I mean that you lose the ability to be influenced by the “urgent” tasks in the middle of the day. As you know you won’t be stung even if you miss the mid-day deadline, you have no incentive to pay the slightest bit of attention to if you meet the deadline or not.
It’s not entirely clear to me what it would mean for a derail to be non-akratic. Let’s take a concrete example, because otherwise this is too abstract. Let’s say you want to beemind doing the dishes. For the most part, you want to wash the dishes every day, but with emphasis on that “for the most part”. You know that pretty frequently (maybe a few times a week) something will come up, and you’ll end up leaving the dishes in the sink to be done the next day. That’s not the end of the world, of course, so long as you don’t let them sit there indefinitely and never actually wash them, ending up with no clean dishes.
You think about it, and you realize that on average maybe twice a week something of that level of importance comes up. Great: you set the slope of your goal to 5/week. That means that twice a week you can beg off doing the dishes, invoking the “something else is more important”, but then that doesn’t turn into a situation where you have no clean dishes, because Beeminder enforces on you that you’re not excusing yourself for every little thing, just for things on the scale of comes-up-no-more-than-twice-a-week.
Is that the scale you want? Maybe. Maybe it’s 3/week. So set the slope accordingly. This isn’t intrinsically unknowable upfront: from your description it seems that you’re not talking about black swan events like a sudden illness or your car breaking down, but rather fairly normal not-quite-day-to-day work things that there is no reason you can’t plan for.
Putting it another way: if you decide you’ll do the dishes X times per week, then sure, great, whatever value of X works for you, if that’s what you actually plan to do. If you say one value of X, but know that in reality it probably will be a different (lower) value, then say that second value in the first place. What benefit do you get from trying to fool yourself into thinking you’ll do the higher number? And if you do so and fail to hold yourself to this (unrealistically high, and known upfront to be unrealistically) standard, you then say that you feel that the derailment was not legit.
This is rather strange. If you actually managed to fool yourself that would be one thing. (Tricking yourself into achieving your goals is a tried and true strategy.) But from what I can tell you aren’t: you know perfectly well what will happen, and that’s why you’re right now discussing your policy of what to do when it does. That’s rather self-defeating, if what you’re trying to do is trick yourself into accomplishing more.