One-dimensional mountain goats actually kind of
should be peripheral contenders though. Just like one-dimensional rouleurs and TT specialists. Tony Martin isn't going to win the Tour anytime soon, right? The purest mountain goats aren't
meant to win unless they're truly special to the level where they can overcome those flaws. Bahamontes won one GT, Julio Jiménez won none. Pantani won 2. Van Impe won 1. Herrera won 1. Fuente won 2. José María Jiménez won none. Charly Gaul is the superstar among them, winning 3.
Yes, in large part the trends in parcours design have reflected changes in cycling:
1) that the domestiques are now much stronger than they were in the past enabling the kinds of rides that those guys put in to only be occasionally possible now;
2) that the climbers and their one-man-against-the-mountains romanticism tend to be more popular than more balanced riders suffering in the slopes (see the '98 Vuelta for the ultimate example of this) so there is a tendency of race organisers to want to have those riders competitive; and
3) you mightn't want to hear this, but Richard Virenque is also to a large extent to blame for the devaluation of the GPM as a competition. In the old days in Spain in particular, the GPM was second in importance only to the victory, leading to dramatic in-race battles over the GPM between those who were reasonably in contention for the win overall and those who weren't because of a lack of balance in their skills. This lured potential GC contenders up the road because of their interest in contesting the GPM and isolated a lot of team leaders earlier. Nowadays in the post-Virenque world, it seems that there is no inherent prestige to the GPM for the climbers, because there is no reason a one-dimensional climber shouldn't be contesting the overall victory, and therefore the GPM tends to go to worthwhile climbers who've lost time early (see Pellizotti Tour '09, Garzelli Giro '09, Majka Tour '14, Landa Giro '17), is won exclusively from the breakaway by riders who are capable but that nobody in their right mind would say are legitimately the best climber in the race (Charteau Tour '10, Lloyd Giro '10, Voecker Tour '12, Clarke Vuelta '12, Pirazzi Giro '13, Edet Vuelta '13, Sánchez Vuelta '14, Visconti Giro '15, de Gendt Vuelta '18), or, especially now at the Tour when to counteract this they've massively over-emphasized mountaintop finishes in the aim of having bigger names wearing the jersey at the expense of making the competition either artificial or killing it completely, being won almost by accident by a GC contender who didn't actively set out to win it (Sánchez Tour '11, Quintana Tour '13, Froome Tour '15, Froome Giro '18) thanks to a points system by which a stage with a Unipuerto HC mountaintop is of more value to the competition than winning every single climb in a stage with an HC, two cat.1s and a cat.2 but no MTF.
If the big problem is that ASO want to break up the monotony of the trains, then they need some super-long mid-race stages along the lines of 'that' L'Aquila stage so that it disincentivises taking the maillot jaune early. The Vuelta saw something we haven't seen in years as neither Mitchelton-Scott nor Movistar wanted to take the pace so the jersey was shipped to a breakaway. This kind of thing used to be commonplace, even Astana in '09 were happy to let Ag2r do the pace making to defend the yellow jersey for a week, and it's precisely for that reason that Pereiro was given half an hour in '06 that let him back into the GC mix; in recent years however, a strong team will take the jersey early and try to carry it all the way, usually successfully. If you make it so that those riders don't want to hold the jersey for too long because their domestiques will be too exhausted, then that opens up new avenues for interest because they'll be looking for somebody to ship the jersey to, and then you've got fights for the breakaway and stories like Thomas Voeckler in 2004 (or even 2011) becoming possible. A series of short stages with easily controllable tempo mountains has precisely the opposite effect - there's no reason to fear any of these stages and you can happily control the race for two weeks unhindered. Inconsistent climbs are a must, also, to disrupt the train techniques. I know that relative to Spain and Italy France has a paucity of these, but it must also be said that the fact the mountains themselves have become brands meaning that the Tour tends to use, use and reuse the same ones year after year after year like it's the Vuelta in the early 90s or something also stifles competition because everybody knows the climbs so well. They may be legendary, but everybody now knows where to dose their efforts, where to watch for attacks, where to preserve their energy. So, for the sake of all that is holy, if you
do innovate or resurrect a long-forgotten climb, do your best to make it relevant (no 2012 Grand-Colombier or Peguère please), and even more importantly
don't spoil the surprise by doing it in the Dauphiné a month earlier!!! Because if riders have only ever done it in training, not in race conditions, they may have come to different conclusions on how to ride it, how to pace, where is best suited to attack. And there's no novelty, no excitement, in it if you've already seen it just a month before.
Now I know that France's best climb for being super hard AND being inconsistent is Croix-de-Fer which is already one of those climbs everybody knows like the backs of their hands, but there's always inconsistent climbs or steep ones in the Pyrenées, and there's much tougher climbs in the Massif Central than have been used in years - where's Pré de la Dame, Col de l'Œillon, Col de la Lusette, Col de Finiels, Col de la Baricaude, Col de la Mure, Col de Charousse, Croix de Chaubouret, Col du Chansert, Col de Béal, La Loge des Gardes, Baracuchet North, Col de Ceyssat, Saint-Anastaise, Pas de Peyrol west with its 2km at 12%?
Sure, many of these aren't likely to be especially decisive on their own (save for maybe l'Œillon, Lusette and Béal, though Col de la Mure worked ok in Paris-Nice) but you could make a really tough transitional stage, one that nobody wants to have to control, rather than the stages into and out of Saint-Étienne we have. You remember the stage that TV Tommy got the maillot jaune in in 2011? That's
not a bad example of the kind of stage I mean. The Giro is better at these, admittedly, with things like
L'Aquila 2010. Or, France could use its generally not-very-interestingly-used-by-Le-Tour southern tips for this kind of transitional stage. Paris-Nice does a much better job in this respect. Remember Contador bonking on
this stage? That would be perfect. No individual climb tougher than Bourigaille, but relentless up and down all day, for nearly 200km. It
can be controlled, but unless you give the break 25 minutes (in which case everybody's going to know it's a day for the break and you'll have loads of people with different skillsets trying to figure out how to win the stage, which will make the battle for that interesting, plus of course the GPM throughout) you're going to pay for the effort taken to control that stage in the mountain stages to come, especially if there is a long mountain stage in the midst of that. I'd go for the following trifecta:
1) stage finishing on climb that is tough enough on its own; as the first mountain stage of 3, people won't go from afar here, so make sure the final climb either has the prestige, is steep enough or otherwise difficult enough that there will be time gaps regardless - Mont Ventoux, Mont du Chat, Col du Granon, Grand-Colombier, Joux-Plane, Plateau des Saix, Super Collet, Chamrousse via Luitel
2) 200km+ stage with several climbs, either finishing on a descent or on a slightly easier summit than the penultimate climb - Galibier-L2A, Romme/Colombière duo, Bonette-Auron, Cayolle-Pra Loup, you know the drill.
3) THIS is where you put your short mountain stage. To complement your strong mountain stages, not in place of them.