Basically, I think cities should grow as fast as they can, up to the happy cap, and up to how many productive tiles they can work at a time, whichever is lower.
If they're not working improved tiles, I think it's a good idea to whip workers (especially as EXP where hammers contribute more to them than food does). They accelerate future growth at all stages.
The math is all "it depends" but I don't think any city that could configure itself with at least +5-6 food surplus shouldn't do so.
An simplified example: say a city is 0-growth at size 6 with a plains hill mine. In ten turns that mine contributes 40 hammers.
If it has a 3f farm available, it could work that instead, produce 30 food over those ten turns, grow and whip to size 6 again (this costs 15 food with a granary), and spend the extra 15 food as hammers (by working on a worker or settler). So it gets 45 hammers out of its 6th pop point instead of 40.
This is just a heuristic to show that a 3f tile plus the whip is clearly better than a 4h tile for raw hammers if we have to choose one or the other! It might be that we want to grow up further past size 6-7. It's also true that it's even better to work the 3f tile only when growing, and then switch to the 4h tile when actually building the settler or worker.
(Whip-cycling cities also effectively lowers the happy cap by 1 turn, another complication.)
In practice we really want cities to be much higher than +3 food, which even further tilts the scales from hammer tiles (like PH mines) to food tiles.
I suspect this general strategy is pretty good:
1. At low size work basically as high food as possible. (If it's simply not possible I made a bad settlement choice, or indeed I shouldn't whip as much).
2. Working an unimproved tile? Whip a worker, or I should have manage existing workers better.
3. At the happy cap? Go into a whip / grow cycle, whipping every 10 turns ideally. (Maybe find a way to increase the happy cap.
)
4. Only when cities get really large (>10 for sure) does it make sense to consider stabilizing cities at lower growth values, favouring high production tiles over food tiles.
Now, admittedly RBMod changes the calculations in favour of production, because 10 turn 2x-whip or even combinations of 2x and 3x whips are fairly godly in BtS for really high food cities. So in BtS there's more reason to want extremely high food cities, even +10 or more surplus.
But even in RBMod I suspect it's a mistake to stabilize pop for a city that could be adding on a 10 turn 1-pop whip cycle to whatever else it's doing.
. 