I think the problem with that is it wouldn't have nearly enough capacity for needed headways of rapid transit (10-15 minutes). Also, I would imagine the ideal configuration for 4 tracks would be:
[local platform]
<----- Local Track
[express/local platform]
<----- Express Track
------> Express Track
[express/local platform]
------> Local Track
[local platform]
so taking one of the locals for bi-directional running would hurt the local track setup and force locals onto one of the express train.
The NEC Infrastructure Improvements Master Plan (IIMP) sticks to pretty consistent platform layouts for MA/RI intermediate stops south of the Hyde Park constriction, with only moderate toughies and shared Amtrak stops deviating from the cookie cutter. Stations are due to have their track layouts uniformly expanded to full future capacity
first before infill mainline tri- and quad-track is completely constructed to join them all together, because passing fluidity at station stops gives much more immediate boost to service density for the $$$ than constructing linear miles of new mainline track. So if they mod the stations first they get big service increases...and then the mainline track expansion can be paid on the installment plan by highest-priority segments to eventually buff it all out to max ceiling.
The one commonality you see everywhere south of the SW Corridor constrictions is that crossing over is sharply curtailed by always having the locals 'bulb out' at a stop, never the expresses. That's important...and that's the one fatal malady that handicaps the Forest Hills-Hyde Park pair + any infills in-between.
Note that some of these designs will inevitably change as the NEC FUTURE commission's traffic modeling for 2040 Superduper HSR came a couple years
after the 2030-focused and Amtrak + NEC member states' DOT co-signed IIMP was released. The next revision of the IIMP is due for 2020 and probably will reflect the updated traffic modeling with some additional capacity expansion touches.
Here's what it's going to be for all non- terminal district (i.e. excluding Back Bay & Providence) stops. . .
--------------------------------------------------
Ruggles (final config under construction)
====< local track
[<- Island platform ->]
====< local & express track >
====> local track
[-> Northbound platform]
(Permanently constrained by the 3-track SW Corridor cut, but the new NB platform ends for-good the crossover games allowing for service increases here without fouling expresses. At Green St. the quad track will start, with southbound local track splitting in two and a crossover from the express track keeping the expresses centered.)
Forest Hills (4-track layout), Hyde Park (4-track layout)
====< local track
[<- Island platform ->]
====< express track / > local track
====> express track
====> local track
(FH is structurally unexpandable, and HP has to trade 3-track w/side platforms for 4-track w/squished island because of lack of space. With additional platforms physically impossible, it requires crossing over the express tracks to reach. Creates lots of conflicts that sharply limit which schedules can use these platforms. Any intermediate stops between FH and HP have to retain the same layout--even if there's room to do better--because more switches means an even bigger mess with crossing over. This right here is the Achilles heel of the whole works.)
--------------------------------------------------
Readville-NEC, Attleboro, South Attleboro, Pawtucket*, Cranston, T. F. Green (candidate final config)*, East Greenwich, West Davisville, Wickford Jct. (final config)
[<- Southbound platform]
====< local track
====< express track
====> express track
====> local track
[-> Northbound platform]
(* - Pawtucket track layout not 100% design-final so *limited* possibility of add'l changes. T. F. Green layout subject to change if Amtrak service added; see alternate build further below.)
(Note how if you DON'T have any squishing at FH & HP, you can simply fan out from Ruggles when the tracks go from 3 to 4 and go off to the races. This is where reanimated Readville is feasible for Regional Rail schedules while HP is not. Any other north-of-Wickford infills in RI outside of current official plans--such as the Olneyville stop in Providence that was dropped because of uncertainty in the schedule for 6-10 Connector interchange rebuild that's a prerequisite for serving up the station land--would follow suit with same quad layout.)
--------------------------------------------------
Sharon (IIMP plan)*, Mansfield (IIMP plan)*, Westerly
[<- Southbound platform]
====< local track
====< express track >
====> local track
[-> Northbound platform]
Sharon (site design tolerances)*, Mansfield (site design tolerances)*
[<- Southbound platform]
====< local track
====< express track
====> express track
====> local track
[-> Northbound platform]
(* - NEC IIMP pre-dates NEC FUTURE traffic modeling, so next revision of IIMP needs to bump Sharon+Mansfield to match 4-track layout of Attleboro, etc. Both station sites have silly-spacious room to do it by expanding on the southbound sides of the station properties, so baffling choice by MassDOT to paper-short the capacity at
lower than the stops further outbound in its IIMP submissions to Amtrak. Treat the current plans as non-final because they make so little sense and the quadding tweaks are so easy to do.)
