[jeffreyrogers]

I think it is the other way around. Cavalry could beat pretty much any group of non-cavalry unless they had a massive numerical advantage, so castles/walled settlements were the way to defend against that, but they only arose by necessity.

Carol Quigley's theory is that technology changes the balance of power, sometimes favoring centralized power other times favoring decentralized power. Cavalry/knights/castles favored centralized power because equipping knights with armor and feeding their horses required a lot of peasants. The invention of cheap firearms led to the masses having more power (and around this time democracy began to spread). Then tanks, ICBMs, aircraft, submarines, etc. centralized power again, where we remain today, although that may be changing with cheapish drones and semi-successful insurgency tactics.

[Balero]

Cavalry could absolutely not beat 'any group of non-cavalry'. Cavalry was a very specilised, powerful and effective tool. But the real world wasn't what the English guy said in Braveheart ("We have the heavy cavalry, we win"). Many knights fought on foot as well.

Cavalry could not break many infantry formations. You were safe in a square.

Cavalry was great at harassing supply lines in raids, and importantly running down fleeing infantry.

This is all purely a European perspective, steepe cavalry was very different, but many of their benefits were at a larger scale than a single battle.

Generally the whole thing is too complicated to draw some simple conclusion from. War is, and always has been complicated and messey, not a game of chess.

[jeffreyrogers]

Cavalry dominated the middle ages except in mountainous areas until artillery became prominent. Yes, cavalry could be defeated, but all things being equal it was far superior.