I always thought that Hold function was supposed to be for when a measurement requires both hands, and full attention to the probes. The DMM was supposed to figure out when the measured value is stable enough, than freeze the measured value for a few seconds, so to give me time to disconnect the probes then watch the display.
My Metrix mtx3283 doesn't work like that on either hold or auto-hold. Hold only freezes the display at the moment it was pressed, while auto-hold doesn't seem to be doing anything. It only holds the value if it differs less than +/-100 counts, on a 100_000 counts DMM.

That's an expensive DMM, it was the top model from that line. It has No-Hold, Hold and Auto-Hold, and it was $800-$1000 new, don't recall exactly.
I remember having the same frustration with some Fluke at a former workplace. I don't recall what model of Fluke it was, but I recall the frustration that it didn't hold the value until it was safe for me to read it. It required two dudes, one to keep the probes, the other to voice read the display.
What's the point of having hold or auto-hold in DMMs? Ever seen a properly working one? Does anybody uses the hold/auto-hold in practice, or hold is in fact just a useless feature, good for marketing hype only?
Why would I need Auto-Hold if I can afford to look at the display while probing?
And why would I need Hold anyways? mtx3283 is slow, 3 measurement/second or so, it's not that the display is so fast that it can't be read.
