David Vizard writing on another person's Quench testing in 0.010 increments:

http://www.enginebuildermag.com/Article/2266/surfacing_equipment_can_increase_horsepower.aspx

sample quote

I knew that minimizing the quench area clearance was also good for normally aspirated engines as well, and had it in mind to do some tests that would establish just what it may be worth. These got put off time-after-time because it is a time consuming and relatively costly experiment for a one-man business to do.
Fortunately, while lecturing at the SuperFlow Advanced Engine Technology Conference in 1998 I met a gentleman who had actually done these tests.
As I remember it, the 350 SB Chevy test engine was in the 380-400 hp range and started with a CR of just 9.9:1. After baselining with the piston-to-head quench at .065; (.040; for the head gasket and .025; down the hole for the piston) the quench clearance was reduced in steps of .010; until the quench clearance was .035. When you are looking at potentially small changes in power data, scatter is a consideration. However, by averaging out the numbers the results appeared to consistently indicate that each .010; reduction in quench clearance was worth 6-7 hp or 5.5 to 6.5 ft.lbs. of torque.
The .030 change in quench clearance increased the CR from 9.9 to 10.6. So I made some relatively sophisticated calculations to establish what the increase in output from this factor alone would be.
The answers I came up with indicate that, at best, the increases seen were only 60 percent attributable to the increased CR.
Other issues of note were that the engine needed less total timing to make the power seen. Along with this, the ignition swings in the advance direction indicated it to be no more prone to detonate even with the higher compression.
......So how close can the quench clearance be run?
Good question.
A friend of mine who ran a machine and dyno shop until recently has, using good stiff race bottom end parts, run down to .022 before contact was seen. I have run a SB Chevy with stock crank and rods down to .025 where it showed that contact had just occurred. This was a motor that peaked at 5750 rpm. In practice I usually build a typical domestic V8 for a customer with .035 total quench clearance and that seems to be safe so long as the pistons are close fitting in the bores. For my own motors I typically go to .028 when using a known combination of good bottom end parts.