is compu polygraph more accurate and sensitve than a older version machine such as those that scribble on paper. does it make a diff towards cms

Polygraphers will tell you that it takes the pass-fail "decision" out of their hands... it's all up to the computer if you pass or not.  Don't blame the tester!

The computer does nothing more, as Bill C. mentioned, than eliminate moving parts.  You have no wires to fray, needles to bend, ink to clot or dry, or paper to get tangled on the spool.

I don't know if the poly industry has tried to implement mathematical formulae into the measurement channels, but I seem to think not.  I have had 2 polys done with a computer, and the interrogator never mentioned anything about having complex math playing a part in his decision.  You would think that he would if anything came into play like that.  However, I wouldn't put it past those people, now or in the near future... except...

I could only see the polywogs seriously trying to implement such a thing on the basis of an amplitudinal (or frequency-based in the case of breathing) standard deviation from whatever "baseline" they would arbitrarily program the poly to establish.  The SD, btw, is a measurement of how far from the mean factors that contribute to the mean diverge, e.g., in 1 2 3 4 5 the mean is 3 and the SD is also 3.

If they could then rig the machine (and that's what they would be doing) to figure if you stayed within, say, 1 standard deviation on BP, GSR, etc. on each question, you would be shown as being "NDI".  Beyond that, I won't even try to guess what they might pull with their "new math".

Given that the ploygraphers, er, polygraphers want to do the analysis on their own, anyway, I find it difficult to imagine that such a scheme would gain much of a foothold.  They may say that the computer does the work, but that's a half-truth to beat them all.  Poly-men still do the tea-leaf reading...