It's no more than 4 trains an hour now at Pickering, and GO still can't integrate even the regular speed VIA trains without delays.
Besides, as day-1 Stouffville service is supposed to be every a MINIMIUM of 15-minutes in each direction off-peak. The idea is to increase service as the year progress. Every 15 minutes isn't the ultimate build-out. It's the minimum. You'll need extra tracks for HSR. Which is presumably why they'd gone back to the Don Valley (and building a new tunnel from Lucien L'allier to Laval).
The capacity of a double-track railway is upwards of 12 trains per hour with ETCS. As I illustrated above, the average speed of GO trains and Alto trains would be about the same between Kennedy and Union, so that's an achievable number in this case.
Even with their old-school signalling system on the Lakeshore East line Metrolinx
currently operates 9 trains per hour westbound on a single track between Danforth and Union:
December 2025 AM Peak timetable, Lakeshore East, Stouffville & Via Rail
They also currently operate 8 trains per hour eastbound on a single track between Mount Dennis and Union including a mix of express and local services (the 3rd track is currently out of service there too)
The reason there are delays at Durham Junction is that
1) It's an at-grade junction, not a flyover as I'm suggesting for Alto, and
2) Metrolinx dispatching is actively hostile to Via trains because they are only evaluated on the on-time performance of Metrolinx-operated trains. They make no effort to minimize delay for Via. There's a similar issue from the scheduling side, as you can see in the Lakeshore East timetable above where Via 43 is scheduled to arrive at a time that is physically impossible.
They aren't spending tens of billions of $ to avoid the freight railways, and then have Metrolinx ***** them up the ***.
The difference of course being that Metrolinx is a public agency, so the public and the feds can force them to stop screwing over other public services. It seems irresponsible to spend billions of tax dollars just because Provincial bureaucrats and Federal bureaucrats can't get along. These are issues that can be resolved with inter-agency agreements that cost basically nothing.
The GO vision for the Stouffville line changes regularly, and I'm not sure that we should bake in a 15 minute headway as the proven upper limit. It seems likely that ML will eventually need and attempt 10 minute or better headways and that will constrain what Alto experiences. Moreover, mixing GO and Alto ties Alto to whatever disruptions GO experiences..... do we want Alto held every tine GO has an ill passenger, stuck door, or whatever ? This is why I do not trust the math around an interleaved LSE line....an express/local division with flyover is a better model than two routes with two tracks each.
I can't see Alto going much beyond hourly headway with half hourly at peak....but I do foresee instances where there might be two Alto's fleeted with say 10 minutes between them. And what if a Peterborough regional or GO service does become a need ? To my mind, keeping Alto away from GO altogether is a prudent futureproofing step even at added cost, as having the two together constrains the upper end of both services.
Sharing tracks through Kennedy wouldn't be baking in 4 tph as the limit for Stouffville, it would be baking in 12 trains per hour as the limit for the sum of Stouffville and Alto.
Like you said, Alto is unlikely to run more than 2 trains per hour, and maybe there would be a GO service from Peterborough too. So that's 4 trains per hour coming in on the high speed line, which leaves 8 trains per hour for the Stouffville line. Are we really expecting more than 8 trains per hour on Stouffville? Seems unlikely. Realistically peak period service would be more like 6 trains per hour for Stouffville and 2 trains per hour for Alto:
Reliability is indeed an important consideration, so if we're considering more expensive options such as the Leaside alignment, we should also consider more expensive long-term solutions to Scarborough Junction, such as one that connects the Stouffville Line between the express and local tracks of the Lakeshore line, allowing Alto trains to use the local track if the express track is obstructed:
This could potentially be built as a 2 km tunnel directly from Kennedy, eliminating the sharp curves through that segment. But this would would be a very long-term item. In the meantime Alto and the Stouffville line can share the existing line with plenty of capacity to spare once ETCS is installed.
I would expect a Leaside route to have some double track added down the Don - if ML could design a whole yard up there, some room must exist. Single track over the viaduct only would not unduly limit operability - certainly much less than the unpredictability of inbound GO and Alto trains vying for track from an Agincourt junction, or arriving off their slot and ending up in the wrong order all the way from Kennedy to Union..
I would not expect speed higher than say 95mph west of Agincourt, but it would be possible to match the Guildwood-Cherry quality track speed from there to Leaside, with a 70 mph flyover/under and 60-70 Mph down the Don Branch. True there is a tight curve at the bottom, but again with many trains running to East Harbour I would expect speed on the LSE/Stouffville route to be comparable in that stretch.
The speed on LSE/Stouffville wouldn't be as slow as 25 mph through East Harbour, it would be 45 or 50 mph. The Kennedy alignment allows pretty much continuous 95 mph running from Kennedy to Union, and it wouldn't take much to upgrade the line to 80 mph from Kennedy to the 401.
While one might be able to design a clever interleaved service and implement a better signalling system to enable that, in the interest in saving some dollars....why would we do that when the result constrains both GO and Alto ? Whatever the added cost, it's the wrong place to value engineer or pinch pennies. It's not a prudent place for a "good is good enough" approach. The less Alto requires from ML, and vice versa....the better inho.
We would do that because it produces a better service in the end. Being able to board a high speed train to Ottawa at Kennedy station would be a massive ridership benefit by opening up the eastern GTA much more than any of the suburban station options on the CPKC line. Furthermore, the actual amount of constraint created by sharing a quad-track corridor between GO and Alto is quite minimal, and I'm not convinced it's much more than a single track bridge on the Leaside alignment.