Thanks for the summary. I've always found it interesting to look back at the assumptions and mechanisms used in historical rules-sets. (Although it's a bit alarming to realize that games of my youth are now becoming 'historical'. ) It's a good reminder that we're probably making questionable assumptions now too, but we just don't realize what they are yet.
Looking further back than Harpoon, there's Fletcher Pratt's naval war game from the 1930s and 1940s. Played with miniatures, it initially focused on gunnery and torpedoes, and then added aircraft in more detail (describing them as "a headache all the way"), although it never got into carrier operations. Ideally it was a refereed team sport, with each player getting a few ships, and eventually they had large groups playing in a rented hall. There were no dice in the final version. The combat system relied heavily on player judgment. Each player had to plot his (or her - they had female players) gunfire, manually estimate the range, and specify the spacing of the salvo, writing it on an arrow tacked to the floor. The referees then did the measuring and assessed hits. Ships were assigned a single numerical value (essentially hit points) using a formula based on a number of factors (armour, weaponry, displacement), and as they were damaged they lost a proportionate amount of performance. If you lost 20% of your ship value, you lost 20% of your speed, 20% of your guns, 20% of your torpedoes, etc. This may seem rather abstract to us, accustomed to detailed damage reports, but apparently it gave results that were reasonably consistent with reality, when you summed it up over an entire battle.
Then we could step back even further, to Fred T. Jane's naval wargame. (Yes, that Jane.) First published in the late 1890s, it was the reason for those detailed diagrams of ship armour layout and gun classifications in the early editions of Jane's Fighting Ships. This game didn't use dice either. Instead, armour diagrams were drawn out, and then the players were given a sort of paddle with a pin attached somewhere off-center on one side. The player would strike at the diagram with the paddle, and the pattern of pin-holes in the paper would show where the individual shells had struck. Because the paddle hid much or all of the diagram, and the player wasn't sure exactly where the pin was, there was a lot dispersion in the shots, until trial and error let the player gradually home in the target. There's an interesting paper from 1899 here in which Fred Jane describes his game to members of the navy, and discusses his conclusions from it. Some of these are quite at odds with actual experience ("Line abreast is the best formation in which to go into action. Line ahead means the destruction of a fleet in detail.").
What bad conclusions are we drawing today?