I agree with the thrust of the article, which is that the decision to buy insurance should be driven by a sensible evaluation of the probability and magnitude of future losses.

That being said, one topic that could benefit from further discussion is the interplay between the deductible (or retention) and the cost of the policy.
Insurance is more economically beneficial where the focus is on "catastrophic" losses - like a home destroyed by fire. Milton Friedman touched on this in "Capitalism and Freedom" when discussing health insurance. One does not insure one's car for needing to have the tank filled each week, but rather for the car being totaled in a collision. Similarly, Friedman's view was that health insurance would be more beneficial if it provided "catastrophic" event coverage (today, that might be anything costing more than $5,000 in a year, for example) - but that coverage for routine checkups and ER visits did not make sense because premiums would have to be inflated accordingly. (A full discussion would also consider the incentives of having a portion of "retained loss" - healthier living habits, perhaps!)

Coverage for "mundane" losses of small magnitude but high frequency add to the administrative cost (the expense part of the combined ratio) and to the premiums. In a way, insuring such items is a "premium multiplier" because one pays premium not only for the actual losses, but also for the insurance co's expense in administering the claim, thus "multiplying" the cost of the ER visit.

For all these reasons, the decision is not just whether to buy insurance, but to set the deductible at a level that avoids "trading dollars" with the insurance company for routine minor losses.