I only donated to one project, which I'm both interested in and I think the developers would actually deliver a quality game. I'm not expecting the best game ever, but for $10 I don't think I'm losing that much if I don't get what I'm expecting.
I would also only be slightly disappointed if I lose that money.
Like with anything, this is a gamble and the best advice with gambling is that you shouldn't gamble more than you can afford to lose.
I set priorities and while $10 isn't nothing, it doesn't even buy me dinner, so I'm losing less than I can afford. That's worst case scenario. Best case scenario, they release the game and it blows my mind.
Most likely case is that they release an OK game I enjoy for 10-15 hours, so that's less than a dollar an hour of enjoyment. I'd say that's good value for money.
Now you should add weightings to these cases, if you really want to act like an insurance guy. So let's assume that for this particular game (but this can be applied to anything) the chance of the game blowing my mind is a paltry 5%. The chance of the game coming out and turning out OK is 35%. Chance of game coming out and turning out bad is 35%. And chance of the game not coming out is 25%.
Now you add worth in terms of hours played for me. 50 hours if the game is great (3 replays). 15 hours for OK game (1 playthrough). 2 hours for bad game(tutorial and first part of game). 0 hours for no game.
5% * 50 + 35% * 15 + 35% * 2 + 0 = 2.5 + 5.25 + 0.7 = 8.45 hours of gaming.
So this means that on average this game is going to provide me with 8+ hours of gaming for $10. Of course, this is my own calculation, but you need to skew the numbers very badly to get less than 2 hours of gaming for $10, which is the price of a cinema ticket
The people who put in 1000 dollars, I don't understand, but to spend $50 or less on something you think might be amazing doesn't sound too bad.
My 2 cents, pennies and yens.