Couple of comments: 1) W/kg refers to body weight in normally hydrated state (i.e., don't weigh yourself after Stage 1 of this year's Epic and use that number for calculations), but if you are trying to improve uphill speed then everything benefits from decreasing weight. 2) A lot of folks do 2x20 or 2x8 (see Chris Carmichael for the math and underlying trials to support these) primarily because if you do a 1x60 properly your body will be jello for days afterwards, which is an unnecessary waste of potential training time. 3) Outdoor vs indoor: I can suffer more outdoors than indoors. I don't know why. So I use my outdoor figures to set training loads for both outdoor and indoors. Keeps me honest. 4) Racing vs controlled environment: nothing wrong with racing in principle but you need to avoid sudden accelerations or decelerations that will skew the results. Racing isn't always conducive to this. 5) Hills vs flats: as long as you can maintain a steady aerobic cadence and pace (see point 4 above), either is fine. I haven't seen any science that precludes hills, unless the hills are so steep that you dip into your anaerobic systems just trying to stay vertical. I've tested on hills up to 8% gradient and the numbers work. Last comment: I suck at testing. I can consistently do 2x8 or 2x20 efforts randomly during rides that are higher than my 'testing' numbers. I don't know why. I get all anxious and just end up blowing it. The point is: you can't fake your aerobic capacity. If you are tracking your data in Golden Cheetah or whatever, it will be pretty obvious over time where exactly your power thresholds are for each Critical Power point. So a lot of coaches (including mine, since I suck so bad at testing) end up forgoing 'tests' and just monitor the entire CP graph over time. When capacity increases, they push the training loads.
