I think it's important for any discussion of carbs to have some mention of glycogen. In most of the paleo-ish literature I read there is precious little about glycogen and I don't really understand why. Take Taubes' WWGF for example. He goes into tedious detail on many things but only mentions glycogen I think 5 times in passing. For someone who is talking about insulin as a primary agent in adiposity, he should be qualifying a lot of his statements with "as long as glycogen is saturated, insulin will...."
Which brings me to the question at hand which is the potential risks of oscillating between low and moderate carb. If you are going in and out of ketosis, like 1 week on, 1 week off, it would be unnecessarily stressful and I would wager unhealthy in general. On the other hand, I have experimented with alternating days and it works really well. On days where I deplete glycogen (MWF in my case) I eat a fairly substantial amount of sweet potato. On days where I don't workout and I'm just walking around, I get pretty close to zero carb. Sometimes I'll feel my liver glycogen dip too low on off days and eat a little bit of carbs, but for the most part, it's an on/off alternating. You end up efficiently recharging your glycogen stores while taking advantage of the anabolic nature of insulin. There's a reason why bodybuilders (stupidly and dangerously) inject insulin ("slin" to use their vernacular) when they're not diabetic.
Glucose gets partitioned preferentially as glycogen because it's a more efficient process (7% ATP loss vs 25%) compared to storing it as fat. Your body will do this all day long if you are continually depleting and repleting glycogen. The truth is that most people aren't doing this to any great extent. You're depleting your liver glycogen constantly, but it's a shallow reservoir of about 100g and I suspect that our consumption of ketogenic fats leads to ketone-body utilization by the brain etc. even in the presence of sufficient glucose, so the exhaustion of liver glycogen is slower (this is simply a hunch based on experience).
If you were able to perfectly predict the saturation status of your glycogen stores and top them up to 99% all the time without exceeding that point, I doubt that you would gain any measurable amount of fat ever. There must be some mechanism whereby excess dietary fatty acids are stored as TGs in adipocytes, but I'm not aware of it. For what I do, the lack of insulin on what are basically zero carb days causes a downregulation of insulin to the point that adipocyte lipolysis suppression is largely halted. With a protocol like this, it is very easy to lose fat while gaining muscle (if you look at a few days at a time). For what it's worth, I feel great doing it.
I advocate pitching the whole idea of X # of carbs/day and instead thinking of it as X # of carbs needed on Y day to recharge Z glycogen depletion. If you lift weights or sprint on a particular day, feel free to recharge with more carbs. If you're just sitting or walking around, even a lot of walking, you can get by with very little. Over time, the latter situation should result in a gradual depletion of most glycogen which could necessitate a weekly recharge or something, but if you're working out every couple days, it's safe for fat loss to load up on those days. The amounts will be highly variable and will require experimentation of course.
