A Perfect Game of SpyParty

I have a lot of stuff queued up to post about, but the other day, tytalus was playing theinfamousnolan and they were streaming both sides,1 and this game happened, and I simply had to write about it.

First, since Nolan’s the Sniper in this game, you can sort of play along with him, so watch this first:

YouTube Preview Image

Now, here’s Tytalus’ side, where he’s playing Spy:

YouTube Preview Image

Now, on first viewing, you might not be that impressed.  “He got lucky”, “Nolan made a mistake and shot too early”, end of story, right?

I don’t think so.  I think this is a perfect game of SpyParty, and I’m going to analyze it fairly closely to explain why.

It’s important to point out the title of this post is A Perfect Game of SpyParty, not The Perfect Game of SpyParty.  I don’t want to imply this is the only or the best way to play.  Any deep competitive game—which is what I’m striving to create here—needs to make multiple overlapping strategies available to the players at any one time, and the players need to constantly react to the current situation to choose the best one, or the least-worst one.  However, this particular game exemplifies the pinacle of a certain style of SpyParty play, and I want to talk about why.

The whole thing’s a set up. A scam, a frame job.

One of the advanced ways to play SpyParty as the Spy is to try to “frame” an NPC, and get him or her shot by the Sniper.  This is quite distinct from the Sniper choosing to shoot the wrong person even though you were trying to complete your missions; that happens all the time, even in elite games.  Framing somebody is a step beyond that, where the Spy spends precious time doing certain actions in a certain way to try to implicate another partygoer, to “help” the Sniper shoot the wrong person. There’s a halfway strategy of just trying to get a lowlight for yourself, or getting a highlight for somebody else, and then using that as cover for doing missions as well.

It’s a risky gambit, because you can waste a lot of valuable time trying to frame somebody, and it’s very hard to recover when you’re running low on time, so you have to be careful if you’re going to choose this path.   Also, you have to have a fairly high level of micro skill so you can be pretty confident of being able to act like the NPCs convincingly while you’re framing another one, and to hit green Action Tests on missions, both of which which are important to effectively frame someone.

I think Tytalus’ play here exhibits incredible observation, strategy, patience, and yes, luck, or more correctly, being able to take advantage of luck when it comes your way.  I would have screwed this game up at multiple different points as Spy, and I’ll try to note those below as I talk about the game.

Observation

Tytalus’ observational skill was in play long before he started talking in this game.  He almost ran into “the Red Dress” (Cybil Disobedience) at the statues immediately on spawn (0:11), but since this is the Time of Chaos™ you can get away with a lot of jerky movement and direction changes during this beginning phase, so he just bails from the statue and goes to hang out in a conversation to see where things settle out.  This is the first place I would have screwed up.  Since the Spy spawned near the statues, I almost certainly would have at least done an Inspect Statues, and probably would have earned a highlight for my troubles, which also would have made Cybil bail from the statue and go do something else.  I’m too focused on completing missions as the Spy to observe subtle details about the party flow.

At this point, Tytalus is watching Nolan’s laser (which you can see if you set the video to HD with the little gear icon) to see what he’s looking at.  He also checks his twitch.tv chats and answers one about contacting the Double Agent.  He notices that Cybil goes straight from one statue to another, which while not exactly rare, is certainly a reason to be suspicious of somebody and probably got Cybil highlit if she wasn’t already from her “statue rush”.  This is the second place I would have screwed up, because I wouldn’t have noticed that Cybil was the object of the Sniper’s attention for the back-to-back statue visits.

Tytalus moves to the conversation group containing the Double Agent at 0:37. 

Strategy

At this point I think Tytalus has formulated a plan for framing Cybil and getting her shot.  At 0:45, from the safety of a conversation, he says, “Now he’s watching the Red Dress pretty heavily, so what I don’t want to do is scare him off right now.”  This means that if Tytalus had gone and done the Contact Double Agent action right then, the audible tell of “banana bread” would have immediately eliminated Cybil as a suspect because she was still at a statue.  This would have made her impossible to frame, so it would have gotten a mission done but flushed the opportunity to frame down the toilet.   I probably would have done that, making three mistakes.

