Koinos mainnet is well suited to the web. The three second block time is great for just about anything you would build. But some games want more.
Game developers like the idea of sub-second block times. They want to know your chain can support their massive transaction volume when their game inevitably blows up.
Don't get me wrong, games can build on slow networks just fine. It just takes a little more work. You need some other infrastructure to hold the transactions before they're submitted to the chain. The real limiting factor is whether you can access enough resources to support your transaction volume.
Fee-based chains make the economics too unpredictable. You can only safely support players who are paying their own gas fees. This naturally pushes developers to launch their own chains where they will always have enough gas tokens.
Some chains treat this as their chosen scaling solutions, and it's not a bad strategy. Avalanche, for example, is designed around subnets. It's relatively simple to launch your own subnet with whatever block time, block size, tokenomics, and level of decentralization you want. Subnet validators contribute back to securing mainnet and it's simple to bridge assets between your subnet and mainnet.
Because of Koinos's upgradeable smart contracts, there's a unique opportunity to adopt a similar strategy that can be adopted by dApps only once it's needed.
DApps could build directly on mainnet for simplicity, then if they outgrow their transaction capacity, they could launch a koinos-based subnet and upgrade their existing mainnet contracts for backwards compatibility.
This would make it easier to get started without worrying about scaling upfront. Win for dApp developers. Win for koinos.
P.S. strategies that offload scaling to other networks only get you so far. Mainnet scaling efforts are still important.