Would having regular passenger services to Okehampton be a suitable pilot to gauge whether an actual extension to Bere Alston would see suitable patronage?
If GWR were to sub lease a class 159 from SWR then I reckon that they could run an hourly service without actually needing to have any extra units.
First thing a train which isn't due to enter service until later in the day leaves Exeter bound for Okehampton, undertake the return joinery and joins a Waterloo service an hour before it's normal diagram. Then repeat an hour later for a second unit which is normally diagrammed to enter service on the next Waterloo departure. By the time of the next Waterloo departure the first unit has returned from Okehampton.
As an example (during the day as it's easier to show) all times are approximate:
11:20 departure from Waterloo
14:43 arrival to Exeter (due to depart 15:25, now covered by the unit from the 13:43 arrival to Exeter)
15:32 arrival at Okehampton
15:40 depart Okehampton
16:25 arrival Exeter (now covering for the unit which headed to Okehampton at 15:43)
19:49 arrive at Waterloo
As such there's no need for any extra units is just extra milage and staff costs.
If there's disruption delaying the WofE services then the Okehampton service is cut back to Exeter.
Crediton would see an Exeter bound service at xx:15 and xx:37 (22/38 split) and a service from Exeter at xx:27 and xx:53 (26/34 split), of the was a 5 minute turn around at Okehampton (xx37 departure) with a 3 minute stop at Exeter before heading to Waterloo Crediton would have Exeter bound services at xx:12 and xx:37 (25/35 split) which would give it a more even split of services during the day.
There's other advantages, such Okehampton being connected with Exeter Central without needing another path through St Davids. Also by needing to start a unit before it's needed it would likely see a morning peak service from Crediton which is more useful than the current service.