correct zerostat technique

I have the Zerostat 3 and was always surprised at the lack of instruction supplied with it. I constantly wonder how to most effectively use it. In the past, a search of the archives turned up a tiangle technique. Basically, squeeze and release at the three points of an imaginary equilateral triangle ~1 foot over the record surface followed by one squeeze(no release) over the center. In max AC season here in Iowa this doesn't seem to be real effective and sometimes I wonder if I'm actually adding charge to the vinyl. Does anyone have a different technique to recommend? Would I have more luck with the Talisman or Furutech destat?(Both of which seem pricey)
I purchased my zerostat relatively recently, with a manual, and it clearly explains that you point it at the record (or whatever object you want to discharge) and squeeze the trigger. Then you point it away and release the trigger. It said the squeeze process removes static charge and the release does the opposite--adds static charge. This seems correct to me--I have used it to statically charge things like balloons. So I think those of you doing the squeeze and release may not be accomplishing much. For me, I often need to do several squeezes, but it always works i.e. the record always ends up virtually free of static.
Thank you Sharpnine. I was hoping I had been using this the correct way for all these many years. I also wondered from time to time if I was doing it correctly, would try both methods. It seemed to me that noticeable dust particles were removed much easier using the point at/point away method, the reason I continued to use that method.
Sharpnine, thank you for pointing us to the instructions in the manual current for the current production Zerostats! I'll have to try this now and my very early production Zerostat.
Interesting, Sharpnine! I guess that would be the correct way to use it. Thanks!

Relax, have a Sprecher Barley Wine and listen to some John Prine...
Thanks for all your responses. I think the variety of replys underscores the confusion that exists on the use of the zerostat. In the end, it seems that the technique suggested by Sharpnine is working best for me at this time. Maybe that will change. Who knows?