate moved a couple of times) (from 5:43 in the evening to 9:40 pm) before being cancelled. If they had cancelled and rebooked immediately, there would be far fewer problems.  The gate agent in fact said at first that the delays were due to mechanical issues with a plane, rather than the “weather”, but then they later started blaming the “weather”, which I believe has financial consequences about compensating passengers. I believe they also deplaned a bunch of passengers from one flight hoping to reuse it for us. In any case, the problem was not the presence of airplanes, it was the presence of crews. There were an insufficient number of either flight attendants or pilots at the right gate simultaneously (take your pick), and then when they changed the size of the aircraft, they discovered they needed  another flight attendant (because ‘rules,’ even though the number of passengers is exactly the same).

To mollify annoyed if not irate customers, Delta gave out pizzas, which is great, unless you are dairy intolerant.

Fortunately I was put on the last flight out of MSP, but too late to make that night’s connection from LAX to SYD. Apparently they officially automatically rebooked me through Houston, but they never bothered to tell me that, so they were very confused when I called them from LA. Their computer databases don’t sync very well, since they did put me on the LA flight, and that should have been known to their system.

In any case, after arriving at LAX, I got a hotel near the airport. I waited in the taxi line, and when my turn came up, the taxi driver advised me to wait in a different line for the free hotel shuttle, which came right along. He didn’t want such a short distance fare.

I called Delta from the hotel to rebook. I was on hold for 3 hours. I went to sleep in this time and just put the phone on speaker, hoping that when a human came on line I would wake up. This actually worked. Even more amazingly, my cell phone carrier (T-Mobile) held a connection for 3 hours, which never happens. They put me on the flight the next night and all was well. I went to Manhattan Beach (next post) and had no subsequent issues getting to Australia.

The three hour wait is generally unacceptable. One wonders why there is not a 3rd party business providing surplus call center people trained in the ways of multiple organisations that can work as temporary staff for different companies. I understand JetBlue uses at-home workers for call-center staff.

Later, I complained to Delta and they said they would reimburse me for the hotel (a check is in the mail). I didn’t ask for the meals since I figured I had to eat anyway (one was built into the hotel, one was with friends). They also gave me, and I guess everyone, 20000 frequent flyer miles, which is not enough for a domestic ticket, but better than 0. I assume the cost of domestic trips in terms of miles will soon quietly get more expensive. It almost says “I’m sorry”.





Recruiting students

As a new academic staff* of the Faculty of Engineering and Information Technology, I am recruiting graduate students.

A current list of my PhD and research master’s project opportunities can be seen here:

PhD and master’s project opportunities


An overview of the application process is outlined at

When lodging an application for admission, you are expected to have already discussed your research proposal with a potential supervisor (me), and will need a support letter.

You will need to draft a statement of research interest, this typically includes a proposed topic area, a review of research already in that area, specific research questions, and a brief idea of how the research will be conducted. I will review that with you, and if I am interested, may support your application. Even with the support of a potential supervisor, you will still need to meet the University of Sydney’s rigorous graduate admission criteria, and opportunities with financial support available are even more competitive.





* I am “academic staff” not “faculty”, the “faculty” is the “college”, a college is basically some hybrid between a “house” (in the Harvard system, without quite the percentages) and a “fraternity/sorority” in the land-grant system. Also the “school” is the “department.”


Hyperloops and circular runways

I will appear on ABC (Australian Broadcasting Corporation) RN (Radio National) show Future Tense Sunday 7 May 2017 10:30AM. They also have a podcast.

In Japan they’re constructing a passenger train system that will travel at speeds in excess of 600 kilometres an hour. Now that’s fast, but it’s not fast enough for some. New Hyperloop technology promises the speed of sound. But can a train really go that fast? And why would it need to travel in a vacuum tube?

Also, we meet a man with a revolutionary new approach to runway design. He wants airports to look and function like velodromes, with planes landing and taking off on a donut-shaped runway.


