I advocate using the Keith Count. It's easy to learn and employ and it's one of the most accurate for normal races. That being said, it never hurts to double check your work with another count like the EPC or Trice or Kleinman or anything. If you have the knowledge check it against other counts. If they end up disagreeing figure out for yourself which one you trust more or why one may be off in the particular position you're looking at.
In my lesson on the Keith Count I give this position.
                        |
| ||||
| XGID=-----CFACA-Aaaa-abaeba----:0:0:1:00:0:0:3:0:10 | |||||
| Blue on roll, cube action? | |||||
eXtreme Gammon Version: 2.19.206.pre-release
If we do the Keith straight up we get 117 to 112 for a difference of 5 making it a no double. There are two separate things going on here and the first is where we can't allow our common sense to go out to lunch. We penalized the on roll player for having no men on the four point. Guess what, that's plain dumb, he's going to put checker(s) there without much of a problem. We need to realize that and figure it in to our analysis now. The same with your position. Most of the time we will get a checker there. (and if we don't with something like [66 44] so be it, great!)
The second thing I'd note is that while the flat Keith Count says it's technically a bare no double, by one pip, and if we assume our four point isn't void it is a double but not likely a big one, is my adjustments to using the Keith Count. I have stated in the past that at the racing length of 80-100 pips we can double exceeding the opp's count by 5, at a length of 100-120 by 6, and 120+ by 7. This race is just inside that window but certainly could still have an effect.
Stick