I think an approximation of such an algorithm could be made. It would be decent in a way that it could tilt enough games in someone’s favor to be noticeable. But, then we:
a) would have proofs of that algorithm, and
b) knowing it exists could play around it, effectively neutralizing it.
And even then, one of the conditions to make even such an approximation, would be to never nerf outline cards because the algo would depend on drawing your highest impact cards on the turn where their impact is known to be (on average) highest. So if you nerf the outliner, you lose your only way of creating such an algo.
Keep in mind additional complications, such as:
- you dont want the player to HAVE his highest impact card in hand before it’s time to play it, to minimize the chances of them playing it at the wrong time. This would neutralize the algorithm
- you dont want the player to lose the game before that time, so you’d have to be executing a Zephrys algorithm before any draw is made
- and that fact alone guarantees you we would know about it. It’s impossible not to realize that your opponents are drawing perfect cards all the time.
Nature of games would change. If you thought THIS meta is repetitive, think about the meta where algo serves you optimal cards all the time. Every game would play out the same and some cards would never see play at all.
All in all, it is practically impossible.