Good example of this? Rolling for the door when people are Muscling Through (DF2, p. 8).
I don't roll for doors anymore.
Instead of rolling for the door, I just assume I rolled a 10. This allows me to pre-calculate the margin of success the PCs need to force the door. I apply that as penalty, usually in secret, if they don't/can't examine the hinges. They roll, tell me how much they made it by, and I announce success or failure.
Optionally, you can smooth these to 0, 10, 15, 25, 60, as shown in the third column. This just makes them a little easier to remember.
Obviously, you won't force a vault door easily. Maybe ever. Those are better dealt with magically or by bashing (IOW attacking the door.) This is true with the original rules - it's not likely for the door to roll an 18, and you'd still need to beat your own roll by more than 52 to pull it off.
Bending bars is the same (not lifting portcullises, though, that's a straight-up comparison of BL to weight to see if you can lift it or not.) For bars:
Optionally, you can smooth these to 10, 15, 25, 45, 60, as shown in the third column.
What about barred/wedged doors? Bars make it tricky, because they replace one or both of the numbers for DR or HP of the hinge or lock. A complex solution is to figure out the net score for the barred door using the better of the DR and better of HP of each of them. A simpler one is to use the bar's number if it's higher.
Simple approach: Use the door's force number or the bar's, whichever is higher. If the door is the same or higher, give it a minimum +2 for the bar. For example, an average door (8) with a light bar (5) would get a +2 for a net 10; with an average bar (17) it would use the bar's number instead. With smoothed penalties, the average door (10) with a light bar (15) would be a (15). You can also use this for improvised bars and wedges even if you use the next method - this avoids calculation in play.
Complex approach: Figure a combined number for both, using the higher of the two DRs plus the higher of the two HPs. The average door with the light bar would get a 10, with an average bar it would get a (20). This is best done ahead of time.
Actual Play Feedback: This speeds up play a surprising amount, since only one margin of success matters. I don't need to roll or do any secret calculation. It's just a matter of looking at the door's number and what the MOS of the PC is. It's been pretty good.
It also feels better. No longer do you try a door once, roll a 9 and have it fail (because I rolled well) and then try again, roll a 9 (or even higher), and this time it works (because I rolled poorly). It takes an active chance out of what feels like it should always be a passive trait of the barrier.
So yes, in actual play, it works as intended!
There, one less die roll per door.