Is there any reason to believe or proof that any of these "nuisance" stops have actually prevented anything? Human error will always be a possibility until there is automated control of the system. And from what I've seen, the places where human error resulted in a crash were not locations of particularly high risk.
There is no easy answer to that. The system isn't foolproof. A human can elude any of the safety warnings. A human can elude any of the safety warnings on any non-ATC system. That's the difference between having ATC and not having it. And they need to address that before another crash ties their hands and forces them to quickly implement an ATC solution without the lead time to figure out the headway management that maintains capacity. That is the true doomsday scenario: get the ATC off-shelf, lose 20% of the service levels because there's not enough time on-deadline to right-size it.
The resignal isn't going to make the Central Subway go faster. Best case for ATC is fighting to par. No "pointless" reds, but your trip thru-and-thru goes not gain one second on the stopwatch vs. what it does today in a non-FUBAR condition. The closest-packed Jetsons-shit ATC system is more inflexible than humans are at exercising discretion in very close spacing situations like when packing back-to-back trains into Park St. So the self-perception of "movement > non-movement" with elimination of those enforced reds gets counterbalanced--and then some--by throttling back to low speed or outright elimination of those on-platform back-to-backs. It's movement, but lower-speed movement. Wow. What...a...difference...that...isn't.
The target they have to aim for with the new signal system design is that any attrition from present conditions in the closest technologically allowable train spacing can immediately be offset by running 3-car trains everywhere all the time. And run 4-car trains to bank for future growth. This is deadly tricky because when Red was re-signaled downtown in 1988 from more permissive Blue Line-style mechanical trips to today's magnetic pulse ATO they gambled on overly cautious signal spacing being offset by lengthening of trains from 4 to 6 cars. And whiffed on their projections, which is why you get to pause and enjoy the view on the middle of the Longfellow in the morning at a pointless 'virtual' red unable to approach Park any closer. Stakes are way, way higher on a Central Subway than they were on Red. If they misfire the slightest on design they lose a crippling amount of capacity that extra cars won't solve. It has to hit nail square on the head: if headways are X seconds wider under computer-regulated ATC, it must be within margins that train capacity can offset at zero aggregate loss of service.
Godspeed with that engineering job. It's not an exaggeration to say that getting it right may take a cumulative 15 years from next survey to prelim engineering to full design to build. And a build that starts endpoints-in before it ever touches Kenmore-GC. That's how delicate and fraught with peril the brain surgery is.
And ATC still doesn't address the Green's main problem: branch bunching KO'ing the schedules before the trains ever descend on Kenmore or Copley and see their first subway signal. That's the real cause of the GL's precipitous decay the last 20 years. The Central Subway has always been a "garbage in, garbage out" operation. It gives you what you feed it. Always has. This current signal system, "pointless" reds and all, managed crush-load traffic just fine for decades upon decades until the branches stopped being functional. When the B chokes on its own bunching and takes the C's and D's down with it, there is nothing the Central Subway can do to restore order. Slotting's all fucked up before trains ever merge onto the same track, and when bunching's dragging an outbound branch schedule down you have crapshoots like "Three D's and two E's before the first B sighting" syndrome @ Park outbound. String up Jetsons-shit signaling all the way from Kenmore to North Station but leave the branches thrown to the wolves unchanged from current conditions and it's exactly the same: garbage in, garbage out.
The only way this gets meaningfully better is taking out the trash at its source: on the fucked-up branches. Like a broken record, all of the obvious B/C/D/E fixes go on the table:
-- Improvements to station dwell times: all-door boarding and PoP as the rule, security cams for random-sample fare enforcement, standardized platform dimensions so entering and exiting passengers don't have to squeeze by each other.
-- Transit signal priority on B/C/E. Now, now, now.
-- GPS tracking everywhere to aid dispatch with a live look at train position instead of intermittent snapshot through radio squawking. (They're almost there with this capability.)
-- Stop recalibration. Not just outright consolidation, but also things like re-spacing for more stops with offset platforms on either side of a traffic light when that makes the starting spots vs. a traffic light much easier to time.
-- Short-turns and alt-routing. Supplemental B turnbacks at a Harvard Ave. mini-yard for surging and throttling around the student-heaviest hours. C's extended to BC to infill the B slots that are turning. D-to-E on-street connection to Brookline Village. Low-cost infrastructure enabling greater elasticity in branch service patterns.
-- Electrical system and platform length upgrades so 3-car trains run as a default condition on all B/C/D and Medford trains all the time except for the far off-peak. Heath-Union Sq. the only deuces left since Heath Loop isn't big enough for triplets.
And the first and arguably most important deployment of Jetsons-shit tech. . .
-- CBTC first on the D and GLX. There is still a human inefficiency with the present-day dispatching on those grade-separated lines (and lines-to-be), with lower posted speed limit out in the open than the vehicles are rated for due to concerns about line-of-sight signal visibility. Get those D's all the way in from Route 128 to Kenmore within a minute of schedule target every time and let the computers auto-adjust the timings through Longwood and Fenway to time the GPS position of the next B or C heading for the portal that it has to slot between. That's what takes the trash out of the Central Subway. Computer brain smooths out still more bunching by leveraging precision dispatching of the full-signaled branch against the variable schedules of the mixed-traffic branches. Kenmore...and especially Kenmore inbound...is where the new tech improves capacity and throughput. Central Subway still takes what you feed it. You're just no longer feeding it a shit sandwich of schedule conflicts at the mouth.
The ATC business will work itself out. For proactive action and advocacy purposes, it's still the ^^fat targets^^ we've known all along. Get the branches and the roads they share debugged, get better operating/boarding practices on the branches, increase their flexibility and establish better elasticity with short-turn or alt route throttling options. Do everything that improves their OTP from endpoint to portal so the dispatching at the portal is more predictable. And repair those rolling ruins.
That's all. Take the garbage out and the Central Subway will do its thing. Keep stuffing garbage down those portals and the Central Subway vomits it back up. It's not hard to put 2 and 2 together on primary cause-effect.