OK, this is what I understood from De la Rosa: a 2-stop strategy is 10 to 15 seconds faster than a 3-stop one. RBR was correct in thinking that they needed to split their strategies if they wanted a 1-2, because getting one of your cars to get past Grosjean on the same strategy was achievable, but getting both was much more difficult - so you needed to try something to destabilise him. As we saw on the race, it worked to perfection; it would have equally worked had the RBR cars's roles been reversed, ie, had Vettel been given the 3-stop. What was also to be expected was that your 3-stop car would never be able to overcome the strategy disadvantage, and end up ahead of your 2-stop car (which the 3-stopper would have had to pass on track in any case, and Suzuka is known as not particularly easy for overtakings). So there you are: RBR knew, from the moment they put Webber in a 3-stop strategy, that he would never end ahead of Vettel.
It sounds odd, but I think I have to challenge De la Rosa here, since I can't believe that a 3-stop is 15s slower than 2-stop. Let's do the superficial math: 3 stops mean that your tyres are on average 4.5 laps fresher (calculation: 53/3-53/4) throughout the race. If a 1 lap fresher tyre gives you 0.05s advantage, then this amounts to 12 seconds in the race (53*4.5*0.05), which would make a 3-stop only 10 seconds slower than a 2-stop. Using a 0.05s advantage would be very much on the conservative side already.
Now let's look a little deeper: How much does a fresh tyre give you realistically? For this, you can roughly look at laps of a driver, especially before and after the pit stops, and take fuel into account. Fuel information can be found here: http://www.f1fanatic...-circuit-japan/ (time penalty per lap of fuel is ca. 0.112s). Lap time information can be found here: http://en.mclarenf-1...r2=Mark Webber# I didn't do the full maths, but even if you roughly compare the laps, you can see that they don't get much faster within a stint, which suggest that tyre deg is not much smaller than the time penalty per lap of fuel. At around 0.09s deg (meaning that a car is 0.4 faster at the end of a 20 lap stint than in the first lap), 2 and 3 stop would be balanced over the race.
I know that those things aren't linear and I also know that, in practice, you have to factor in more traffic on a 3-stopper. That is the main reason why team simulations might say that a 3-stop is worse. However, due to the large gap behind the top-3 traffic behind them was not a significant issue. Considering that, it seems that a 3-stopper isn't actually slower than a 2-stopper - at least not by a significant margin.