Although the TTC Board takes a long siesta through the summer, the CEO produces a monthly report even in months when the Board does not meet. The August 2019 edition was recently posted on the TTC’s site.

This report continues a format established some time ago by CEO Rick Leary in which the focus is on measures of system performance. There is no financial information here, and only summary ridership numbers with no sense of the associated revenue. All of the financial reporting was hived off of the CEO’s report into a quarterly CFO’s report, but we have not seen one of those since April 2019, and that covered the year ending December 2018. One might have expected an update at the July Board meeting, but the abrupt and unexpected departure of CFO Dan Wright in early June appears to have iced the short-lived report.

I spoke with Wright at the Audit & Risk Management Committee meeting the day before he left the TTC, and he gave no indication of his impending departure. We spoke about the problems of counting “rides” in an environment where there is only a tenuous link between fare payment and the actual number of trips (or trip segments) taken.

The TTC has been wrestling with the possibility that ridership has been over-reported for some time, and the situation is further complicated by Metropasses, the move to Presto and the two-hour fare. Just what is a “ride” for statistical purposes? I had planned to follow this up with Wright for an article here, but alas he vanished like so many other senior TTC staff in the past year. The problem is summarized in the CEO’s Report:

Higher PRESTO adoption appears to have affected measured ridership in two ways. First, we now have more precise ridership data compared to counting tokens and weighing paper tickets. Second, more than 25% of our former monthly pass customers have converted to PRESTO pay-as-you-go e-purse each month in 2019, likely to take advantage of the two hour transfer and for some, the TTC/GO discounted co-fare. This would affect measured ridership to the extent that these customers may ride less often than the monthly average of 71 rides per adult monthly pass. [p 25]

Josie La Vita, the Executive Director of Financial Planning for the City of Toronto, replaced Wright on an acting basis pending recruitment of a new CFO, a process expected to take 12-24 months.

The CEO no longer reports financial data on operations or capital projects, and the absence of a CFO’s report leaves a major gap in information available to the TTC Board and Council on the system’s actual performance.

Moreover, riding counts (i.e. vehicle occupancy) are only rarely reported at a route level, and a “crowding report” has not recently seen the light of day. When crowding data are released, they are averaged and this gives no indication of the effects of bunching and gapping on individual vehicle loads. Performance metrics that do appear in the CEO’s Report do not fully describe the service quality actually experienced by riders particularly on surface routes.

Ridership for 2019 to the end of June is reported as 267.8 million as against a target of 271.9m and a 2018 figure of 270.3m (down 1.5% and 0.9% respectively). For the month of June itself, ridership is on target. As the chart below shows, the shortfall in 2019 came primarily in the winter months which were unusually cold. June’s ridership was boosted by the Toronto Raptors Championship Parade without which the number would have been down 0.6% compared to 2018.

Weekend ridership is down, and this is thought to be due to various factors including the number of subway shutdowns. These are not going to end any time soon with the ongoing signalling projects and other infrastructure upgrades, and at some point the TTC cannot treat their effects as unexpected. A follow-on problem for the TTC is the perception by riders that “the subway is always closed” and this can affect ridership as much as the actual shutdowns.

Presto fare card usage continues to increase and by June 2019 the adoption rate reached about 80%. Presto ridership for the first half of 2019 was 214.2 million out of the total reported ridership of 267.8m.

Vehicle Reliability

The fleet benefits from improved preventative maintenance and from the retirement of older vehicles. Before the onset of hot weather, there was a concerted effort to get air conditioning systems in good order, and AC failures were rare even with the particularly hot weather in 2019. (Personally, I never encountered a “hot car” on the subway, a first for several years running.)

On the streetcar fleet, the decline in the proportion of service provided by the nearly 40-year old CLRVs and the disappearance of their ALRV cousins (of which a few are still officially active but never seen in revenue service) contributes to a reduction of in service failures. The Flexity fleet has its ups and downs for reliability, but even running below its target, these cars are much more reliable than those they replace.

On the bus fleet, the retirement of old buses and the recent purchases of hundreds of new vehicles has substantially lowered the average age of a bus and increased the proportion of buses that are “spare” relative to service needs. The fleet is over 2,000 vehicles, but the peak requirement in June was 1,641 including buses used on streetcar lines. This generous spare ratio has benefits in service reliability but also a cost in both capital and in garaging.

