That's a consequence of the limiter having been introduced as afterthought on engines designed originally to rev quite a bit higher than 18k and then allowing only small retuning (but without changing basic geometry) during the freeze. When you have an engine designed to rev up to 19.5k or more, and meant to spent most of its WOT time in the uppermost 2-2.5k rpm, limiting to 18k means disallowing access to most of the rpm range it was designed to use, so power near the limiter is still climbing quite steeply, being far away from where the peak power originally was (roughly 500-600rpm under peak rpm, that for most was around 19.5k-19.7k, notable exceptions being Ferrari, lowest revving at max 19k for races and few hundreds more in qual, and Mercedes + Cosworth, both approaching 20k in qual).
FWIW this is a collection of the total gear ratios (expressed as speed [km/h] @ 18k rpm) used by Vettel (plus Spain data from Webber) since 2011:
Consider that these come from looking at the engine rpm drop in upshift (engine rpm extracted from engine sound) and are obviously subjected to error, especially in the shortest gears (where engine rpm variation is quicker hence precision is limited) so when the difference between two ratios is of just couple of km/h, you can safely take it as the ratio is actually the same (as probably is first gear in most of cases). Obviously when there's no indication of first gear it's because it's not used during the lap in that track.
As everybody should know then, Red Bull traditional tendency is to gear towards the short side, so maybe these aren't really representative of the average car, but it's the most complete set of data I have for same driver/car (the onboard laps more likely posted on youtube are for pole position, and Vettel had a few recently...) so I went for it still.
As for next year selection I share marcush's opinion, not big issue, with 8 gears available it shouldn't be difficult to cover all needs, it would be possible with current engines, let alone with next year's that should have a more favorable power delivery.
For instance, if you look at the data and compare Monza with Monaco, you'll see that the first two gears are basically identical, while 6-7 of Monaco are quite close to 5-6 of Monza.
With 8 gears having something that works for both tracks is far from difficult, just take Monaco's ratios, add as 8th the 7th of Monza and it's done... the problem is if anything use in other tracks these extremes as you'll end up having a 7th that is too short and an 8th that is too long, so in reality a more useful starting point could be probably based on something like Bahrain ratios for 1->7 + Monza's 7th as 8th, and if it comes a bit too long for Monaco, who cares, so be it, it's one race out of 20.
Obviously teams will easily find the solution via optimization from lap simulations etc. compared with other stuff they have to optimize that's barely worth mentioning.
