Short Term vs Long Term

One of the more interesting types of decision you have to make in Prismata is whether to invest in the long term or the short term. Compare Tarsier and Grimbotch:

Tarsier is a long term investment, since it takes a while to build and sticks around for the rest of the game. Grimbotch is a short term investment, giving a lot of value for its cheap cost over the course of four turns, and then dying. Both are good in different situations.

Why should you ever buy Grimbotch?

Grimbotch is a random set unit, while Tarsier is a base set unit. Random set units are typically balanced to be slightly stronger than base set units, so Grimbotch is supposed to be slightly stronger than Tarsier. But is it? Although Tarsier takes an extra turn to build, it can keep attacking for the rest of the game. A typical Tarsier might do 7 damage before the game ends. Meanwhile, Grimbotch has lifespan 4. This means that you can attack with it a maximum of 4 times. If you hold it back on defense on its last turn of lifespan – the most common way to play it – then you attack three times for one damage each, and then hold it to gain 2 soak, for a total of five value to Tarsier’s 7 (the 7 is a randomly chosen number – sometimes it’s lower, but sometimes it’s much higher, depending on how long the game goes on and how early the Tarsier was made).

Holding Lifespan units back on defense on their last turn of life is a common tactic

So why is Grimbotch stronger? Grimbotch is stronger because it’s faster. It hits right away, while Tarsier takes an extra turn, and this puts the opponent on the back foot. Say that you make a Grimbotch, and your opponent makes a Tarsier. The whole time your Grimbotch is up, they’re down one health – they’ve had to make an extra point of defense to defend the Grimbotch. Then on the turn your Grimbotch dies, you get to hold it for +2 health – while their Tarsier only does 1 damage – for another +1 health advantage for you, for a total of +2. After this, your Grimbotch dies and you’re down an attacker, but the short term health advantage you gained is more than enough to compensate for losing the attacker. This is because Prismata is a game where you are constantly making investments – either in Drones that make Gold, or in attackers that damage your opponent – and so getting temporarily ahead on health is really valuable, because it frees up resources that you can invest in attackers. So the short term advantage ends up turning into a long term advantage.


Even though Grimbotch is the stronger unit overall, it won’t always be the right choice. Here are three rules of thumb you can use to decide what to go for:

1. Pre-absorb, long term units are better

Term 1.2
Tarsier, Tarsier

If you aren’t over the absorb barrier yet – in other words, if you aren’t yet doing more damage than your opponent can absorb – then you want to stay away from short term units, and focus on long term units instead. The reasoning for this is pretty simple: the advantage of short term units is in getting temporarily ahead on attack. Getting temporarily ahead on attack is worth much less when that attack is getting absorbed.

This is also why longer build-time units, like Tarsier, are excellent choices for your first attackers: they are cheaper at the cost of taking longer before doing damage.


When you’re in a high pressure situation where fast damage is really important, this is a bad trade. But early on in the game, when any immediate attack you make will just get absorbed, giving up early damage in order to get cheaper attackers is a fantastic trade. Even burst units, like Endotherm Kit, can be a good choice to build pre-absorb, when they have a high build time.

2. When block is efficient, long term units are better


This one isn’t intuitive at first, but it should make more sense as you gain experience with the game. If it still doesn’t make sense, then you can just use rule 3 as a substitute.

Exactly how much one point of defense costs will vary depending on what type of defense players are using. For example, if they are defending with Walls, then three points of defense costs 5B. A Blue is worth approximately 1.5 Gold, so three points of defense costs 6.5 Gold, meaning that one point of Prompt defense costs around 2.2 Gold. Suppose instead that they are defending with Rhinos. A Rhino costs 5R – a Red is worth approximately 1 Gold, so if players are defending with Rhinos, one point of Prompt defense costs 3 Gold – a lot more!

Remember, the key to short term units is to either force your opponent off of making good investments by doing damage fast, or to free up your own resources to make good investments by helping out on defense.


So the more expensive it is to defend, the better short term units are. Their value comes from getting a short-term lead and snowballing that into a concrete, long term lead. This is a lot more effective when your opponent is being forced to make inefficient defense like Forcefields or Rhinos (or, for a defensive unit like Sentinel, if the defense it makes is saving you from having to make Rhinos and Forcefields). If players are defending with efficient soak like Polywall or Aegis then you aren’t getting enough of a short term lead to compensate for your loss of long term value.

People will tend to make their most efficient defense first, and only make less efficient, ’emergency’ defense like Rhinos or holding Drones, if they are facing a lot of damage. This leads to the next rule of thumb:

3. When pressure is higher, short term units are better

The more pressure you put on your opponent, the more valuable further pressure is. So combining Tia Thurnax with Gaussite Symbiote explosions allowed Arkanishu to take a decisive lead and win the game.

When both sides are having a harder time defending, they are forced to make less efficient defense. This means that short term units are more valuable, because those short term units are forcing out even more inefficient defense. Because of this, combining burst damage can be very strong, provided you pick the right moment to do it.

4. Long term units are good early and short term units are good late

Pressure naturally rises as the game progresses. Every time a player adds a permanent attacker, their opponent will have to make one more defense per turn to avoid a breach for the rest of the game. Because of this, it’s natural in a game of Prismata for players to keep having to make more defense each turn as the game progresses. Pressure rising over time naturally means that short term units tend to be good late game, while long term units are good early game.

Another way of thinking of it: if a game lasts until turn 20, then a Tarsier built early on, on turn 5, will hit a lot (14 times) before the game ends. On the other hand, a Tarsier built on turn 16 will only hit 3 times before the game ends, so you might as well make a Grimbotch which will get more value.

5. Advanced units are still better than base set units

When I first put this article up, a lot of good players pointed out that when choosing between a short term, advanced set attacker like Kinetic Driver, and a long term base set attacker like Gauss Cannon, the advanced unit is nearly always the better choice.

This is especially true for the “short term” units that last 4 or more turns – those units would be more accurate to call “medium term” and tend to be pretty good at any stage of the game, although they still aren’t ideal pre-absorb.


Killing an opponent by combining a bunch of burst damage is really fun, and so is slowly outvaluing someone over a long game with a bunch of long term attackers. And sometimes you get to do both – gain a long-term edge early on, and then finish them with short term attackers like Grenade Mech or Tia Thurnax. This post should get you started on learning how and when to play for the long term and short term.

3 thoughts on “Short Term vs Long Term

Add yours

  1. great blog! i love this kind of analysis.

    i have <1 hour on prismata, and i'm learning by losing to the expert AI. i just played a game where the AI opened (going second) 1. DD 2. DA and then rushed out 4 grimbotches, 2 per turn. there was no good absorber in the random set, and i forget my opener, but i quickly got two tarsiers, remembering this blog entry.

    the rest was a blur, but i lost, quickly falling behind in the arms race, and i ended up running out of walls. if grimbotches truly are a weaker early investment than tarsiers, what is the optimal play to punish 4x grimbotch openings? am i over-relying on walls for blocking, when i should uae more engineers by using 2 engineers + 1 wall to block the 4 grimbotches every turn?


    1. Grimbotches are overall stronger than Tarsiers, which means they can be good to get even in situations where Tarsiers ‘should’ be better. If you had a Blastforge up, then leading off with two Grimbotches would be a mistake by the bot, because you can just Wall them off. However, if you had gone for Natural Animus, you wouldn’t have a Wall up, meaning you wouldn’t be able to fully absorb the two Grimbotches. In this case leading off with two Grimbotches would make a lot more sense for the bot, since defending them is a lot harder for you.

      Sets with no absorb in the random set tend to be aggressive, and Grimbotch is a good attacker, which often makes them even more aggressive – where ‘aggressive’ means that it’s correct to switch to making attack fast. Like you found out, falling behind in the arms race can be a big problem in aggressive sets. In base plus Grimbotch – aka, playing with just the base set and Grimbotch – it’s actually correct for player one to go 1. DD 2. A, in order to get attack going faster! 4x grimbotch will usually be a mistake to open with (if you have a Blastforge up), but not a huge one, so it’s mostly just a matter of playing solidly – getting your absorber up asap, not overteching, and other generally good plays – and you should win. But, if it’s a highly aggressive set, then the mistake of going too slowly could outweigh the mistake of going for too many grimbotches too early.

      Liked by 1 person

  2. thanks for the reply! your concise, step-by-step writing style where you assume little undue foreknowledge (for example, someone reading a blog post about set reading shouldn’t expect you to explain why absorb is good) from the reader, and clearly communicate ideas without leaving many unexplained leaps in logic that require the reader to read your mind, is why i’ve read each of your blog posts several times.

    i’ve learned a lot in the past ~10 hours of gameplay (mostly against bots in casual play for daily rewards) since i posted that comment, but still have a long way to go. the three major causes of improvement for me have been:

    1. using an intermediate absorber. i was so sold on the power of having a huge absorb barrier that i would shun inferior absorb, to the point that i was giving up my engineers (but never drones or attackers) and absorbed nothing while i rushed out that mega absorber that was probably overkill for that point in the game anyway

    2. realizing just how much of an investment in economy it is to get a third engineer. i got the impression that buying a third engineer early on (before the point where you’d make it for granularity on defense) was just something you did eventualy, but i’ve since seen the power of playing 2-engineer games, especially since i’m mostly playing base+5 for now, and there are fewer random units to empower powerful economy with diverse tech

    3. linked to the first two points, and this is probably a lot more relevant against AI, but i adjust my earlygame based on whether they get a third engineer or fast tech now. if they tech fast i’ll probably make a blastforge asap so i’m ready to make that intermediate absorber wall, or if they make third engineer when i think the set’s aggressive, i’ll stay on two engineers and exploit that casual mode AI’s simplistic programmed playstyle

    my current focus of improvement is improving my gameplans, which too often involve me overteching (like trying to make apollo and then add on shadowfang as consistent damage) or just not playing certain units correctly. in 3 feral warden sets, i thought more feral wardens was better, and that since animus provides 2R, i should go double conduit + animus to make two every turn. it’s obvious to me why that doesn’t work now, and made me lose all those games; the strength of feral warden is in absorbing on a new (prompt) ONE every turn. by massing them, i would always end up having to either use a fresh 3 HP feral warden to soak, or overtech on 2 engineers by having animus, 2 conduits, and a blastforge (usually at least 2) to make walls just to soak with


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at

Up ↑

%d bloggers like this: