The Yankees should have learned their lesson about long term contracts and try to avoid them whenever possible while overpaying in the short term, if necessary (The Sox obviously had a lot of success with this in the 12-13 offseason and still are having success with the plan, not on the field this year, but with payroll flexibility). This mean extending the QO to Robertson would be a no-brainer because they can easily live with either result: he takes it and they overpay for one year (but only one year!) of a "proven" closer (I use this term loosely) or they get a draft pick when he signs elsewhere.
From a risk management standpoint, one year of Robertson at $15 million is better than, for instance, signing him, or a similar player, to a 4 yr/$35-40 million contract. The first deal has no chance at having much of a negative impact. The latter does, as we have seen time and time again with Yankee and other long term contracts.
Small market teams have to take chances with longer contracts because of strict yearly payrolls. Teams like the Yankees (and the Sox, too) shouldn't if they don't have to, and with Robertson they don't.
The other financial advantage the MFY have here is they can overpay for a closer. Giving Robertson $15 million represents about 7% of their budget. If, for example, a team like my Rockies signed Robertson for the same price, they would be spending about 15% of their budget on a closer, which would be foolish to the extreme.