Capacity and Reliability of Lines 1 YUS and 2 BD

The CEO’s report sets out statistics from the subway lines using two separate metrics:

• “On Time Performance” is measured at terminals and counted on an all-day basis Monday-Friday from 6 am to 2 am. A train is “on time” if it departs within 1.5 times its scheduled headway.
• “Capacity” is measured at several key points on the lines Monday-Friday from 6 to 9 am and 3 to 7 pm. This is based on the number of trains passing through the measurement points compared to the scheduled value.

Note that there is a substantial difference in the two charts and this can mislead readers.

For the On Time Performance chart, the scale is 0-100% and the variation appears small on either side of the 90% target line. For the capacity chart, the scale is 80-100 and the variation appears to be greater. OTP is measured over the entire day and only at terminals. This includes many off-peak trips that are outside of the sample captured by the capacity chart.

Subway Vehicle Reliability

For the month of June, the reliability of the TR fleet used on Line 1 YUS and Line 4 Sheppard fell well below the target level. However, this must be seen in the context that the number of failures in service resulting in a delay greater than five minutes is usually fairly small. This causes the MDBF (mean distance between failure) to swing up and down a lot around the target. June had an unusually large number of TR failures, 14, of which half were due to door problems.

The T1 fleet serves Line 2 BD and as these are older trains, their reliability target is lower. Even so, these cars have achieved much better reliability than their target and sometime rival the newer TRs. There were only six delays in June of five minutes or more, and four of these were due to door problems.

Status of the Bombardier Flexity Streetcar Order

The mean distance between failure for the Flexity fleet fell in June 2019 to 19,405 km, well below the contracted target of 35,000 km.

There are actually four separate reliability targets depending on the severity of the failure as shown below. The 35k target applies to the second most severe type of failure where a service delay results.

As of the date the CEO’s Report was written, 165 new cars had been commissioned, although nine of these are back at Bombardier as part of the major repair program for correction of faulty welds and flood damage. The table below shows plans for ramping up the number of cars scheduled for service to a total of 162 (out of a fleet of 196 available) by January 2020.

The current peak scheduled streetcars sit at about 160 including the remaining CLRVs, and with the 505 Dundas and 502/503 Kingston Road routes operating using buses. This shows quite clearly that the TTC will not be able to field a full streetcar service on its network even when all of the new cars are here and after reducing the operating spare ratio to 15%.

An important reason for this, which I will explore in a future article, is that new streetcars have been consumed at a rate greater than planned because of service improvements and because scheduled trip times have been extended both in response to traffic conditions and to give operators more terminal recovery time. (There is a similar problem with the bus fleet where some of the apparent increase in fleet size has been consumed not for more frequent service but for extended scheduled travel and recovery times, and to increase the spare ratio.)

Streetcar Service Quality

A major problem with all charts showing service quality for surface routes is that they summarize data for at least Monday-Friday if not for the full seven day week and all hours of service. This masks variations over the day – peak vs off peak, weekday vs weekend.

For surface routes, “On Time” is defined as departing from terminals no more than 1 minute early or 5 minutes late. On routes with frequent service, this means that buses or streetcars can begin their trips running as pairs or even triplets, but still be considered to be on time. A further problem is than any unevenness in headways from terminals is quickly magnified as vehicles travel along their route with bunching getting worse and gaps between groups of vehicles getting wider. There is no service metric for this condition even though most riders do not experience the comparatively reliable service at terminals.

A further problem is that “on time” is meaningless on most routes where it is a regular spacing of moderately loaded vehicles riders want to see. They do not care about the official schedule because, after all, “frequent service” should provide just what it says.

The OTP metric for streetcars has not changed much in the past few years and, if anything, is getting worse despite repeated schedule changes to pad running and recovery times.

In the chart below, the legend is incomplete. The orange line is for 2015, and the black one is for 2016.

Finally, because the performance charts summarize data from at least Monday-Friday, if not all days of the week, there is no breakdown of service

For several years, the TTC has attempted to wrestle the number of short turns down to a minimal level. There is no question that in years past, short turns were used excessively as a line management tool, and the amount of service delivered beyond common short-turn locations could be much worse than the advertised level. That said, there are cases where short turns are inevitable and a target of zero is impractical.

On many routes, the TTC has extended running and recovery times so that in theory short turns should almost never be required. However, this has a cost both in extra vehicles and in wider headways for all riders, not to mention bunching of vehicles at terminals during periods when the extra time is not actually needed.

