Ask HN: What game do you wish existed?

In other words, two AI "commanders" (perhaps along with separate AI "squad leaders") duke out the battle with the players being simply pawns. I've always imagined it in a WWII setting, but I suppose really any setting could work.

Imagine -- you spawn in, and your commander tasks your squad with taking control of a church in a nearby village. You and your fellow squad members enter the village and come under machine gun fire. A teammate tags the window where they saw the fire coming from. You all take cover, as the squad leader starts barking commands -- a few squad members stay back and start providing suppressing fire, while you and some others start advancing slowly, hopping from cover to cover. A teammate is hit, and another is tasked with dragging him to safety. You and the few others still advancing finally get to be in range to toss a grenade in that window. You peek from cover, grenade in hand... to see a tank rounding the corner down the road. You tag the tank, and chaos ensues as the squad leader screams at you to retreat. A mortar squad in a nearby section of the map is alerted to the tank and starts dropping mortars as you fall back to your original position, explosions and bullets flying all around you.

Your squad regroups, a few members down, but the strategy is adjusted and you go back in, eventually destroying the tank, capturing the church, and getting control of the village.

At the same time of this battle, similar battles are potentially unfolding in other parts of a larger map, as the attacking and defending commanders dynamically wage war.

Sometimes I want to think strategy in games, but sometimes I just want to shoot stuff. I think if executed properly this could strike a good balance of both reaching meaningful objectives and also focusing on dynamic, moment-to-moment action. Games like Squad can be great -- if you can find a good squad leader or group to play with consistently. As I get older I find I don't really have time for that, and the probability of getting matched with a good squad leader by chance is pretty low. The setup of this game minimizes the risks of poor teamplay and makes the "Squad" sort of experience more accessible.

You play against your two selves, and flip back and forth between two modes. In "mayor" mode you have the birds-eye view and you are trying to build a functioning and safe city for your citizens.

Then at any time you can flip to "street" view, where you are no longer mayor, but a criminal leader. Here you are trying to expand your criminal network and evade police.

If the "mayor" your reduces crime to near 0 then doing anything as the "mobster" you will be extremely difficult. If the "mobster" you is super successful then the mayor has a very poor rating, and citizens complain of crime. This should also have physical change to the look of the city as well (graffiti, car's on blocks on the streets, cars with smashed windows, stores with smashed windows, or boarded up, parks generally trashed).

Personally, I like to work teaser math problems and algorithms, like the Traveling Salesman Problem, set sorting problems, or whatever. It's so much more fun to know you might by happenstance, fumbeling arounds in math space, find something actually beneficial to the world. You'll never contribute to society play a "bounded game" like candy crush or whatever.

I guess what I'm saying is that I wish there were such a thing as an "unbounded game" that truly allowes you to discover. I think protein folding crowd sourcing comes close, but how fun is that, really? (I'm literally asking, I don't know, I've never partaken)

How do you make a game that also contributes to collective knowledge?