I think that in my version of the game Matthew Goss is very much overrated when it comes to climbing, even tough he's only 65 at climbing.
He can easily follow average climbers like Valverde, Rolland and Cataldo on cat. 1 and 2 mountains. He even won the 5th stage of Paris-Nice (real top 3 was Westra, Valverde, Wiggins, to give you an idea) and got 6th in the Vuelta a Catalunya.
Does this problem accur to anyone else?
Well, two main factors:
First, you said his "climbing" is 65. I believe you were only looking at the MO stat, but the MO stat is not the only influencing a rider's ability to climb, there's also the Hill stat (and the secondary ones, such as stamina, resistance, etc). Specially in these early races on the year where there are hardly any big mountain stages like the ones in the GTs, the HIL stat is usually much more used than the MO. Each stage has a ratio of how much of HIL is used when climbing and how much of MO (see here). The Mende stage from Paris - Nice, f.e., has 0.43 on this, meaning that the HIL stat will be more relevant that the MO stat. Additionally, because of the 1:10 scale the Mende stage is probably easier in PCM compared to real life.
Secondly, it's probably related to his form as well. P-N and Catalunya are very close in the calendar, so it'd make a lot of sense that if he had great form in P-N he'd also have it at Catalunya. And both these races are 'around' MSR, so it'd make a lot of sense if his calendar is built with MSR as one of his goals and thus he's in near-perfect form by then. Specially in these earlier parts of the season, form is a big factor on race results because not few riders are in good form.
Lastly there's also the daily form, but this isn't the best explanation since he had good results on several hard days, and not only one or two.
Anyway, if you think Goss is climbing way too well you could always decrease a bit his climbing stats using one of the editors from the downloads section