As the number of older streetcars declines, the frequency of calls to repair or replace a vehicle in service has also dropped. The target for both streetcars and buses is 1.5% of peak scheduled service which, for streetcars, would be about 2.5 events per day (based on peak service of 160 cars). The actual numbers are improving, but they are still above the target level.

Bus Reliability and Service Quality

The TTC’s move to lower the average age of its bus fleet and to retire old vehicles after 12 years of service rather than 18 has translated into a much improved reliability value for the fleet overall. Another important part of this change was a move to better preventative maintenance so that, for example, the inspection cycle for vehicle systems was shorter than the typical failure interval so that problems could be caught in advance rather than causing an in-service breakdown.

The bus fleet has a large spare ratio, well over 20%, and this gives the TTC the luxury of keeping older problem vehicles off of the road. Whether that spare factor is excessive for a younger fleet leading to artificial constraints on peak service is a question the TTC should address in its fleet planning. Now that the oldest of its fleet is retired, a higher proportion of vehicles should be available for service. Of course that would have a budget impact because a bus sitting in the garage does not incur the cost of an operators, fuel and running maintenance.

The “on time” values for the bus network are generally better than for the streetcars because, considered on average, there are a lot of bus routes and periods of service where congestion and other random service disruptions are rare. Even so, the TTC only manages around 70% “on time” at terminals and this number has not moved in several years.

As with the streetcar network, there has been a push to reduce short turns. The numbers stabilized through 2017-18, but there has been a further decline with scheduling changes in recent months.

Thanks to the reduced age and improved reliability of the bus fleet, not to mention the improved preventative maintenance, the number of in-service failures requiring a road call or a replacement (change off) vehicle has fallen over the years and the rate is now sitting at the 1.5% target.

Presto Equipment Reliability

Although the Presto fare system equipment reliability has slowly improved, the CEO’s report raises many questions about the quality of equipment provided by Metrolinx/Presto. These issues did not arise overnight, but the political need to praise Presto at all costs appears to be much reduced now that the TTC revenue stream can be affected by fare collection system failures.

From the description of problems with fare gates, it is clear that the original design and materials used by the vendor Scheidt & Bachmann were not up to the task. This begs a question of both the original spec and of the acceptance of faulty gear.

We continue to work with S&B to address ongoing hardware and software issues. A number of programs have been developed and are currently being implemented. These include:

• The program to replace the industrial computers in the fare gates. S&B has a second generation industrial computer with a new Solid State Drive (SSD). This new computer with this drive will provide a number of improvements including: extending the hard drive capacity, improve and protect the hard drive sectors, increase the hard drive speed (faster read / write – start-up time will be improved), extending the data logging, and help address USB disconnect issue we are currently having with the fare gates. This program is ongoing and will require both hardware and software testing to be implemented;
• New software deployments. The next software update will: improve passage detection leading to a more reliable interface for the customers, provide an upgrade to the motor control interface improving reliability of the motors, and resolve one of the major issues we experience with the card reader. This upgrade will be available for deployment in Q3;
• With the ongoing issues with the current fare gate motors, S&B development teams completed an in depth field review. The team is currently reviewing the information obtained and developing recommendations for next steps. The report is expected to be completed shortly. Once their recommendations are reviewed, an action plan will be developed.

These plans will help to address the following issues: screen freezing, tap/no entry, card reader failures, motor and heater failures. We have additional software updates scheduled, which will add functionality and provide further fixes to know problems, improving the gate availability to the customers. [pp 65-66]

This is not a trivial list of problems, and one cannot help wondering how S&B gets customers anywhere.

As for Presto card readers, the CEO’s report notes:

Availability data from Metrolinx may be subject to inaccuracies …

TTC staff continue to conduct in-field observations and validation of card reader availability. Our initial assessment indicates there is some variance in the availability data provided by Metrolinx, attributed to card readers going in and out of service or being non-responsive while appearing to be operational. [p 67]

Regular riders are quite familiar with the problem that “the lights are on but nobody’s home” on Presto readers.

With respect to fare vending machines (FVMs), self-serve reload machines (SSRMs) and fare and transfer machines (FTMs), there is a common refrain:

TTC staff continue to conduct in-field observations and validation of [machine] availability. Our initial assessment indicates there is some variance in the availability data provided by Metrolinx, attributed to instances where some equipment repairs are not being reported and captured consistently. [pp 68-70]

Whether the TTC will ever get Metrolinx to accept full responsibility for revenue losses thanks to the shortcomings of Presto remains to be seen.