(Fewer tracks, but same deal...locals bulb-out, expresses stay centered. The only places where you'd have to cross over is an Amtrak NB vs. Amtrak SB meet...and those meets are inherently rarer than Purple vs. Amtrak or Purple vs. Purple meets so it's not all that problematic. Especially once you get down to into the wide stop spacing of the South County, RI hinterlands. Any other un-proposed infills between Kingston and Westerly such as a reinstated Shannock station would follow suit with same tri layout. As above, Sharon/Mansfield plans a cinch to get revised to quad-track instead.)
--------------------------------------------------
Kingston (new Amtrak config)
====< local track
[<- Island platform ->]
====< Amtrak track >
====> local track
[-> Northbound platform]
(Only difference here vs. the tri- layout of Sharon/Mansfield/Westerly is that Kingston will always be a NE Regional stop so platform has to have access to the center track while commuter locals always hug the outer tracks. Uncongested segment so no pressing need for quad out here, though a 4th track does already exist in the form of a lumber yard freight siding adjacent to the NB local track making future quad-modding possible if necessary.)
T. F. Green (proposed Amtrak config)*
[<- Southbound platform]
====< local track
====< express & Amtrak track
[<- Amtrak island platform ->]
====> express & Amtrak track
====> local track
[-> Northbound platform]
(* - There's room to do this, and that's likely what RIDOT is going to shoot for in the end. Amtrak is warm to the idea of an NE Regional stop here, and the schedule tradeoff would probably only entail them permanently dropping Westerly + Mystic from the Regionals schedule once expanded RI + CT commuter rail take up residence there. More congested spot with more overlapping schedules and only the Acela skipping, so needs platform touches of all 4 tracks.)
--------------------------------------------------
Route 128 (IIMP plan)*
====< local track
[<- Island platform ->]
====< express track >
====> local track
[-> Northbound platform]
Route 128 (official facility future-proofing)*
====< local track
[<- Southbound island platform]
====< Amtrak track
====> Amtrak track
[-> Northbound island platform]
====> local track
(* - Totally baffling that the IIMP shorted the 2004 station's actual prebuilt expansion tolerances for 4 tracks/2 islands by only going for 3 tracks, because that's not nearly enough for NEC FUTURE modeling with all Regionals + Acelas stopping here. South Coast Rail already has to perma-skip this stop at loss of job access to the 128 belt for Stoughton commuters, which is an unacceptable demerit. There's not a chance that the shorted layout will hold up to scrutiny, so virtual guarantee that the 2020 IIMP revision will add Track 4.)
Route 128 (outer design tolerances)**
====< local track
[<- Southbound island platform]
====< local track
====< Amtrak track
[<- Amtrak island platform ->]
====> Amtrak track
====> local track
[-> Northbound island platform]
====> local track
(** - For the NSRL era there remains extra slack space on the east side + room under the 128 overpass to add still one more island for 6 total station tracks if track switches are reconfigured so the middle 4 tracks are the mainline and the outer 2 tracks are bulb-outs. This is how you'd terminate all manner of northside tunnel slots at 128.
Nobody is proposing anything this grandiose yet, nor does NEC FUTURE thru service require >4 tracks. It's solely a future-proofing consideration for NSRL. Note also that Tracks 5 & 6 would
not enable a Fairmount Line extension to Westwood; the need to re-engage Amtrak dispatch at Readville Jct. would hiccup a few too many "Indigo" headways when they conflict with priority Amtrak slots. A Fairmount-to-128 extension will always go via Dedham Corporate instead of Westwood because staying uniformly under one T dispatcher is crucial for maintaining rote consistent all-day headways without interruption.)
--------------------------------------------------
Canton Jct. (stet per IIMP plan)*
[<-Stoughton northbound platform]
====< Stoughton local track
====> Stoughton local track
[<- Island platform ->]
====< NEC local track
====> NEC local track
[-> NEC southbound platform]
Canton Jct. (site design tolerances)*
[<-Stoughton northbound platform]
====< Stoughton local track
====> Stoughton local track
[<- Island platform ->]
====< NEC local track
====< NEC express track >
====< NEC local track
[-> NEC southbound platform]
(* - Another head-scratcher in the IIMP that'll probably have to be revised in the next edition. Stoughton NB platform overspills the junction switches and has to be moved back by necessity; South Coast Rail skips this stop on half its peak runs because of that chintzy oversight. NEC side used to have a Maintenance of Way track turnout into the parking lot, so the junction is already provisioned for a 3 NEC x 2 Stoughton track setup. By moving the southbound platform back and taking a single parking row they can turn the SB NEC platform into a bulb-out and gain an *extremely* crucial Amtrak passing track before Canton Viaduct to eliminate conflicts. Don't see any other way to manage this pinch point, and the alternative for not fixing it is punitive service cuts on the Stoughton side and probable service caps + increased skip-stoppage on the Providence side. Since it's easy enough to construct that turnout + express passer that's most likely what you'll see the 2020 IIMP revision call for.)