Secrets (2017)

Game Details
NameSecrets (2017)
Accessibility ReportMeeple Like Us
ComplexityLight [1.49]
BGG Rank2875 [6.44]
Player Count4-8
Designer(s)Bruno Faidutti and Eric M. Lang
Buy it!Amazon Link

TL;DR: Definitely not for me. Maybe not for you.

Secrets makes a good first impression, on several levels. For one thing, the box art is very striking – bright, bold colours with an effective, cartoonish graphic style that evokes some of the light-hearted Cold War paranoia of 80s television shows. For another, it’s a game by Bruno Faidutti and Eric Lang and that is a dream team of designers. When you open the box you find the game components are as straightforward as a deck of card and some wonderfully chunky identity tokens that make the same seductive music as a pile of poker chips when rattled together. It’s a beautifully presented package.

Secrets box

Its name is a good reflection of its design too – it’s a social deduction game where the currency of play is hidden information. It’s also a very tractable game, made up of simple rules that give the concept room to breathe. At the start of the game you’re dealt out a role – CIA, KGB or hippy. These represent team loyalties – CIA and KGB members win together on the basis of the combined score of each member. Any hippies in the game are in it by themselves – their job is to have the lowest score of the table. Points are awarded by control of intelligence assets – journalists, scientists, assassins and so on.

Role tokens showing CIA, KGB and hippy allegiences

Loyalties are, as is typical in games like this, obscured from the start. You’ll know who you are. Depending on the number of players you might also know the identity of one of your neighbours, and perhaps even the unassigned identity token that sits in the middle of the table. Aside from this, you don’t know who is who – all you know is that identities are forever shifting and you need to ensure you’re paying attention to the implications of everything single card that is played.

Every round the active player draws intelligence assets from the draw deck until they have two different types in hand. They secretly choose one of these as an ‘offer’ and select a player to which that card will be slipped. ‘I’m giving you an assassin, use her wisely’, the active player might say. The target of the card can accept, in which case it gets added to their tableau. If they reject, it gets added to the active player’s tableau instead. Any effects on the card are actioned, and play moves on to the next player. If two cards of the same type are in a tableau at the same time they get flipped and no longer contribute to scoring. The game ends when the fourth, or fifth (in smaller games), card is added to a tableau.

Roles in Secrets

It’s in the interplay of card effects that we see some of the game’s most interesting behaviours. The scientist permits its controller to swap the identities of two other players, and they don’t get to see what they ended up with. As a double agent you can swap your current role with the unassigned role in the centre, and then check who you are. The politician doesn’t get added to your tableau – it gets added to someone else’s and you get to look at their token.

These roles are useful and effective. You can find out loyalties and then selectively share (mis)information about your findings. You can get rid of dead weight on a team with an unexpected betrayal, or you can continually throw disadvantage towards someone you believe is an opponent. Or indeed, throw advantage their way in the belief they’ll attempt to redirect it back to its source. There are eight different roles and they’re the hammers and chisels that you’ll need to construct a case solid enough upon which to rest your loyalties. You want to help your allies, hinder your opponents, and must do so under circumstances of intense uncertainty until you can extract the necessary context to make meaningful and important decisions.

I mean, let’s say someone offers you a politician and the token glance it will permit. It’s a lens on the game – a precious reliable navigation point in a sea of uncertainty. It’s also worth three points. It’s a good offer, an offer you’d only give to an ally, assuming they don’t already have a politician stinking up their display. It’s definitely worth accepting…

… if you believe that they are an ally. Or at the very least, they believe that you are an ally. Or if they believe that you believe that they believe they’re an ally. It’s the Vizzini staircase once again – that elaborate mechanism of bluff and counterbluff that underpins many games of this type. And, as is usually the case, it doesn’t really make for a satisfying gameplay mechanism.

Essentially a Vizzini staircase is built upon a grounding of information – a fixed starting point that has no obvious upper bounds to its escalation. It’s like the set of positive numbers – there’s no place you can stop counting and say you’re done because there’s always one more number after that one. It begins at 1, and doesn’t end. On a Vizzini staircase, there’s a similar situation – no matter how complex your thinking is, there’s another level that goes ‘Unless you think that I think that you think that I…’. There’s no real point at which you can say you’re done thinking.

Growing player display

In Poker this is sometimes called leveling. In essence, you want to think one level above your opponents. Any more than that is over-thinking. Any less has been taken into account by your opponent. You want to think as hard as they have, but one step further. The problem is – you have no idea how much thinking they’ve done and if you’re on two different levels the results are indistinguishable from simple probability. Both players throw a dart at a level in their mental dartboards and if they’re both in the same ballpark interesting things emerge from play. If they hit radically different numbers, then it’s a weighted coin-toss. Every bid, every raise, every bluff – that’s all information that helps an expert player assess the level at which others are thinking.

Effective leveling requires players of approximately equal skill levels and a similarly matched tendency towards cerebral consideration. That then needs to be coupled to a gradual release of important information that permits for clarity to gradually and reliably emerge. This is why it’s a difficult mechanism to get right. It’s a mechanism though that’s easy to make feel right – even if all it gives is the illusion of control that can be enough to be a satisfying experience. The problem in Secrets is that it doesn’t actually do either well. It incorporates neither effective scope for leveling or a sufficient misty veil of illusion that would convince people it does.

Effective leveraging of leveling is tightly bound up in familiarity – with the people, with the game, and with the solidity of new information. All three of these are problems with the way Secrets is pitched and designed.

The first is with the people – Secrets is a very straightforward game to explain. It’s like a gateway game, as much as I loathe that term, aimed at new players. Players who have cut their teeth on games like ONUW, the Resistance and Coup will have already seen pretty much everything they can hope to get out of Secrets. ‘It’s like games you’ve already played but a little bit different’ is a terrible pitch for an experienced, jaded audience. Secrets is simple enough to play with people that haven’t experienced better but not distinctive enough to be an easy sell for anyone else.

There’s definitely room in this market for an introductory social deduction game, but the problem with that is that Secrets actually requires some pretty deep game literacy to spark off fun. It’s got several interlocking gears inside its possibility space. You need to master the offering and acceptance or rejection of deals. You need to navigate the cumulative scoring of uncertain and changeable teams. And you need to be able to manipulate role identities in your favour. Any one of those has depths within depths and until you explore them a little it’s a game where you’re presented with seemingly meaningless choices and asked to make seemingly meaningless decisions. I watched this game absolutely die on its arse on several occasions with several groups of varying composition for this very reason. It’s easy to learn Secrets but it takes a lot longer to get the hang of all these interlinking and interdependent game systems to the point that you feel like you have any handle on what’s going on.

So, ‘a player needs to get good at Secrets to play it well’, right? Huge surprise buddy. Next you’ll be telling us games with lots of dice rolling need you to be lucky to succeed.

Bullet cards

Yes, to an extent this is true of every game but let’s go back to Poker and how leveling works. Every bid you make in Poker gives out some information, but the meaning of that information is what’s contentious. I bid ten dollars. What does that say about what I have and what I think everyone else has, and what I think everyone else thinks everyone else has? That’s the puzzle, and it’s why everyone that says Poker is a game of reading people is ultimately simultaneously right and wrong. Yes, people reading skills are essential but you’ll almost never be doing it at the level needed to turn it into genuinely reliable actionable intelligence. Almost never.

That massive uncertainty space about meaning begins to contract with the rest of the bidding, because everyone is testing everyone else’s defenses. You raise to fifteen. I raise to fifty. Knowing what you know of your cards, what are the chances that my cards are compatible with a bid of that nature. Knowing my cards, what would I deduce from your behaviour? Am I trying to frighten you off, or trying to make you think I’m trying to frighten you off? Am I bidding big to make you sense weakness? I mean, if I genuinely had a good hand I wouldn’t want to over-bid and make everyone else fold. I’d want to make small bids and escalate them gradually to increase my pot.

Every new bid though is a commitment, and that’s what essentially makes leveling in Poker a useful lens. I can’t spend money without giving information, and I’m constantly having to spend money. As such that data begins to acquire predictive weight as well as actual financial weight. The bids I put out in the game are simultaneously propaganda and investment.

You raise to one hundred. Bill to my right raises to two hundred. Jasmine raises to three hundred. And then I reach over and swap my hand for Jasmine’s. And then Bill and I share the pot because together we have a flush of seven cards.

Wait, what?

Can you imagine what that would do to everything in the game? But particularly, what it would do to any attempt you might have to link success to behaviour? The tightening of context that’s implied by this kind of game is rendered just one more uncertainty when nobody is tied to the consequences of their decisions – when I can win by swapping your cards with mine. In Secrets you can swap roles and take unassigned roles. You can pass cards to people that you know will do them permanent damage unless they end up being the hippy, which might happen through no choice of theirs. When your cards, your role, and your relationship to other players have no permanence it completely decouples accomplishment from activity. It becomes impossible, in my view, to become meaningfully better at the game because every single part of it is in too great a state of unknowability to permit control. You might win, but I’d be willing to bet a fair vivisection of the game session would reveal it was down to factors outside of your control. It might feel earned, but I don’t think it would be earned. Moreover, failure feels anarchic and incoherent, inconsistent with the mental model people were building of the game. All meaning in Secrets is temporary. All accomplishment driven by unaccountable whorls and eddies in the game state. Secrets is all activity, very little achievement.

United Nations token

The end result of all of this is that Secrets is a game that doesn’t permit for the building of mastery, but requires players to be masterful in order for it to feel meaningful. It doesn’t have anything really to recommend it to experienced players, but it asks too much of novices to encourage them to even attempt to build expertise. All of this is presented in a package that is deceptive – that looks and feels like something more controllable than it actually is. A game of this nature should be tense. It should be uncertain. It can certainly be funny but it shouldn’t be anarchic. This is the Fluxx school of design as applied to social deduction and you may recall I was absolutely not a fan of Star Fluxx when we reviewed it back in the day.

In the comparison to other social deduction games, it also suffers. It’s not meaningfully simpler than One Night Ultimate Werewolf or the Resistance but and yet a good deal more demanding in terms of game literacy. It lacks tension, because ultimately the constraints between which you are being stretched are too loosely anchored to be in a position to apply the necessary force. It all just feels a bit ephemeral and unsatisfying, and that’s not the kind of emotional response I want to have to a game of this nature.

I couldn’t recommend Secrets over any of the other social deduction games we’ve discussed on this blog because in the end it doesn’t offer me any of the things that I think are critical to that kind of experience. Worse, it doesn’t even really give me a get-out in the conclusion. The best I can say is that if being able to puzzle out roles and motivations is something you don’t want in a deduction game, Secrets is a decent Fluxx-esque version that is impressively well presented. And it’s from two game designers with individual track records that has earned them the benefit of the doubt over a career curmudgeon like myself.