Zachary McClelland – project leader, VicHyper, RIMT University

Edouard Schneiders – team leader, Delft Hyperloop, Delft University of Technology

Dirk Ahlborn – CEO and founder, Hyperloop Transportation Technologies

Steve Artis – Director, Ultraspeed Australia

Professor David Levinson – School of Civil Engineering, University of Sydney

Hesse Hesselink, Researcher, Netherlands Aerospace Centre

Update, with Transcript

Anthony Funnell wrote this up: Hyperloop or hyper-loopy? The race to make high-speed tube travel a reality. I abstract my bits below.

‘Human factor’ a complication

Both HTT and Hyperloop One are yet to do field tests using human passengers.

For this reason, the University of Sydney’s David Levinson argued it was premature to be building a business case.

“They’re going to put people into a sealed container and accelerate them at very high rates of acceleration, sometimes around curves,” he said.

“We don’t know how normal people will react to that because we haven’t done that before with normal people.

“This is sort of test pilot territory. People aren’t going to be really excited about being in a rollercoaster for a very long period of time.

“Yes, people have flown at 1,000kph in the Concorde — but there’s a different type of acceleration and deceleration profile associated with that.”

Dr Levinson, a professor of transport engineering, also has safety concerns.

“Track inside a tube is really pretty vulnerable to attack,” he said.

“How does such a thing respond if the tube gets punctured? What happens to the capsules that are inside of it? Does it gracefully decelerate or does some sort of implosion happen?

“People are going to expect something like this to be as safe or safer than trains or aeroplanes if they are going to be entrusting such a company with their lives.”

Using Temporal Detrending to Observe the Spatial Correlation of Traffic

Recently published: Figure2.jpg

This empirical study sheds light on the correlation of traffic links under different traffic regimes. We mimic the behavior of real traffic by pinpointing the correlation between 140 freeway traffic links in a sub-network of the Minneapolis – St. Paul highway system with a grid-like network topology. This topology enables us to juxtapose positive correlation with negative correlation, which has been overlooked in short-term traffic forecasting models. To accurately and reliably measure the correlation between traffic links, we develop an algorithm that eliminates temporal trends in three dimensions: (1) hourly dimension, (2) weekly dimension, and (3) system dimension for each link. The correlation of traffic links exhibits a stronger negative correlation in rush hours, when congestion affects route choice. Although this correlation occurs mostly in parallel links, it is also observed upstream, where travelers receive information and are able to switch to substitute paths. Irrespective to the time-of-day and day-of-week, a strong positive correlation is witnessed between upstream and downstream links. This correlation is stronger in uncongested regimes, as traffic flow passes through consecutive links more quickly and there is no congestion effect to shift or stall traffic. The extracted correlation structure can augment the accuracy of short-term traffic forecasting models.

Toll Roads – a view after 25 years

David Hensher at ITLS Sydney writes: Toll Roads – a view after 25 years. I excerpt an interesting bit below, emphasis added, but read the whole thing.


One of the great errors in the current tolling model has been the political decision to prescribe a unit toll rate which is indexed over time by the consumer price index. This has resulted in ring fencing on a crucial mechanism that is capable of recognising the need to adjust the toll to ensure that the travel time savings are delivered commensurate with the value (to the users) of those time savings relative to the non-tolled route(s), given travellers’ value of travel time savings. Consultants have struggled to establish the best outcome in relation to patronage forecasts because of this seriously problematic imposition. Added to the fact that consultants associated with the bidding consortia are often told to improve the patronage forecasts in ways that require what might be best described as imaginative (‘long tail’) futures, extending the range of time related benefits (such as the toll quality bonus) in the search for even higher patronage forecasts for a fixed toll regime. This point aligns with Bain’s 21 ways to inflate toll road traffic forecasts2. This becomes a commercial proposition in contrast to a network efficiency solution, resulting often in the loss of network welfare gains. Unfortunately, there is no incentive for the operator of a stand-alone asset to think ‘network’ There currently exists a complete failure across all tolled roads in Australia to optimise the level of toll and I believe this is generally opposed to by the operating companies of tolled roads on many grounds, but specifically their liking of the greater certainty of revenue flows even if these flows are a mismatch in delivering a better performing road network.  Only the state thinks ‘network’.  This is a key issue.  The state gives away pricing controls and then finds it difficult to optimise the network when it only has control over this important lever for part of the network.  This is of little concern when there are small, isolated sections of privately-operated toll roads.  It suddenly becomes a massive concern when these privately-operated toll roads ‘become’ the network!

