Tom Dumoulin 2017 Giro d'Italia
Ryder Hesjedal 2012 Giro d'Italia
Tao Geoghegan Hart 2020 Giro d'Italia
Alberto Contador 2008 Giro d'Italia
Denis Menchov 2009 Giro d'Italia
Arguably Nibali at the 2010 Vuelta a Espana and even Fabio Aru at the 2015 Vuelta a Espana. But Rodriguez simply produced 2 horrible TT's back then.
Leipheimer nearly stole the 2008 Vuelta a Espana in the TT's though.
Arguably Roglic 2020 against Carapaz in hindsight.
Evans at the 2011 Tour de France & Wiggins at the 2012 Tour de France of course
‘In the past 10 years’. Up until 2012, agree there were plenty of GTs weighted, sometimes heavily, towards the time trialists.
Dumoulin 2017 - would argue that that was a fairly balanced GT, and even then Dumoulin only needed the amount of TT there was because of him taking a **** at the bottom of the Umbrail. Without that, he likely would have been in pink going into the final TT with only 40k of TT being raced at that point in a race that had quite a lot of climbing. And he would have lost handsomely had Nibali or Quintana - the only two riders in that GC battle who also ever finished higher than third in a GT - been anywhere near their best.
TGH 2020 - GC was tied with less than 50k of TT done in what was probably the hardest Giro post-Zomegnan, also the only time Hindley dropped TGH on a climb was on stage 3 when neither was even thinking about winning the race, or even about each other.
Aru 2015 - 39k of TT only, Rodriguez couldn’t put any real time into Aru uphill and, as you said, did a very poor TT.
Roglic 2020 - lol, please. There was a grand total of one TT, 34k in length and finishing on freaking Mirador de Ézaro. That Vuelta was, if anything, skewed towards the climbers.
Examples the other way:
Vuelta 2013 - Nibali took 1:29 on Horner in the only TT, sub-40k with enough climbing for Pozzovivo to finish third, and only lost by 37 seconds. In a Vuelta with nine legitimate MTFs plus uphill finishes at Lobeira, Fisterra and Valdepeñas.
Giro 2014 - would have been a lot closer if there was more than a 42k hilly TT to balance all the MTFs and the MTT. Add the time Quintana controversially took in the Stelvio descent and it’s very much flippable.
Vuelta 2016 - extrapolating from the gap in the only TT, Froome needs a shade below 60k of TT (including the existing TT) to flip the GC. In a race with ten MTFs, including the brutal Tourmalet stage, that is still balanced.
Giro 2018 - barely 40k of TT total, Dumoulin wins it with a little more time-trialing in the first half when Froome was riding terribly
Tour 2018 - Dumoulin probably needs too much TT to overturn the lead, but it would at least have been a lot closer if there isn’t only one TT that’s only 30k and hilly…
Tour 2019 - Thomas lost to Bernal by 1:11, having gained 1:22 in 27k of TT. About 50k of total TT distance needed to flip it, which seems a lot more balanced to me.
Tour 2020 - literally 30k of time trialing that wasn’t on a categorised climb. In a race with ten mountain stages, not even counting PBF in that same TT - clearly not balanced. Roglic could have won it with a mid-length flat TT earlier in the race, although I will concede that the main reason he lost was his/Jumbo’s own mistakes.
Giro 2022 - only 1:18 in it in a race with with 28k of TT, if the gap between Carapaz and Hindley in the TTs was slightly bigger that one also flips on a balanced route - as it is, Hindley would most probably have got away with his defensive riding from second place anyway.
TL;DR: the only times in the past decade that a disbalanced route has helped decide a GT was on routes lacking in TT.