There's hysteresis and then there's stiction. I understand hysteresis in mechanical movements is represented by the difference in the contact force in the postive direction Vs the negative. If an indicator develops 10 grams of contact force going towards the work and 9 grams going away then 1 gram is the difference and represents the hysteresis.
Stiction is the product of elasticity in the measurement system deflected by hysteresis.
Whenever I try for refined measurement using apparatus with mechanical movements or with set-ups having some elasticity I've been taught to "dither" it; that is apply a small vibation or shick to settle the system to its lowest energy state. My usual dithering tool is a wood pencil I use to tap someplace on the set-up.
If you know the calibration error, the behavior of the instruments and apparatus under various conditions, and have a good sense of Kentucky windage it's possible to work well within the minimum advertised specifications of the available equipment.
Jo blocks for example are calibrated and their errors are logged in the calibration sheet. Using the calibration sheet and a little arithmetic, the stacking error can be determined. Simply add the errors correspnding to the blocks assembled and add/subtract from the stack.
I've used plain vanilla dial indicators in large awkward horizontal set-ups. If the weight of the assembled dial indicator set-up is known it's possible to determine deflection from gravity using a sensitive spring dynamometer.
So if you are looking for ultimate in accuracy and reliability, remember that much depends on how the apparatus you're considering is employed and what measures you take to compensate for system errors from mechanical stiction, temperature, elasticity, gravity, etc.
All that said Mahr makes first rate equipment and its accuracy and reliability is in part measured by the price.
I have a 50 millonths graduation 0.0020-0-0.0020" dial indicator, a clone of the famous Mahr "Comparmess"(?) made in China sold by Shar. As near as I can tell using my Jo blocks and a granite base comparator stand the Chink indicator is on the money going in 0.00005 increments. And it const $115 Vs over $400 for the good one. Is it any good? I don't know, fine apparatus is a little out of my general expertise.
So what you buy depends on how long you expect to use it and what confidence you require. And don't forget annual calabration and serviceing if you have to meet documentation and third party verity.
High accuracy, high confidence readings is not for the timid or the poverty stricken.