Note: Bain, R. (2012) Twenty-One Limitations & Shortcomings with Traditional 4-Step Models, “In Australia, a number of the toll road concessions were awarded to the bidder offering the largest upfront payment to the state. That’s a recipe for disaster. Without checks and balances in place the bidding process simply turns into a competition on traffic numbers. Toll road traffic generates revenue, and the largest upfront payments can be justified by those with the highest traffic forecasts. The whole process becomes skewed and the numbers get bent out of shape in response. Considerable pressure is placed on traffic consultants to come up with the ‘right’ numbers; numbers that meet the requirements of the financial model.”


I made a similar point in my post “setting the right toll“, which was mostly theoretical in nature, but has clearly become problematic in nature here in Australia.

Review of ‘Forecasting Urban Travel: Past, Present and Future,’ by David Boyce and Huw Williams

Book Review:

BoyceWilliamsForecasting Urban Travel: Past, Present and Future, by David Boyce and Huw Williams. 2015. Cheltenham, U.K. and Northampton, Massachusetts: Edward Elgar. 650 + ix. ISBN: 9781848449602.

David Boyce and Huw Williams have written the definitive history of travel demand modeling to date. The book really begins (chapter 2) with Douglass Carroll, whose innovations in Detroit, and then in Chicago, developed the foundational four-step transportation planning forecasting tool that is widely used and abused in metropolitan areas across the globe. The aim at the time was to develop forecasts of future traffic—how many trips, where are they going, how many would drive, and which routes would they use—that could be used to locate and size freeways being deployed with the upcoming Interstate Highway System. In one sense, it was enormously successful, as the model spread from the Midwest of the United States across the globe, and has been used to conduct analyses, inform, and justify projects. It also spurred enormous methodological advances, one of which earned Daniel McFadden a Nobel Prize in economics for his work on developing random utility choice models.

After developing the framework, the book explores the formalization of spatial interaction models, especially the work of Alan Wilson (chapter 3), as well as the emergence of choice models in the 1960s, through development and generalization of the multinomial logit in its many, many forms (chapters 4 and 5), and the unification of the two. I have experience working as a travel demand modeler, so while it was familiar territory, it was also clarifying in many ways. Reading the book, I understood the linkages far more deeply than before. One of the merits of the book is the hiding of most of the mathematics in the end notes. (Note 1) The book also provides a very detailed history of developments in the United Kingdom and some discussion of developments in other European countries, in addition to the history from the United States’ perspective, providing a more global perspective on the issue. I was not aware the East Germans had developed their own modeling methods, and remain curious about practices elsewhere in the Eastern Bloc.

The book turns to the present trend in microscopic models, including activity-based models (chapter 6), which model tours of individuals rather than trips of aggregate flows, and the Transims project (chapter 8). From a methodological perspective, this is clearly an improvement, and the direction most applied models seem to be moving toward. By recognizing time-space constraints as formulated by Ha ̈gerstrand, the models become more realistic. If it were presented clearly to the public, it would probably be more intuitive and comprehensible than traditional aggregate models. This has not improved their predictive value, however, and Transims was, as the book describes, a diversion of resources with little to show for it.

After dealing with the first three steps of the travel demand model, the book then discusses network equilibrium and solution methods for route assignment or route choice (chapter 7). These evolved independently from the work of Beckmann, McGuire, and Winsten in the 1950s (coinci- dentally also in Chicago, though not interacting with Carroll’s group), were operationalized with steadily improving algorithms in the 1970s, and were mainstreamed into practice by the late 1980s. While equilibrium for route choice is now standard (whether or not traffic is in equilibrium, and whether or not route choice depends solely on travel cost), there is no standard “feedback” method to ensure that the travel time inputs to the travel demand process are consistent with the travel time outputs of route choice procedures, one of the many problems with actual transportation modeling practice, and a problem that as the authors note has received insufficient attention in the literature (the authors notably excepted).

