OK.
I don't dispute that Thomas had 32 1/2in hands. Just interesting how disparate the data is.
Oh, no question that it's not a quality-controlled environment. You're absolutely correct that it's all somewhat unreliable data. I'm sure the people measuring think they're doing a fine job, but other than a centralized standard for a couple hundred players (the Combine), the rest of the measuring and timing is performed by dozens and dozens of scouts who all probably learned a different set of rules for how they measure.
What I find more interesting is looking at general correlation trends between measurements and performance; granted, individual measurements may not be accurate, but the aggregate of data points can give a relatively reliable basis for running regression analysis. Turns out there's not a lot of support for the idea that longer limbs = better players. Yes, long arms can give more margin for error in creating space and landing an effective punch, but timing, hand placement, technique, quickness, core strength, etc, are always more important for success.
Anecdotally, I've done a fair bit of hand to hand sparring and also swordfighting (from traditional foil fencing to German broadsword in the HEMA tradition), and while reach can help at times, it's not always the great equalizer that people assume. I'm about 5'8, and my main sparring partners are 6'5 and 6'2, so I'm always at a reach disadvantage, but whether armed or unarmed, I can make up for the difference with footwork, core power, body positioning, efficiency and quickness of movement, craftiness, etc. I don't lose very much.