Tytalus says he wants Cybil to go into a conversation and talk, so then he can hit a green Action Test on the Contact, making it look like he’s not talking but she is, and probably increasing Nolan’s suspicion of her.  Tytalus doesn’t know for sure he can hit a green Action Test, because it’s hard,2  but it’s worth a shot, and if he waits for her to talk, even if he hits a white action test and talks himself, she’ll still be talking and be more suspicious than he is due to the statue visits.

Patience

“Red Dress should chat…c’mon Red Dress, chat!  Chat, damn you!”

Although frustrated that Cybil isn’t talking, Tytalus is incredibly patient here.  He’s in a conversation with the Ambassador (who is in range for the bug!) and the Double Agent, and he still plays it cool.  He knows Nolan is not seeing any real tells, so is stewing in his own juices, manufacturing suspicion and starting to panic that he’s missing something.  He knows this because this is what Snipers do—this is what human brains do—they see patterns even if there aren’t any.  Similarly, Nolan knows Tytalus is higher ranked than he is, so he’s panicking even more that Tytalus is just invisible due to their skill gap, but is still accomplishing missions, and Tytalus knows Nolan is thinking this.

The Ambassador and the Double Agent leave the conversation, and he still plays it cool:  “I don’t even care, I don’t care if there’s no…” (1:03) 

I would have chased one of the two of them, or gone and done a seduce, or a statue mission.  I would have been sweating the clock, since I’d wasted so much time trying to frame Cybil.

Luck

But, Tytalus is patient, and stands there as the clock ticks down.  He knows Cybil has to go talk at some point, and at 1:21 his ship comes in.  Cybil chases the Double Agent and the Ambassador, and then sidles right up to the Ambassador, as if she wants to cuddle (or bug) him.

Boom, headshot.

“YES!”

Poker versus Go

Now, you might say luck was a huge part of this, that it was the thing that won the game, but I don’t think so.  Luck played a major supporting role to be sure, but it was Tytalus’ read on the situation, and his steadfast refusal to screw up an opportunity for a good thing to happen that won him this round.  I bet, even if Cybil hadn’t landed right next to the Ambassador, or had gone to a different conversation circle, that Tytalus would have still gotten her shot.  He’d played brilliantly up to this point, dealing with bad luck (like Cybil not talking for 30 entire seconds, an eternity in this game), and so I don’t see why he wouldn’t have responded to whatever else happened with the same virtuosity unless things had gone completely unpredictably pear-shaped.3

In poker terms, he was dealt a good hand, he turned it into a great hand with consistently excellent play, and he got dealt the perfect card on the river to take him unambiguously over the top.

I am a strong believer in luck and I find the harder I work the more I have of it.

Benjamin Franklin

Go has always been my favorite game, what I would consider the most beautiful game designed by humankind.  When I first started designing SpyParty, my goal was to aim for Go, with its crystaline precision, its complete lack of randomness that would—I believed at the time—dilute the player-skill aspects of a game.

Over the years, I learned more about the very different beauties of Poker, and I began to realize regardless of where I wanted to aim design-wise or what my own preconcieved aesthetics were, when I allowed the game to lead the design naturally, when I listened to the game, SpyParty was actually heading towards Poker, with its beauty coming from probabilities and expected values and seizing the opportunity as it’s presented.

I certainly don’t expect SpyParty will ever be as well-designed as Go or Poker, but I am happy that it’s inching in the direction of allowing play at the level exhibited in the game above, and I plan to encourage this and take it even deeper.

If you’re interested in this topic from a game design standpoint, you really must watch two lectures:

If you’re interested in watching games at this level, you should follow some of the elite level SpyParty streamers on twitch:

And, yes, of course playing the game is good too!  More invites will go out tonight, and the open beta is coming soon!


  1. …which is awesome, I love it when people stream both sides! []
  2. I haven’t run the numbers in a while, but current estimates among elite players are around 30-50%, there’s some old data here. []
  3. …which, of course, can happen too! []
  4. I’m going to start streaming soon, as well!  More on that later. []

23 Comments

  1. Sergio says:

    Oh gosh, this is brilliant. I can’t wait to play and review it in my Brazilian pc game site!

