Based on my travel around Union during the day, a 5 car train would be leaving people behind.
Not necessarily during offpeak, if the trains are frequent (15-minute-and-better). Frequency can compensate. Consider the 6-car single-level TTC subway moves far more people in one hour at peak than Lakeshore West does at peak (8 trains 4:45-5:45). Some European cities have run commuter trains at subway-like headways (3-minutes!) at peak.
Again, how do you figure? For one, there is absolutely, positively no information whatsoever on what form these supposed EMUs will take. Hell, we don't even know if Metrolinx will choose EMUs at all. And beyond all that, as I wrote above, EMUs aren't going to change how the railroad operates.
This is very true, but look at the large number of planned frequent services planned for the Georgetown Corridor. If Metrolinx intends to run express trains AND allstop trains (UPX, diesel Kitchener trains, SmartTrack GO RER or just GO RER) in both directions, then it starts warranting EMU performance to reliably slot trains between trains (e.g. between UPX trains). Otherwise, Metrolinx would have to scale back the Georgetown corridor plans. UPX is already a Multiple Unit, and it dramatically outperforms the existing GOtrains, so the performance profiles make it difficult to slot UPX and pre-existing GOtrains together on the existing tracks. I've made a long post in regards to this respect (two pages long). Simply put, the railroad mathematics of the Georgetown Corridor points to at least one more EMU-route, if UPX continues to stay a MU (EMU) and they intend 15-min bidirectional service with multiple trains. At peak there would be 15-min-or-better service in five directions contending a 4-track corridor; involving SmarTrack/GO RER 2-way, UPX 2-way, and Mt Pleasant peak-direction diesels, all 15-min-or-better). We haven't even talked about corridor capacity used up by Barrie RER or Milton peak trains, which presumably will use separate tracks (cramming 6 tracks through Bloor will be needed). Only EMUs allow you to tightly slot train services between trains with short headways. There's a big reason why nearly all subways use EMUs
And we haven't even merely *thought* about potential future HSR service increasing pressure towards EMUs (There is an feasibility/EA under way for this corridor now and Libs want this to happen in ten years, it's more realistically 2030s+ (delays+density+viability) to the point where HSR might replace UPX during its end-of-life as a London-Kitchener-Pearson-Kitchener route (HSR to Pearson is presumably accomplished in theory via a Woodbine Racetrack stop to a discontinued UPX spur that is converted to run a LINK extension to all Pearson terminals). This requires two of the tracks to be dedicated to HSR, bumping or mergeing services to the remaining track per direction, and EMU performance required to slot services with sufficient performance for frequent service on all.
Even without HSR -- the Georgetown corridor mathematics is going to either cause EMUs
or broken promises (e.g. greatly scaled-back RER as an example). Very predictable two possible outcomes.