News   GLOBAL  |  Apr 02, 2020
 11K     0 
News   GLOBAL  |  Apr 01, 2020
 43K     0 
News   GLOBAL  |  Apr 01, 2020
 6.6K     0 
Is there a reason why all intersections chosen are ones where there is competition with other streetcars?

I can't speak for staff on this, except to say that the potential delay reductions from TSP are greater at major intersections than minor ones. The three intersections chosen all represent the maximum possible savings, with only one omitted, and that's Queen.

The non-streetcar crossings don't produce as great a time savings. (that is not to suggest they shouldn't be done, merely that there is a logic to placing them lower in priority.)
 
Presentations from yesterday's board meeting. Linked the pdfs below but they really don't do the presentations justice which aren't long to watch. Some interesting Q&A after each

Opportunities to improve Transit Signal Priority in Toronto from Narayan Donaldson, Integrated Mobility Consultant:
slide deck: https://www.toronto.ca/legdocs/mmis/2026/ttc/bgrd/backgroundfile-264729.pdf

Developing a Surface Transit Revitalization Plan from Jonathan English, Founding Principal, Infrastory Insights
slide deck: https://www.toronto.ca/legdocs/mmis/2026/ttc/bgrd/backgroundfile-284274.pdf

I really don't agree with a lot of what he is saying here, because the assumptions and basis is upon what he is thinking is a fair playing field. Its not.

The problem with any conditional TSP is that the assumption that the TTC schedules for streetcars being on time is fair. They are not. The schedules are insanely slow and need to be sped up. This is not about streetcars being "late" this is about streetcars being too slow EVEN WHEN THEY ARE TECHNICALLY ON TIME.

Furthermore, all this BS about maximum allowable green times is secretly about cars. Fuck cars. If a streetcar is coming to an intersection, the light needs to be held green, period. If it takes 4 minutes, fine. That light should be green for 4 minutes until the streetcar clears. Or 1000 minutes. Whatever it takes. If thats technically not TSP but signal pre-emption, great. Implement signal pre-emption. I don't care what you call it.

This is the problem with bureaucracy, they smell their own farts. They love to talk about things like "oh see, we can't do this and that" because of the own rules that they have imposed themselves within.
 
I really don't agree with a lot of what he is saying here, because the assumptions and basis is upon what he is thinking is a fair playing field. Its not.

The problem with any conditional TSP is that the assumption that the TTC schedules for streetcars being on time is fair. They are not. The schedules are insanely slow and need to be sped up. This is not about streetcars being "late" this is about streetcars being too slow EVEN WHEN THEY ARE TECHNICALLY ON TIME.
The schedules are created by observing the travel times and setting the schedule to an above-average time. If you set the schedule to the average time, half the streetcars will be late. The reason the schedules are so long currently is that the travel times are very inconsistent. Improving TSP therefore allows the schedule to be dramatically shortened.
Furthermore, all this BS about maximum allowable green times is secretly about cars.
Maximum allowable green times exist to avoid creating unlimited delay for other traffic, such as streetcars/buses on the cross street, and pedestrians trying to get to/from streetcar stops.

Sure it might sound nice in theory to inflict unlimited delay on pedestrians and perpendicular bus routes to get rid of the edge cases where a streetcar might get stopped, but in reality the imperceptible difference in travel time for the streetcar simply cannot justify the impact. In the real world we need to consider costs and benefits, and the benefit of saving less than a minute on a streetcar route simply isn't worth the cost of adding several minutes of delay to pedestrians.
Fuck cars. If a streetcar is coming to an intersection, the light needs to be held green, period. If it takes 4 minutes, fine. That light should be green for 4 minutes until the streetcar clears. Or 1000 minutes. Whatever it takes. If thats technically not TSP but signal pre-emption, great. Implement signal pre-emption. I don't care what you call it.
Again, it's really not about cars. The biggest constraint on TSP in Toronto is pedestrian signal timing (and the desire to avoid excessive pedestrian delay). See for example how TSP is allowed to extend lights by 30 seconds, but there's a limitation that only 16 seconds can be during Solid Don't Walk. If Transportation Services wanted to prioritize cars, they would just use Solid Don't Walk extensions everywhere, since they create much less impact on car traffic than Walk extensions. That is because Walk extensions need to be implemented well before the end of the green light so end up being longer than necessary to provide a margin of error to make sure the streetcar doesn't miss the green light. The longer green extensions associated with a Walk extension disrupt signal coordination (for cars) much more than a Solid Don't Walk extension would.

