So I have one for towards the exhaust and the indexing washer web site says towards the inlet , So its in the middle for me ,I guess it makes sense not to hide the spark plug gap from the incoming charge, so gap pointing any where from inlet to exhaust should help.
Beyond not having found a credible published study of the effect, if any, of indexing the plugs, remember that the valves are nearly half the bore diameter, so there are plausibility factors to consider. Saying the ground electrode should be pointed at, say, the inlet valve to mask the center of the plug still leaves a lot of the "unmasked" part of the plug pointed at the inlet valve. Since the mixture enters through all 360-deg. of the inlet valve, at best only that coming in through the area of the valve nearest the plug could conceivably be diverted in some way by the ground electrode.
But, the spark doesn't happen when the charge is rushing in. The inlet valve closes (the exhaust valve is already closed) and then the mixture is compressed by a lot before the spark finally happens. During this time the charge is free to continue move and/or change how it is moving within the chamber. Even without a deliberate squish band, the edge of the piston comes closer to the head than does the center. Because of this, it pushes mixture toward the center of the chamber as it rises. This motion imparted to the mixture is not a small effect, and it takes place after the valves could no longer be having any effect, but the arguments for indexing totally ignore it and only consider what might have happened relatively long ago, when the valve was open.
When the spark does happen the pressure rises to a very high level as the flame front ignites the mixture in the chamber. Except, through some mechanism that hasn't been explained, we are told that if the spark plug is "correctly" oriented there is an additional 1% h.p. The only way this can happen is if there is an additional 1% of fuel that otherwise didn't burn that now adds to the pressure rise. Since the combustion chamber is a sealed system, why would this 1% burn if the spark plug were pointed one way, but have managed to keep itself from burning if the plug were pointed in a different direction?