I suspect you are trying to plot and measure speed with the 3 minute 15 seconds rule, correct?
Well this is good, but do not measure between successive plots! Instead measure between the 1st plot-point and the 5th, so there are 4 spaces of 3.25 minutes in between there. Then measure this distance and divide the number by 4. Now you have a speed value (each 100m represents 1 knot in 3m15s period) that is on average accurate to 0.25 knots. Do it over 10 spaces (easier to dive by) and you have averaged the inaccuracy in speed down to 0.1 knots. Better than you can set the target speed dial.

But 10 spaces take 32 minutes. All it takes is patience.