Yo,

Just to remind non-believers...
Each bow, each limb setup and each arrow shot with any point/head, any weight, any vanes etc. behaves differently.
I have seen a bow energy graph in detail and know a lot of physical principles. Let me enlighten some of you about some theorycraft and share a few thoughts.

Energy:
A bow is most energy efficient when its draw cycle is equal to release cycle (that's 100% energy). Most of today bows can reach somewhat impressive effectiveness (above 80%) but none is close to the ideal line.
Now to the energy transfer... Well, some bows get most energy transferred to heavier arrows while other bows do better with the lighter ones. Dependencies? Loads of things. Will post as many as I know if you wish me to. Basically, a bow with a given setup will shoot light arrows faster then the heavy ones.

Penetration:
Penetration will depend on the energy transfer rate. For an arrow which is to light, the energy can't transfer efficiently enough to secure the theoretical penetration. Also, the risk of breaking arrows/limbs grows proporcionally. Is that any different with heavier arrows? No. An arrow with to much weight may take all release cycle energy from the bow but the effectiveness of the bow will be smaller.

So what do we get from this? And I really counted only this 2 factors.
At 70# the bow could shoot the 300 gr arrow with more penetration power as the 422 gr arrow.
The normalized weight should be 350 gr for the arrow (or more). As the 300 gr arrow is only 50 gr lighter then the normalized arrow I do believe, the arrow had more penetration as the 422 gr arrow had. The 77 grains difference for the heavy arrow could have justified it.

Final words:
I do believe there are many other factors which "interfered" with this theory yet a complete graph of the actual behavior would be the best proof. What do I mean? Get arrows. As many as possible with weights from 300 gr to 450 gr in let me say 10 gr steps with at least same spine and utility (vanes, broadhead, knocks). The diameter would be hard to maintain however. Now shoot them all 5-10 times, get the values, make the graph. You will discover a simple thing. There is no winner on any of the graphs ends.

HOUGNH