NavList:

Tony wondered "how accurate was DR with traverse boards and those tables?"

I don't know much about traverse boards, but guess they were mostly used when sailing by the wind, where the course was more or less constantly changing due to variations in wind direction and wind speed. Then perhaps 1 point resolution was sufficient. Otherwise the helmsman was ordered to steer a given course.

Steering compasses were usually graduated in quarter points (2.8°) which is about as near as you can steer a sailing vessel. And traverse tables used to have a resolution of a quarter point. One exception is Roswall from 1824 that has a traverse table for every 1/8 of a point. 

The main uncertainty was, I think, in logging the speed with a chip log. The distance between the knots on the line varied a lot. Reworked for a 30 seconds glass, Norwood (1636) gives 15.24 m, Roswall (1824) 14.84 m, Bowditch (1851) 14.70 m, Pettersson (1861) 13.95 m, Nares (1882) 15.38 m, Kusk Jensen (1924) 14.12 m. And how accurate was your 28 s or 30 s glass? How did you calibrate it in early times?

So the accuray of DR was not that great. 
Lars