In 2017 I started learning about Ethereum. Once I understood it as a virtual machine, the idea hit me that one could build a game. About that time, CryptoKitties hit and became an example of what could be built. Shortly late, a tutorial called CryptoZombies helped teach "a way" to build crypto collectable style games on distributed virtual machines like Ethereum.
Like many people, I went through the tutorial and I started working on my own game. A big difference between myself and many others is that many others only slightly deviated from the CryptoKitties formula. The inspiration for my game, Orbiter 8, came from an old BBS game, so I deviated a lot.
I worked on Orbiter 8 for a couple of years and released two demos before I bothered with ERC20/ERC721 interfaces. The game concept has an in-game currency and ownable assets so I created interfaces making those things tradable outside of the game. This was sort of a no-brainer. By using ERC721, I can use open markets like OpenSea to trade my game assets as NFTs. By using ERC20 I can put the in-game currency on SushiSwap and let the public trade the currency as they wish.
I didn't intend to, but having done those things now means I am building a Play To Earn game. That is, of course, assuming my game assets and in-game currency actually have value and a value higher than the cost of creation. But for all intents and purposes, the emerging model is a pretty decent Play2Earn model.
Most games are about progression. You typically "grind" in any game to acquire better items and stats. The only step to make any of these games P2E is to allow those things to be traded on open markets. So if you build a game natively on one of these networks, P2E is an extremely natural emergence.