Let me try..(PS..I have been under nuclear attack from all sides so show some patience).
The time to event works like this...As enrollees enter the study at different times (the case in R-I) the enrollee's date of entry is recorded and the clock is started...So the time in trial is not the same thing as the time the trial has run...During the course of the trial all patients are evened up by looking at their time in the trial..So you can look at all enrollees that have been in the trial for one year (or other intervals) and compare the number of events in each of the two arms (placebo vs active)...in a similar fashion you could compare all enrollees at two years and increasing amounts of time...This allows you to make what are called Kaplan-Meyer (KM) graphs and you can determine the HR for a specific period of treatment.
The beauty here is that you can compare enrollees that entered the trial at different times and get an accurate assessment of the treatment over a period of time. And determine such factors as how long does it take to produce therapeutic effects, whether the effects continue over a longer period of time, or even the reverse...
All this with having enrollees in the trial for differing periods of time..
You are correct - It is theoretically possible that the total events between groups could be similar, but the V group could take longer to event, which could still be a statistically significant beneficial effect.