Travel demand models have their uses. Long-term forecasting—their putative rationale—is probably the weakest. Yet, this entire enterprise was nominally driven by the desire to forecast travel demand, not simply understand or model it. It is indeed in the title of the book. On that score, the field has, in my opinion, failed miserably. Boyce and Williams note that accuracy is poor, and not improving over time. While one can understand the naivety of early modelers in the 1950s and 1960s (who undoubtedly well understood the limitations), by the 1970s (and certainly by the 2010s), the futility of accurate forecasting should have become apparent to those both within and outside the field. The forecasts are driven by the assumption that behavior in the future, given identical characteristics, will be the same as today. Culture is outside the scope of models, with good reason, but if culture matters, or anything else that is also outside the model’s data, there will be misses. Modelers may claim data issues, or poor inputs, and those certainly matter, yet as the authors note, estimation of models across time is never done in practice. There are always reasons— incompatibility of surveys, time, budget, and so on. The excuse for using cross-sectional analysis in the 1950s was that there was no time series; only one survey (at most) had ever been done in any metropolitan area. The excuse today is what?

In addition to behavior being static in these models, technology is as well. As the book notes, use of stated preference models to examine what would happen given a new technology attempts to push the boundaries of this, but it fails to say what technologies will actually be around, which will affect demand in ways we just have to admit that we cannot accurately foresee. This issue is increasingly important as new modes like shared autonomous vehicles are being considered, and autonomous vehicles (even if unshared) change the character of automobile travel. That a forecasting tool considering 30 years into the future cannot consider the possibility of such change in any reliable way suggests that it is probably not the right tool.

For this reason, these travel demand forecasts, at one time the most sophisticated analyses done by humans with their early use of mainframes (described in chapter 10), fall into the same trap as much simpler forecasts: underestimating growth in the early years of a technology’s life cycle and overestimating in the late stages. Boyce and Williams note the many problems with forecasts toward the end of the book (chapter 11), but remain more optimistic than I would. I believe the field should take its tools and apply them where they might be useful: short-term analyses of minor changes, scenario analyses of alternatives, but most definitely not forecasts. This requires changing evaluation procedures and government regulations. However, there are enough problems today that remain unsolved, so that looking for problems 20–30 years down the road seems futile.

In the end, this book serves as an excellent history of ideas about the research and its application in travel behavior and travel demand modeling forecasting, and should be widely read by researchers and practitioners in the field, and owned by their library. In my mind, however, it is a history of proceeding very deeply down the wrong rabbit hole, of systematic application of mathematical methods providing a veneer of science to cover, in the end, for political decisions.

— David Levinson

The Review appears in Journal of Regional Science June 2016 issue:


1 As Alfred Marshall wrote: “But I know I had a growing feeling in the later years of my work at the subject that a good mathematical theorem dealing with economic hypotheses was very unlikely to be good economics: and I went more and more on the rules—(1) Use mathematics as a short-hand language, rather than as an engine of inquiry. (2) Keep to them till you have done. (3) Translate into English. (4) Then illustrate by examples that are important in real life. (5) Burn the mathematics. (6) If you can’t succeed in 4, burn 3. This last I did often.” pp. 427–428 of Memorials of Alfred Marshall, edited by A. C. Pigou. One edition is: New York, A. M. Kelley, 1966.

Sydney Train Stations Need Two Exits

One of the things about the trains of Sydney is that, in contrast with the light rail and Metro stations I am used to, so many of the train stations have only one entrance or exit, and it is on one side of the train station (rather than the middle).

This is fine if that is the only direction you are going, but in many cases (maybe even half, maybe more in a poorly designed stations) someone is interested in going the other way. So for instance, let’s say you have an east-west line, and an entrance/exit at the east end of the station. Let’s say further you live or work west of the station, or you are on the west end of the train and need to transfer to another platform. You would wish there were another entrance and crossing.

