I certainly defer to you, and even more-so to practical sailor's research... real world testing trumps any wacky theories I come up with, and I completely understand that. There are two simple reasons it surprises me (though, again, I admit I'm clearly misunderstanding something since the real-world test has proven me wrong). Firstly, I (incorrectly) feel like the only link of the chain that is exposed to the "blade" part of those hooks is the one that can pass freely through the hook, and is not obstructed by it in any way... which in my brain puts the chain-vs-hook load on the "next" link of the chain, which can't touch the "blade" based on it's being turned 90 degrees and pressing on the side of the hook. I'm clearly wrong though. Additionally I'm surprised that it would create any more point loading than the link-to-link strain of every link in the chain pulling on the next link. In my head, the contact area of a chain link touching another chain link is pretty close to being a single point. and the contact area of a link touching the hook is at the very least two points, and possibly bigger. I'm not arguing that I'm right and the test is wrong... just the opposite... I'm clearly wrong, and working through how to fix my faulty logic. Anyhow... good information to know... and largely irrelevant to me anyway since I use a short length of chain on the end of a nylon rode anyway, and don't need a hook for anything under strain at all... so I'm good.