Still not saying anything too detailed on the ins and outs of Israel/Palestine, but this piece gets a lot of credit simply because it ignores everything except the raw data. Said data says Israel has breached the peace 80% of the time such a peace could plausibly be said to exist.
I did wonder whilst reading the article whether it's fair to judge a ceasefire only ended by a confirmed kill, rather than by an actual violent act, but I can see the point that one rocket per month from a area holding 1 500 000 people, many of whom are furious and desperate, probably doesn't count as an intentional attempt to ignite a conflict by the Palestinian Authority. From this data it's impossible to know whether each rocket had the blessing of the government or not, but certainly had Jean Charles de Menezes been carrying the bomb the police apparently thought he had, it wouldn't have meant a declaration of war by Brazil.
Moreover, the second commentator also reminded me that while the metric used may seem unfair to Israel (since their attacks are better organised, and use better technology), the flip side is that you could also argue for including the deliberate contribution to a death through inaction, which would mean considering military strangleholds on infrastructure and imports.
Given that, my first instinct is that it's pretty much a wash, in some respects, and that a death through "military" action is probably not a ridiculous metric. Certainly what the graphs show us is that when there are rocket attacks in sufficient number to cause casualties, then four times out of five there's already at least one dead Palestinian on the ground, killed by the IDF. It's been obvious for a long time that the Palestinians take far more losses than the Israelis in these matters, but this suggests that the Palestinines often take the first losses as well.
h/t to Mahablog (LOTS).