Roast turkey

There are many ways to roast a turkey.  I'm sure there are a number of ways to get good results.  There are also many different kinds of turkey, which react differently to different methods.  We had a 19.3 lb. fresh turkey from Ashley Farms.  I was reading up before Thanksgiving on the recipes I'd used before, and other advice on the web.  There are a number of decisions to make: to stuff or not to stuff?  To baste or not to baste?  To truss or not to truss?  Cover it with cheesecloth?  High temperature or low temperature, or high and then low?  Some recipes may be suitable for an injected supermarket turkey but not a natural bird.

The main methods I considered were from the Joy of Cooking (1975 edition), which has you put the bird into a 450°F oven and immediately turn it down to 350°; Barbara Kafka's high temperature recipe which has it at 500° F the entire time; and Alton Brown's brined roast turkey which starts at 500° for 30 minutes before being reduced to 350°.  The other common method, to keep it at 325° the entire cooking time, is also mentioned in the Joy of Cooking, where it is said to be foolproof and "has been rumored to entail much less shrinkage", but that the flavor is superior when starting with high heat.

I decided to go with 450°F for 30 minutes, then turn the oven down to 350°, unstuffed and untrussed.  The Joy of Cooking says 13-15 minutes per pound, but since I would keep the oven hot a bit longer and was aiming for a lower temperature (the USDA recommendation is now 165°F interior temperature, it used to be 180°F), I was ready for it to be a bit shorter than that.  I wasn't prepared for how fast it went.

Since I had just acquired a remote temperature probe, I recorded the internal temperature at irregular intervals throughout the cooking.  It turned out to have been done in just over 2 hours, about 6½ minutes per pound!  I then had to keep it warm for a couple hours and warm it up before serving.  It was very good, though, tender and juicy with a nice skin.

This is not something I do often enough to do systematic experiments, but here is the time series data for this trial: