Two-player is different

Almost everything written about this game assumes three or four players. You two play head-to-head, and several deep assumptions change. This page is the translation layer — and it matters enough that it’s worth reading twice.

Every swing counts double

In a four-player game, hurting one opponent helps two bystanders. In two-player, every point you gain is a point he doesn’t have, and every resource he loses is pure profit for you. Attack cards, tile denials, and award steals are all roughly twice as strong as the four-player advice suggests. There are no bystanders; it’s all margin.

Attack cards are always aimed at you

Plant-removal cards (the asteroid family and friends) have exactly one target tonight. Two practical consequences:

• Never sit on 7 plants. Bank plants below the visible threshold or convert the moment you reach 8 — a fat plant pile is an invitation. If he’s holding events and you’re at 6+ plants near generation end, place the greenery early even if the spot isn’t perfect.
• Your own removal cards rise a tier. Advice written for four players calls them situational; head-to-head, eating 5 of his plants is close to a 5-plant gift to yourself.

Milestones and awards are half-price fights

Three milestones, claimed for 8 MC each, 5 VP apiece — and only two of you racing. Head-to-head, milestone racing is one of the highest-return habits in the game:

Track his milestone progress out loud. If Matthew is one city from Mayor or a couple of building tags from Builder, either beat him there or make peace with the 5 points and take your compensation elsewhere. Being surprised by a claim is the only unforgivable version.

Awards flip the other way: in the official two-player rule, second place is not awarded — lose an award and you score nothing, even if you paid to fund it. So a misjudged fund is the full fee burned plus 5 points handed to him. Fund late, fund only what you’ve already won, and treat funding an award he’s winning as 5 points of charity with a cover charge.

The board is emptier — placement is richer

Fewer players means more open hexes for longer. City-plus-greenery clusters have room to breathe, the best placement-bonus hexes stay available deeper into the game, and tile denial is a real move rather than a luxury. The arithmetic that makes denial rational: head-to-head, a point denied equals a point gained. Taking a hex that would have been worth 5 VP to him is a 10-point swing — which prices a 25 MC “inefficient” standard-project city at 2.5 MC per effective point. If you build cities, commit to greening around them; an unclustered city is mostly wasted.

The game runs longer — engines get better

Two players push the global tracks slower: expect 11-14 generations against a four-player game’s 7-10, and roughly fifty research-phase cards seen across the game. That quietly buffs card draw, discounts, blue action engines, and accumulating cards — more activations, more turns to compound. The strong-player consensus goes further: at two players, the long game is the default winning game, and several high-level players say most 2p games are decided by who builds the better discount-and-draw stack and cashes it in the final generations. A rush can still beat it — but only a disguised one that ends the game before the engine player expects (see Tempo & the clock); slow, visible terraforming just feeds the engine more generations.

No drafting (unless you house-rule it)

The base two-player game deals research cards straight — no draft passing. You can’t hate-draft his combo pieces away; you can only watch what he plays and respond on the board. (If you two ever add the drafting variant, card denial becomes a real skill — but that’s not tonight’s game.)