There are mantis shrimp, there are an absolutely staggering array of colours and, apparently, there is a fairly active underwater economy based entirely on taking things from your neighbours.
That, broadly, is Mantis.
On your turn you make one choice: either attempt to Score from your own collection (your Tank) or attempt to Steal from someone else's. You then reveal the top card and find out whether probability is on your side, is against you or has simply decided that Steve across the table is due for another card.
After having spent a bit more time with Mantis the cleverness of it really comes to the fore. What at first appears to be a very simple matching game has a rather slick layer of informed risk to it. The real question for the second look isn't whether Mantis is easy and fun (it most certainly is). The question is whether that little decision stays interesting once the initial novelty has gone.
Two Rules, More Decisions Than Expected
The central mechanism is absolutely superb in its economy.
Each Mantis card has a single coloured character on the front. The back of the card has three different colours and the colour on the front will always be one of those three.
That means the card you're about to reveal isn't entirely unknown.
Before you even get to the top card you have to decide whether you're going to be trying to Score or trying to Steal. If you're going to Score you simply reveal the card into your Tank. If you match a colour in your Tank then all of the cards of that colour (including the one you've just revealed) go into your Score pile. If you don't match then the card just stays in your Tank.
Stealing is the nasty alternative. You reveal a card into an opponent's Tank and match one of their colours and that entire colour group (including the card you just revealed) goes to your Tank. If you don't match you've just given your opponent a free card. The important thing to note is that the card you've stolen isn't scored immediately.
That final detail is what gives Mantis its bite.
A successful theft will of course make your Tank worth a fortune, but until those cards get into your Score Pile they're still just sitting on the table with a very large 'PLEASE ROB ME' sign on them.
The Card Backs Are the Real Game
With a bit more exposure to it I think what I like most about Mantis is...
You're gambling, you're just not gambling blind.
Suppose the top card is green, yellow and purple. Your tank has green and yellow, someone else's has purple.
Scoring provides two obvious routes to success. Targeting the other side is one of them.
Easy decision?
Not necessarily.
Maybe that purple stack has four cards in it whilst your colours only have one each. Suddenly the prospect of stealing looks very attractive indeed. Alternatively there may be another player closing in on the finish and you're going to need to get your own cards into points before you start thinking about anyone else's aquarium.
It's a very neat decision in fact as almost everything you need is right in front of you.
Even better really, attempts to steal have their own penalty. If you guess wrong the player you've targeted simply keeps the card you revealed. You haven't just wasted your action, you've potentially improved the position of someone else.
Nothing creates family-game diplomacy quite like:
"You gave him that!"
"I didn't know it would be orange."
Why Stealing Doesn't Get Old
Take that mechanisms can go rather badly with age.
The first game is all laughs. The fifth game is all vendettas. Eventually Aunt Carol is going after you for something that happened three Christmases ago.
Mantis avoids some of that problem as its aggression is linked to the visible information.
You're invariably going after someone because the three colours on the card make them the obvious choice. The robbery then doesn't seem quite so random. You're not necessarily picking on Dave. It's the statistics that are picking on Dave.
Obviously that won't stop players from making emotionally driven decisions. We're still just humans sitting around cardboard so grudges are effectively part of the game.
But the system incentivises opportunism rather than pure destruction.
It also has a nice flow to it. Large tank stacks are of course useful assets as well as very attractive targets. You want sufficient matching to enable a big scoring turn, but the more you accumulate the more interest your personal pile will generate in the other players.
Replayability Comes From People
Here's the big caveat after a number of repeat plays: Mantis doesn't expose some huge hidden strategic terrain.
The official description of the game is that it's a classic matching game with more thought-provoking set building decisions. BoardGameGeek has it categorised as a push your luck, set collection, hand management and take that game. [boardgamegeek.com]
But it's still very much a light game.
There are no evolving player powers, no complex combinations and no radically different strategic openings for game 12. You're just going to be looking at three colours again and deciding whether the best risk is to point inwards or at someone else.
So why continue playing?
Because the players keep changing the equation.
One person has a nice pile to hoard. Someone is ahead in the scoring. There's a point at which another player is suddenly close enough to the win that a steal which would normally be a sub-optimal play becomes very attractive. The rules underneath are pretty much static while the state of the table is constantly in flux.
That's perfectly in line with a game the publisher reckons will take about 10 minutes. Mantis isn't designed to be a game to fill an evening. It's the game that somehow becomes three games because "one more" doesn't sound like a huge amount of time at all.
The Race Matters More Than It Initially Appears
In the normal game the first player to get 10 or more cards in their score pile wins. The cards that have been scored remain face down, though you're required to tell people how many you've scored when you're asked. That's a major pressure release.
If Mantis was all about building the biggest collection possible then everyone could just go on and on with ever more ludicrous Tanks. The game isn't like that though. If someone got to seven, eight or nine cards they were going to change the feel of the game.
Suddenly efficiency matters.
The only way to get cards into your Score Pile is to be trying to score so there's a very real distinction between amassing wealth and actually converting that wealth into victory. That's probably Mantis's most astute design decision.
Stealing is useful in that you get the cards, but it doesn't actually get you any closer to winning. In the end you're going to have to subject those delightful stolen goods to another probability check.
Greed has paperwork.
Two Players Is a Different Beast
Mantis makes explicit rule changes to two players.
Rather than be racing to 10 scored cards you're after 15. More importantly, each successful steal gives you another turn. Keep successfully matching during the steals and you can go on to take further turns.
That produces a more combative game.
With a number of players the attention is going to move around the table and the potential victims are multiplied. At two there's no where to hide. Every attempted theft is going to be directed at the one increasingly suspicious individual.
I'd prefer Mantis when you've got multiple personalities going at one another, the official two player tweaks just stop it from feeling like a half-baked multiplayer game.
BoardGameGeek currently lists it as a 2-6 player game and the community poll is in favour of 4 or 5. That seems to be about where the social chaos is most evident. Again though that player number is just the community poll result rather than a rule.
A Few Bumps in the Rainbow
The very simplicity that makes Mantis such an easy recommendation is also its limitation.
Repeated plays won't somehow turn it into a strategic heavyweight. Good decisions will improve your position, but it's the revealed card that ultimately decides things. There will be instances where the sensible play fails. There will be instances where an opponent makes a hopeful play and gets precisely what they need.
That's not a design flaw. That's the source of the game's drama.
It means that the players who in fact hate luck are not going to come around to liking it after 20 games. If your reaction to a loss is to go off and check your spreadsheet to see exactly where your economic engine went 3.7% inefficient, Mantis is probably not going to be the game for you.
There can also be occasional targeting. The rules make attacking a profitable rather than just a spiteful thing to do, but a clearly lucrative Tank is going to be targeted. Anyone who doesn't like having their progress just snatched away by another player should be very aware of what they're getting themselves into.
The Second-Review Verdict
More plays have failed to reveal any hidden complexity in Mantis.
They've done something even better actually, they've shown you that it doesn't really need anything at all.
Its best trick is still the relationship between the information on the back of the card and the unknown colour beneath. That one idea is what turns what could be nothing more than blind card flipping into a series of tiny probabilistic decisions.
Then Mantis attaches those decisions to theft.
Lovely.
It's fast, immediately comprehensible, interactive and just tactical enough to make you feel personally accountable right up until the card comes along and proves you absolutely weren't.
Who should play Mantis?
Families and mixed experience groups are the obvious market. The publisher has it at 7+ with 2-6 players and a playtime of around 10 minutes. It's also a very good game for people who just want something quick to play between the bigger games.
Fans of push your luck and take that games will find plenty to enjoy (particularly if you're the type of group that enjoys the dramatic reveals and the 'I knew it!'s that follow events which in fact no-one actually knew about). Push your luck and take that are both listed as game mechanics on BoardGameGeek.
Who could possibly not enjoy it?
Players after all looking for a deep long term strategy might find the decision space quite narrow. And if direct theft actually makes your group miserable rather than just theatrically outraged then the entire 'Steal or Score' steal half is a fairly significant red flag.
Final Thoughts
Mantis survives the second review pretty well.
There's actually less randomness in it than the rather garish array of shrimp at first sight would lead you to believe and considerably less strategy than its most flamboyant comparisons would suggest. It's a success which sits comfortably between those two extremes.
Every turn gives you information. Every turn gives you a decision. Then every turn gives the deck a chance to make you look absolutely ridiculous.
After multiple plays that's still quite enough.
Mantis isn't the game you'd bring out for a proper night of serious warfare. It's the game you'd bring out when people are all arriving, after you've done something more substantial, or when someone says the awful 'we've only got ten minutes' thing.
Because ten minutes later someone will have ten cards.
And everyone else will be demanding a rematch.
