To measure total time played stopwatch is good enough anyway, or online stopwatch, or run counter, whatever works for someone.
Problem with run counter is it doesn't register first run you do after you pause it or restart it. However if you pause run counter, open char in gomule/atma, then resume run counter, it will start from 0:00 current run and that first run will be registered normally. Don't worry though, total time is registered normally on run counter regardless of conditions.
Example: you start run counter and do 2 runs where both are 60 seconds long. Run counter will show you did 1 run with 60 seconds average time and 120 seconds total time. He registered first run only in total time invested, and didn't count it in average time nor in runs done. That's why players who wish to know exact number of runs they did, they simply s/e first run immediately and continue running normally after that.
Example: you s/e first run which took like 10 seconds total, and then do second run which was 60 seconds long. Run counter will show you did 1 run with 60 seconds average time, but 70 seconds total playing time because run counter registered you did s/e first run which took 10 seconds, but didn't register that run ever happened in number of runs done nor in average time per run.
That's why we usually use average time * runs done as a measure of invested time rather than using info of total time played from run counter. So, since I will do this s/e thing this RFO because I want to know exact number of runs I did, that means my total playing time in the end will be over 21 hours on run counter, while real playing time will normally be 20 hours as per rules. I will lose that 1 hour on s/e procedure while not doing any runs in the process. And in general that's the reason why when you divide total time displayed on run counter with runs done, you will NEVER get correct average run time nor will you get average time displayed on run counter.
So, there are two correct methods of keeping track of total time:
One method is to start run counter and simply run. Total time will be displayed correctly, but average time and runs done will be incorrect, but irrelevant for this RFO.
Second method is to s/e first run anytime when you restart or pause-continue your run counter. That way you will have correct info about runs done and average running time, but total time will be wrong since it included all those seconds when you s/e first run (didn't run), so total time should be calculated with avg. time * runs done.