Redfern-GoogleEarthShown in the figure is Redfern station, nearest to the University of Sydney. While there is an entrance to the southwest at Australia Technology Park, it takes you to Platform 9 and 3/4 10, which is seldom served, meaning you have to go all the way to the front of the station, (climb the stairs) find your platform (climb down), and position yourself in the desired car, which might be the other end of the platform depending on your destination. There is a lot of backtracking. While climbing the stairs up and down may not be avoidable, the backtracking is.

Backtracking is not mere inconvenience, it is a loss of time for want of staircase and overpass, and thus it is  a loss of ridership and a loss of fare revenue. It is also a loss of accessibility thus and a loss of land value.

Further, this is not a small amount of time. The length of a train set is about 163 m, the railway platform is about the same, maybe a bit longer (I estimated 180m before I looked up the length of a train).  It takes nearly two minutes walking at normal speed to go from one end of the platform to another, and a passenger may need to do that twice in the extreme case (though they may have to do that second one anyway making the long walk either before they board or after they exit, because of the location of exits at the arrival train station, it could very well be that the problem there is the mirror, with the exits at say the west side when the traveler’s final destination was to the east). (And likely the process occurs twice a day.) Four minutes walking at 1.5 m/s is 360m (nearly a quarter mile). Eight minutes walking is 720m.* And this is inside the station area, leaving aside access to the station.

Now, I recognize that platform configurations may not always permit insertion of staircases or lifts near the unserved end of the station platform due to space reasons, especially when the track curve as shown at Redfern. The platforms often narrow at the end, and may themselves be awkward retrofits. But that does not mean they never do. And some expense is justifiable to the train operator if the benefits in ridership or property uplift are sufficient and can be captured, and to society if they cannot.

I am not going to make a list of the Sydney train stations where this issue should be investigated, but somebody should, and some assessment of their feasibility and cost would be warranted.

At a minimum, a few stations can be retrofitted as a pilot project and the effects on ridership, land value, and investment monitored.


Update: A reader writes:

Yes interesting comment about Redfern.

Now in fact when I went to Sydney Uni there was a footbridge exactly where you suggest – I recall that it was a somewhat ricketty timber structure but very handy none the less

This reference says:

“In c1994 the southern footbridge was removed as the Eveleigh railway workshops were gradually closed down and the footbridge was no longer required” which seems to neglect the traffic that used to access the university.

I suspect that additionally there were costs associated with fare collection management etc – which today could be covered by Opal Card readers – which the Govt of the day wanted to avoid.

Wikipedia also sort of supports this saying:

“Until 1994, Redfern had an overhead footbridge at the Eveleigh end of the platforms, connecting platforms 1-10 by stairs. This was demolished because the funds for its maintenance were not available.[1]”

Interestingly your comment certainly could have also applied to one of the busiest stations on the network – Central Electric i.e. the suburban platformswhich lacked a southern concourse until about 15 -20 years ago. That concourse is linked to the Devonshire street tunnel and provided access for thousands of -……you guessed it ……………uni students going to UTS. And workers going to the Railway square precinct which also happens to house the offices of TfNSW!



Note: There is also a potential control issue, as this would lead to more than one entrance to the station, which might have security issues. A real criminal would be willing to leave the platform and traverse rail tracks if necessary, so I think this is largely spurious. It might however warrant extra staffing if people (police) are needed to monitor entrances and exits, as occurs at some stations. I suspect most of this can be dealt with using cameras.

* I know, walking is supposed to be healthy for you, but this isn’t about health, it is about convenience and time.

On Benefit / Cost Analysis and Project Selection in Transport

Before I left Minnesota I was asked by a Representative in the Legislature how to improve Department of Transportation project selection (following up on this presentation in February). I wrote back this (I revised and extended my remarks for the blog):

A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King
A Political Economy of Access: Infrastructure, Networks, Cities, and Institutions by David M. Levinson and David A. King

What are transport “needs”? It’s simple, a project is a “need” when the full benefits exceed the full costs. [Clearly very few projects are a “need” in an existential sense, but what we are talking about are more than “wants” in that they are net benefits for society, by definition.]