    Damn open beta, arrive soon. :)

  2. dfan says:

    Spy Party will never be anything like Go, as far as I can tell. But it is a hell of a lot like poker, and that is a high compliment. Embrace it!

  3. Nolan says:

    Awww. I’m happy and sad about this. I’m glad you got to make a post out of it, but sad I was made a fool haha. It’s all good though because I learned a lot from those matches

  4. David says:

    Hi,

    I received a Beta invitation but I regret to inform you that I will never pay 15 dollars to beta test a game. Maybe for the final game, yes.

    • checker says:

      No problem, just hit the “unsubscribe” link in the invitation and the scripts will leave you alone from now on.  I tried to make it really clear that it’s a paid beta on the signup page (bold red text above the form and all), but sometimes people still miss it.  Anyway, all kinds of fans are welcome, beta testers, blog followers, stream watchers, etc.

    • Lee says:

      It also says that the $15 you spend gets you a copy of the final game once it’s finished. Seems like a good deal to me…

    • checker says:

      Some people don’t like the idea of paying to beta test a game (even if you get the full game on release), which is totally a-okay with me.  Luckily, Minecraft and Frozen Synapse and other indies have helped pave the way for a lot of people to want to do this, so it’s a viable funding model, which is awesome because I think it’s a very healthy way to fund the game for the developer and players.  But, nobody should feel obligated, which is why I try to be really clear it’s a paid beta on the signup page.

  5. I thought a proper response was in order regarding the series of events, because they help shape why I did what I did.  After Checker noted to me that he might use the match referenced in his blog post I made sure to note why I did what I did so I could better frame the context that led to me choosing to play in this way.

    First, the entire stream is available here:  http://www.twitch.tv/tytaluswarden/b/339702909 .  I understand why Checker chose not to post it… it’s the longest stream I’ve ever made (nearly 3 hours) and it’s a beast to navigate through.  I’ll post other links below with timestamps to reduce the burden should anyone be interested in viewing the scenes I am referencing.

    So at or around 2:17:35 ( http://www.twitch.tv/tytaluswarden/b/339702909?t=2h17m35s ) I am the Spy and we’re on Courtyard.  The first thing I do is go to the statues, where I earn a highlight.  I know this because Nolan shot me a minute or so later and the summary clearly indicates the highlight.  Summary is found at 2:18:40 ( http://www.twitch.tv/tytaluswarden/b/339702909?t=2h18m40s ).

    At or around 2:20:30 ( http://www.twitch.tv/tytaluswarden/b/339702909?t=2h20m30s ) I am the Spy again, and we’re on Courtyard 2.  The first thing I do is go to the statues, where I earn a highlight.  I know this because, again, about 2 minutes later the summary shows I’m dead and Nolan highlighted me.  Summary is found at 2:22:25 ( http://www.twitch.tv/tytaluswarden/b/339702909?t=2h22m25s ).

    At or around 2:26:08 ( http://www.twitch.tv/tytaluswarden/b/339702909?t=2h26m08s ) I am the Spy yet again, and we’re on Ballroom.  As you may now predict the first thing I do is go to the statues, where I earn a highlight.  Summary of my death can be found at 2:28:16 ( http://www.twitch.tv/tytaluswarden/b/339702909?t=2h28m16s ).

    With these three losses in mind I realized Nolan had figured out my current play style.  He was hyper-focusing on anyone approaching the statues while on these maps and eventually (correctly) shot someone who had both been at the statues and performed one suspicious activity (i.e. walking near Ambassador, walking even a little odd, being in a conversation circle when banana bread/contact DA was performed, etc.)  I had to drastically change this up or he’d snipe me every time!

    We were playing a “cycle”, which is my loose term for playing one time as Sniper, then one time as Spy, on a particular map, then switching to the next map in the “map cycle”.  We were currently on Ballroom and I wouldn’t be Sniper on that again.  The next map is Balcony, no statues there.  The one after that is Veranda, a map that’s way too big to go framing someone who “rushes” the statues.  Then we got to Courtyard…  that’s when I thought it’d be good to try and frame someone rather than playing to complete all objectives.  Both strategies are “playing to win”, and in the past we’ve had a lot of fun doing this.  (For a quick example with some explanation, Viriflod has intentionally failed seductions on balcony to get me to shoot the seduction target… the seduction target leaves a conversation circle when the Spy fails the seduce action.)

