It really depends very heavily on the playstyle, if we’re being objective.
As the specific engagement scenario matters a lot and as such without specific examples it’s hard to determine.
I could run the numbers, but I doubt that it would sway many people, and would need a significant quantity of data, further a heatmap on a per map basis would be nice to have to allow for strict metrics to be produced on a per map basis as to workable positions and their map coverage and their potential impact on a given round.

Since as with the survivability onion if you aren’t identified, while holding an angle or don’t try and counter-peak someone looking at you there is no reason to need to return to cover. At least in a 1 on 1 engagement. Since, a pre-ranged first shot should be expected for a good position;
As to what you should expect from a 1 vs many encounter, do you really believe that it is fair to expect to be winning those without a misplay by the opponent? And as such be the metric to judge the impact of reverse speed.
If you do, you expose yourself for a short period time, which is dependent on the Hide’s gradient (-5 means locations with a grade of less than ~8.75% are suitable (8.75 meters in altitude per 100 meters of travel) for the tank in question, -10 is ~17.5%) , the tank’s speed, acceleration, gear ratio’s and “height over bore” to effectively return to cover, crew’s driver’s skill all make the difference.
4km/h is ~1.1m/s and Height over bore is about ~3 gun diameters so ~0.4m (125mm * 3) of vertical distance needs to be covered thus ~4.6 meters of incline does as a rough approximation it would take T-Series a an approximate minimum (assumes instantaneous velocity change occurs, ignoring acceleration) of just over 4 seconds (in 5 seconds, the T-series covers ~5.5 meters, so Grade limit is ~7% for a covered reload) , though a lot of positions don’t allow for maximum performance, which narrows viable options, It’s still less than a reload for the majority of tanks that will be faced with any regularity.
Note that the above calculation presumes that you only need to reverse, numbers may be fairly different if peaking or otherwise outside a vacuum, Also this assumes well aimed shots and so precludes accuracy and post penetration damage assumptions as this will change from player to player, but not where the breakpoint falls is as to which is better off for any specific individual encounter.
Further if you really wanted you could rev the engine and subsequently manually shift out of neutral to get best performance, but that’s a lot of work for very little actually improvement.
And is it anything is it not on the player to know and work around the limitations of their lineup? Outside the obvious “erroneous” modeling of their vehicles.