Measuring benefits and costs can be tricky, but it is not impossible to get a first order estimate, and the general principle is straight-forward. Sadly almost no agency requires actual benefit/cost analysis.

So I would suggest rules something like:

  1. All highway, transit, airport, and port projects that are considered in project-selection processes involving expenditure of state or federal funds above $5 million shall undergo a consistent, peer-reviewed, monetized benefit/cost analysis that would
    • Consider the full benefits and full costs of the project (in comparison with a no-build alternative) incorporating changes in: number of passengers and freight, travel time and travel time reliability, accessibility to employment and workforce, land value, wider economic benefits, crashes and crash severity, air, water, land, noise, pollution costs, and carbon emissions, public health (including both physical activity and pollution levels), vehicle operating costs, as well as the costs of building, maintaining, and operating the project over time.
    • Consider these costs and benefits distinctly for the population as a whole as well as any relevant transportation disadvantaged groups
    • Consider these costs and benefits not only for the project, but for the relevant portion of the transportation network, including related transportation sections both upstream and downstream of the project and competing with the project.
    • Consider uncertainty bounds in the estimation
  2. These analyses must be performed according to a standard methodology published by the Department of Transportation (DOT).
  3. The methodology and analyses shall be reviewed every two years by a national panel of transport and economics experts convened by DOT.
  4. The results of these analyses, including both the final results as well as the component estimations, shall be made public and posted on the DOT website in a readily accessible manner. An Annual Report of considered and selected projects shall be provided with the full benefits and full costs reported, and justification provided for any projects that were selected over other projects with higher expected benefit/cost ratios.
  5. In order to improve travel and cost forecasting, and provide an understanding of the accuracy of such forecasts:
    • The project-delivering agency shall review project cost estimates made at the time the project was approved for construction upon completion of the project, and report to the Legislature a table of expected and actual cost expenditures for all projects.
    • The agency shall review travel demand estimates made at the time the project was approved for construction 5, 10, and 20 years after completion of the project and report to the Legislature a table of expected travel and actual travel for all projects.

Sydney’s Ferries

Sydney has most of the usual modes of transport. It also has ferries.

I'm on a boat
I’m on a boat

Ferries of course are not unique to Sydney, but they seem to be more significant here than in any US or western city I have seen (Sydney ferries served about 15 million passenger journeys per year, fewer than the State of Washington as a whole, but more than Seattle, the leader in the US, compared with some 100 million passengers in the entire US). Ferries have steadily declined in importance in the US, where they would be replaced by bridges wherever feasible to ensure continuous (or nearly continuous in the case of draw bridges) rather than scheduled service. In western civilization, ferries harken back to the ancient Greeks, who preferred then much faster water transport over land transport (also Greece has lots of islands). Charon even charged the dead for transport over the River Styx, so this was a precursor to toll bridges and toll roads.

Around Sydney Harbour with Public Transport
Around Sydney Harbour with Public Transport

I had a good lunch recently with Robin Sandell, who blogs about Sydney’s Ferries and has a twitter feed devoted to the subject. His idea of running the service as a hub and spoke system with timed transfers, like the Swiss Railway Taktfahrplanto encourage use is really interesting. How much additional demand would such a system induce? [Hint: Research Project]. He also notes that in Sydney Ferry subsidies are less than rail subsidies, and comparable with bus. The fares are not inexpensive, except on Sunday, when they are. While like most of Sydney public transport, my experience is that it seems to adhere well to schedule,  I have heard complaints about random scheduling and schedule adherence issues on ferries

The Ferry Service has at least one private competitor, the Manly Fast Ferry, which you guessed it, is a faster ferry to Manly, northeast of the CBD (for more money). This seems a perfect case to do a Value of Time revealed preference study.  [Hint: Research Project]

From the tourist, or new arrival’s perspective, they are fantastic, a good excuse to enjoy transport. Even from the regular traveler’s perspective, while waiting for a ferry at the wharf might not be your favorite thing to do, remember, you are at a wharf, overlooking the water, waiting for a ferry. This beats a bus stop or a train station or sitting in traffic.  I would bet customer happiness on ferries is on average significantly higher than other modes. [Hint: Research Project]