How about you try advocating for pre-emption and I try advocating for making the TSP work as well as possible, and let's see who achieves more benefit for streetcars.
This is the problem with bureaucracy, they smell their own farts. They love to talk about things like "oh see, we can't do this and that" because of the own rules that they have imposed themselves within.
And yet TSP systems like those in Rotterdam and Waterloo almost always line up a green light for the tram, despite having constraints that prevent the system from inflicting unlimited delay on others...
 
Last edited:
I really don't agree with a lot of what he is saying here, because the assumptions and basis is upon what he is thinking is a fair playing field. Its not.

The problem with any conditional TSP is that the assumption that the TTC schedules for streetcars being on time is fair. They are not. The schedules are insanely slow and need to be sped up. This is not about streetcars being "late" this is about streetcars being too slow EVEN WHEN THEY ARE TECHNICALLY ON TIME.

Furthermore, all this BS about maximum allowable green times is secretly about cars. Fuck cars. If a streetcar is coming to an intersection, the light needs to be held green, period. If it takes 4 minutes, fine. That light should be green for 4 minutes until the streetcar clears. Or 1000 minutes. Whatever it takes. If thats technically not TSP but signal pre-emption, great. Implement signal pre-emption. I don't care what you call it.

This is the problem with bureaucracy, they smell their own farts. They love to talk about things like "oh see, we can't do this and that" because of the own rules that they have imposed themselves within.

While I appreciate your passion, the three intersections being piloted have streetcars approaching from all directions, (4). You literally cannot prioritize them all at once.

There are real world limits.

A further note, when you back up car traffic on a road where cars are permitted on the streetcar tracks, such as College, Dundas or Queen, you also backup the streetcars.

****

Be assured those presenting to the TTC are true believers in transit, they're passionate, they're riders, they know what they're talking about.

****

As to the need to go further than what's proposed to date............. absolutely.

There are some discussions ongoing behind the scenes.

That said, I will be posting in the days ahead on things people should be asking for here, so that pilot work is as successful as possible.

The three intersections on their own, absent switch improvements will only reduce travel time modestly. Its a start, not the end product, and we do have to watch that any benefit isn't dismissed for being small when the scope effort isn't fulsome.
 
I can't speak for staff on this, except to say that the potential delay reductions from TSP are greater at major intersections than minor ones. The three intersections chosen all represent the maximum possible savings, with only one omitted, and that's Queen.

The non-streetcar crossings don't produce as great a time savings. (that is not to suggest they shouldn't be done, merely that there is a logic to placing them lower in priority.)
I wonder whether Queen-Spadina is excluded because of the Ontario Line construction at that intersection. There are frequent lane closures/reconfiguration and paid-duty traffic control while equipment is being delivered etc. It would complicate evaluation of the TSP changes.
 
While I appreciate your passion, the three intersections being piloted have streetcars approaching from all directions, (4). You literally cannot prioritize them all at once.

There are real world limits.

A further note, when you back up car traffic on a road where cars are permitted on the streetcar tracks, such as College, Dundas or Queen, you also backup the streetcars.

****

Be assured those presenting to the TTC are true believers in transit, they're passionate, they're riders, they know what they're talking about.

****

As to the need to go further than what's proposed to date............. absolutely.

There are some discussions ongoing behind the scenes.

That said, I will be posting in the days ahead on things people should be asking for here, so that pilot work is as successful as possible.

The three intersections on their own, absent switch improvements will only reduce travel time modestly. Its a start, not the end product, and we do have to watch that any benefit isn't dismissed for being small when the scope effort isn't fulsome.
All the City has publicly stated so far is that they are imminently implementing phase rotation on Spadina at Dundas, College and King, followed by rollouts across Spadina, Finch W and Eglinton E. Which is obviously good news, but it's not enough on its own to produce the level of priority that we really ought to have on Streetcar/LRT lines.

