when I detect I'm missing position, so I can use the external hardware I have to move back to where I should be
Since I would only be a few pulses out of position

What i have found is that the position indicated based on "pulses" by the Mach DRO is reliable.  If the PC is doing it's job
 then Mach is doing it's job in reguards to the position. Pulses relate to position, but......there are other influences both
electrical, mechanical , and axis tuning which come into play. When you say you are missing a few pulses, I ask
 myself...... how does he know? What is the basis of a "few"?

Let me clarify based on some testing which we did when doing threading tests.

A pulse counter ( measures to parts per million / custom made ) was hooked up to simulataneously measure pulses in five places
and could show the delta difference between the readings ( "P"  in red). Three indicators of axis location ( blue color) are shown ,
but a measurement based on a standard was done. Now to relate all the data can be tricky since the axis system comes into play.
So one needs to know what he is relating the data to and the variances of the all items in the "system" .  

Simply put, when you get into finding or measuring small number of pulses/ movement it can get very demanding and most will not even
 have the know how or equipment to do it ( including myself ) correctly.

A little off the thread .......See attached pic 

RICH

I have a shaft encoder attached to the end of the stepper. Both the step signals and the outputs from the rotary encoder are fed into a microprocessor. From the difference between commanded step ( from Mach3) and the actual rotary movement of the stepper shaft I can subtract the two to work out a missed step. I have run the combination stepper/encoder at a step rate that equates to 4 times faster than Mach3 would ever step the motors and can confirm after a few hours of running ,I have at most lost 2 step in many hundreds of thousands. it when the stepper is installed in the machine and the additional effect of inertia and mechanical loss come into play that I notice the occasional missed step

Since I trying to get the best accuracy/repeatability any loss of position is undesirable