I definitely don't have all the answers, but I'll post a few things here and hopefully Rick can correct anything incorrect that I've said in addition to filling in the gaps that I leave, although he's on leave this week, so that might take a little while.
Starting with the old R1 offset, my understanding is that for an annual model, the equilibria should be equal to
Unfished equilibrium (associated with startyr - 2):
for a = 0: N_{a,startyr-2} = R0
for a = 1, 2, ..., Amax - 1: N_{a,startyr-2} = N_{a,startyr-2} * exp(-M) [exponential decay from R0]
for a = Amax: more complex stuff happens to account for Amax as plus group.
Initial equilibrium (associated with startyr - 1):
If R1_offset = 0 then the numbers at age should exactly match the unfished equilibrium. If R1_offset is non-zero, then R1 = R0*exp(R1_offset). Otherwise, the rest is the same.
for a = 0: N_{a,startyr-1} = R1
for a = 1, 2, ..., Amax - 1: N_{a,startyr-1} = N_{a,startyr-1} * exp(-M) [that is, exponential decay from R1 instead of R0]
for a = Amax: more complex stuff happens to account for Amax as plus group.
Time series (startyr and beyond):
for a = 0: N_{a,y} = f(R0, SB_y, h, etc.) * exp(-0.5*b_y*sigmaR^2 + recdev_y),
where f() is the stock-recruit function, b_y is the bias adjustment for year y and recdev_y is the recruitment deviation for year y.
for a = 1, 2, ..., Amax - 1: N_{a, y} = N_{a-1, y-1} * exp(-Z_{a-1,y-1})
for a = Amax: N{a,y} = N_{a-1, y-1} * exp(-Z_{a-1,y-1}) + N_{a, y-1} * exp(-Z_{a,y-1})
What's different with 3.30 and the SR_regime option?:
For any given year, including startyr - 1, the R0 is replaced by R0*exp(SR_regime[y]). So to mimic the R1_offset, you need to put a block on SR_regime for y = startyr - 1. But if SR_regime has some change during your main time series, that change will be filtered through the stock-recruit relationship into an impact on the numbers at age for whatever years are impacted. This can also apply to the forecast. A block on SR_regime is also easier than some of the old dummy environmental variables that were created in the past to adjust recruitment up and down for long periods.
Hopefully this makes some sense and I haven't made too many mistakes.
-Ian