Here are the current signal timings at Spadina & College, which I obtained through an information request in December 2025. You'll note that there is no TSP whatsoever, so if the only change they make is to add phase rotation, there will still be long delays for streetcars. For example, a streetcar arriving shortly after the light changes to yellow will still need to sit through the minimum phase duration for College (which is very long due to the pedestrian crossing distance across Spadina). If they allowed streetcars to extend the green by up to 30 seconds like they do in the current signal timings at Spadina & Nassau, streetcars could mostly avoid that situation.
 
Last edited:
Of course the remedy for mixed-traffic streetcar is to make either no more on-street parking along streetcar (and bus) routes in the downtown, and/or ban non-residential automobile traffic in the downtown AND allow deliveries before 10:00 in the morning only.
 
While I appreciate your passion, the three intersections being piloted have streetcars approaching from all directions, (4). You literally cannot prioritize them all at once.

If two streetcars come to one intersection the lights should both go green and then the drivers play a game of chicken
 
The schedules are created by observing the travel times and setting the schedule to an above-average time. If you set the schedule to the average time, half the streetcars will be late. The reason the schedules are so long currently is that the travel times are very inconsistent. Improving TSP therefore allows the schedule to be dramatically shortened.
Dont you see the circular logic in this based upon what the guy was saying in the video? Its a feedback loop if its conditional TSP.

TSP sets the schedules, but then the schedules set what kind of TSP is used. He is saying "we dont need uncontidional TSP because if the trains are on time to the schedules, which are influenced by TSP, then we only need conditional TSP.

You can't have the influencer of a metric to be set by that metric. You have to independently set what you think a reasonable travel time is and then modify the device to meet that target independently.
 
While I appreciate your passion, the three intersections being piloted have streetcars approaching from all directions, (4). You literally cannot prioritize them all at once.

There are real world limits.

A further note, when you back up car traffic on a road where cars are permitted on the streetcar tracks, such as College, Dundas or Queen, you also backup the streetcars.

****

Be assured those presenting to the TTC are true believers in transit, they're passionate, they're riders, they know what they're talking about.

****

As to the need to go further than what's proposed to date............. absolutely.

There are some discussions ongoing behind the scenes.

That said, I will be posting in the days ahead on things people should be asking for here, so that pilot work is as successful as possible.

The three intersections on their own, absent switch improvements will only reduce travel time modestly. Its a start, not the end product, and we do have to watch that any benefit isn't dismissed for being small when the scope effort isn't fulsome.
Personally I just hope that they won't conclude that TSP delays the other streetcars and use that as an excuse to shelve the entire thing :/
 
Dont you see the circular logic in this based upon what the guy was saying in the video? Its a feedback loop if its conditional TSP.

Reaper has a very solid understanding of the issues.

TSP sets the schedules,

No, TSP does not set the schedules, it can create conditions for better schedules (or worse) depending on how its implemented.

Yes, these things are inter-related, but TSP is not setting any schedule.

That's Laurence Lui's job (head of TTC Service Planning).

Yes, the schedule will have to be changed or little to no benefit would be achieved.

The changes here are not about one thing, but about multiple things done together, thoughtfully.
 
Personally I just hope that they won't conclude that TSP delays the other streetcars and use that as an excuse to shelve the entire thing :/
Which only impacts a handful of intersections on Spadina, and none on St. Clair, as they'd need not activate northbound TSP for the out-of-service runs.
 
Personally I just hope that they won't conclude that TSP delays the other streetcars and use that as an excuse to shelve the entire thing :/

The impacts on cross-lines can be partially or fully offset by using 'compensation time' for the cross street within the signal's cycle; as well as by implementing some priority measures on the cross street itself.

While this will ultimately be important.....any delays fro crossing streetcars (affected by just one intersection with Spadina) will be less than 1 minute and that's a worst case scenario. Not a real concern at this point.
 
Dont you see the circular logic in this based upon what the guy was saying in the video? Its a feedback loop if its conditional TSP.

TSP sets the schedules, but then the schedules set what kind of TSP is used. He is saying "we dont need uncontidional TSP because if the trains are on time to the schedules, which are influenced by TSP, then we only need conditional TSP.

You can't have the influencer of a metric to be set by that metric. You have to independently set what you think a reasonable travel time is and then modify the device to meet that target independently.
The schedule is set based on the slowest streetcars (i.e. the late ones). If you speed up the slowest streetcars, your schedule automatically gets shortened as part of the schedule review process.
 

Back
Top