How Toronto's traffic signals work
A surprising chunk of your trip time evaporates at red lights, and Toronto won't tell you exactly why. The city keeps signal phase timing to itself, so we reverse-engineer it from each signal's published traits. Here's how the city's signals actually run, and how we turn that into a delay estimate.
The network today
- Active signals
- 2,549
- Transit-priority signals
- 438
- With pedestrian head-start (LPI)
- 1,724
- Control-mode types
- 9
Control modes: how each signal decides
A signal's control mode determines how much delay it imposes. Coordinated corridors are engineered for through-flow (a “green wave”) and hold traffic least; fixed-time signals can't respond to demand and hold it most.
| Mode | What it means | Active signals | Model base delay |
|---|---|---|---|
| SA2 | Semi-actuated (2-detector) | 1,300 | 5s |
| SA1 | Semi-actuated (1-detector) | 471 | 5s |
| SAP | Semi-actuated + pedestrian | 362 | 9s |
| FT | Fixed-time | 251 | 10s |
| PA - MPS | Plan-actuated, master-coordinated (green-wave network) | 101 | 3s |
| SAV | Semi-actuated, volume-density | 38 | 7s |
| Unknown | Unclassified / not published | 13 | 8s |
| PA - IPS | Plan-actuated, isolated | 12 | 4s |
| FA | Fixed-actuated | 1 | 10s |
Base delays are per-signal estimates in seconds; unclassified signals use a 8s default.
Intersection size
More approaches mean more signal phases and more waiting. Most Toronto signals are four-leg intersections; the rare five-plus-leg crossings need an extra phase, so the model adds 15s there.
Time-of-day plans
Signals run different cycle lengths and phase splits by period. Peak plans use longer cycles coordinated for the dominant flow, so the opposing direction waits longer. Overnight, many signals drop to actuated-only (sensor-triggered) or flash. Our model scales each signal's base delay by a time-of-day multiplier:
- AM peak (06:30–09:30)
- 1.5×
- midday (10:00–15:00)
- 1×
- PM peak (15:30–18:30)
- 1.6×
- evening (18:30–22:00)
- 0.8×
- overnight (22:00–06:00)
- 0.4×
Transit signal priority (TSP)
438 signals can hold or extend a green when a streetcar or bus approaches, with priority on corridors like King, Queen, and College. TSP speeds transit but adds variable delay to cross-traffic, so the model adds 10s to a transit-priority signal.
Leading pedestrian intervals (LPI)
1,724signals give pedestrians a head-start walk signal before parallel traffic gets green. It's a safety win that costs vehicles a few seconds, so the model adds 3s where an LPI is present. Unusual 5+ leg intersections need an extra phase and add 15s.
Why we estimate, not measure
Signal phase timing is operational city data and isn't published in open data. So we can't read the exact seconds a given signal will cost on a given trip. Instead we estimate it from each signal's published control mode, transit-priority and LPI flags, and leg count, scaled by time of day. It's an approximation: good for comparing routes, not a substitute for a live signal feed.