    The article was written about the gameplay starting at or around 2:46:36 ( http://www.twitch.tv/tytaluswarden/b/339702909?t=2h46m36s ).  The game before this I had an opportunity to add time, which blew up in my face and forced me to attempt to complete objectives to win, since I thought eventually he’d decide to shoot *someone* and I (correctly) thought he had already highlighted me.

    I hope this additional explanation provides some enjoyable reading for those who bother to read blog comments!

  6. Dan says:

    Let me just say: I really want this game. It hits everything I want from a strategy game, it has an incredibly interesting concept and what looks to be quality execution.

    The problem? The pricing.

    I’m perfectly willing to pay 15 dollars for this game. My friends, who are a little less game-savvy and a lot less inclined to fork 15 over for a low-fi high-concept game like this (You should have heard the response when I suggested we all get sleep is death), are not. This is a game that would be made much much better with friends, and I hope you explore bundling together 2 for 20$ or 2 for 25$ so that I have a much better shot at convincing my friends.

  7. Dan Jammin says:

    I don’t have many games I’m really excited about at the moment, but SpyParty and Monaco are two that I’m really looking forwards to playing with my friends.  This match was fun to watch, and I can almost imagine the tension I’d get from trying to balance mission progress/visibility.  I missed my first email invite, but as soon as I get my next one, or the open beta starts, I’ll be in there, probably killing civilians by accident.

  8. aureliend says:

    Hi ! Do you know when open beta will be available, even if it is only an estimation like “maybe in two months”? I have searched around a little and did not find the answer. By the way, even if it is an open beta, I think it will still be a “paying” one, isn’t it? In fact, I am eager to test the game, but I saw you plan (maybe) to sell double licences for discount price. This last possibility is very interesting to play with a friend. 
    So I have a crual dilemma as I really want to play the game but I don’t want to pay 15$ for a single licence if in one or two months I can have two of them for only 20$ (for example). 
    Sorry for the question if I missed the answer, and sorry for the English mistakes as I am French :) 

    • checker says:

      Your English is way better than my French, so no worries! :)

      I’m almost certain to do a Frozen Synapse-style discount-on-two-copies at some point, but it’s not going be for a while, probably, there’s a lot of stuff to do in the meantime and I’m very slow. This is partly why I really dislike sales and stuff, and wish they didn’t “work” as well with people as they apparently do. It feels unfair to buy something and then have it go on sale right after, and as a creator the last thing I want to muck around with is pricing strategies and whatnot, it’s such a distraction. But, people love sales and deals and whatnot. Maybe I could do something where I charge existing beta testers the difference for a two-for-one deal or something.

      Anyway, I hope to have the beta open within weeks, not months, and then let that stabilize.

    • aureliend says:

      Thanks for the swift answer. I think I will try my luck with the beta so :)

  9. Matt says:

    When is the next round of Beta invites going out? I signed up on the 22nd of November, and have been checking my email daily for the invite. I’m dying to play this game.

    • checker says:

      I try to do about 500 or more a week recently, as I test things for getting the beta open. Hopefully everybody signed up will get an invite within the next few weeks, but I’m notoriously bad at estimating how long it takes to do tasks. But, soon!

    • Matt says:

      Alright. Eagerly awaiting my invite.

    • LowTierSteve says:

      I think about half of the invites checker sends in each batch are random, the other half go from the front of the list forward, and I think you might be about 10,000 people down the list (at least), based on what I remember from his last set of numbers. Maybe even more like 15,000. Signed up months ago, still waiting. But I’m super excited for the chance to play, probably when the open beta comes, and check back to the site daily for updates.

    • Specifically, he typically does a 60/40 (60% in-order earliest first, 40% random through the rest of the queue).  There’s been certain situations where he’s strayed from this format/formula, but almost always this is the case.

Leave a Reply


five × 3 =