Deception: Murder in Hong Kong

My love of deep strategy games may never fade, but I am finding social deduction games to be the most fun. They can fail spectacularly based on who you’re playing with, but they seem to have both a high average and high highs.

Deception is the latest one I learned, and it is a lot of fun. Players are detectives, and one of them is the murderer. Find the murderer and the evidence to win (or else the murderer wins). That is more or less the setup of most social deduction games, what makes this one special?

The game starts with information. One of the investigators is the forensic scientist. The game setup is that each player has four murder weapons and four clues in front of them. After everyone closes their eyes, the murderer points to one of his weapons and clues. Now the forensic scientist knows, and they can communicate only by selecting one from a set of clue boards like “scene of the crime” or “relationship to the victim.” Players then discuss these clues and try to deduce the murderer. (The forensic scientist is likely the least fun role, as it misses all the discussion of the game and needs a poker face to avoid giving away information. The forensic scientist can be the MVP or the cad based on quality of information provided.)

The fun of the game is trying to deduce what was supposed to be communicated, given limited communications options. Which murder weapons seem most like the cause of death, and would you have chosen that cause of death as a description of this weapon? For example, “severe injury” covers a lot more range than “poisoned,” and the murder weapon probably was not plague if anything other than “disease” was chosen (if “disease” was a choice). Committing murder during surgery is obvious if the location option “hospital” comes up, but what do you pick as the forensic scientist if that wasn’t one of the options this game? It matters what the scientist “said” and what was not said, along with how they expected those answer to be interpreted, mixed with the open question of whether that clue was supposed to refer to the blue or the brown card in this murderer’s combination.

The other great fun is trying to put together a story for the murder based on fairly random evidence. “Okay, the victim was killed with … a locked room, where the clue left behind was … timber. So, what, he was left to starve to death at a construction site?” Okay, that one was dull, but weapons include mad dogs and chainsaws, and evidence left behind can be quite random. That story also mutates with the clues, because the forensic scientist is trying to fill in the blanks, and some of the clue cards will have nothing useful at all, so then folks are wondering how “winter” fits into this story.

Balance seems to be pretty good, in that online commentary says the murderer almost always wins and that the murderer almost never wins. Granted, I won five games in a row, murderer and investigator, so I might be biased and would start thinking otherwise if my group was a bunch of foolish investigators who kept throwing it to the murderer. But our small set of games showed that both sides could win, under a variety of game sizes and circumstances (for example, optionally adding the accomplice and witness roles).

Hardest problem: avoiding meta-gaming. It does not seem to be a big problem, but we openly acknowledged around the table that some of the things we were doing were not exactly 100% fair to both sides, like discussing the forensic scientist’s clues while they were picking. Acknowledge that you could spoil that and take steps to avoid meta-gaming, such as making noise to mask movement while the murderer reveals hidden info.

: Zubon

Orders of Magnitude

Balance is easy to recognize by a feeling of indecision. If you have several options that are more or less equally attractive, and they are still equally attractive if you are fully informed, that is a well-balanced design.

In the idle game Realm Grinder, you can get a bonus for watching an ad. They are not mutually exclusive, so you can keep watching ads, but the balance between them is interesting. During an event, one is a very slow stream of the event currency, which is nice. Most of the time, what you want is doubling your production for 4 hours. That is orders of magnitude better than the next option: +10% mana for 10 minutes. There are a few points in the game when that +10% is the difference between being able to complete a combo, but it effectively means that you get 1 minute’s worth of mana. Compared to doubling your production for 4 hours.

The last option is faction coins, which scales with your production. Ever so slightly. I clicked the button for an ad, and if I pick faction coins, I get 3500. I need 2.655*10^18 faction coins to get my next spell tier, and that is after using rewards to cut its cost. My costs are expressed in scientific notation, and the ad reward is 3500. Assuming 30 second ads, I would need to watch more than 721 million years of ads. Wait, no, there are 8 types of faction coins, and I need that many coins from two factions, so the number is a but below 3 billion years of ads.

Sometimes we say that an optional activity in a game is effectively required. These ads are very much not required.

: Zubon

Precipitating Events

Sometimes it takes just a little poke to make you wake up and rethink the whole thing. This is a brief story of a rethink and then the poke.

I gave up all my mobile games. I deleted all of them. The real start of that was having added one of the many farming games, which was quite absorbing but quite a timesink. I deleted it after narrowly missing a minor goal and thinking of how much babysitting the game needed to reach goals. And then that it is legitimately “babysitting” instead of playing. And then Spellstone, the P2W CCG I had been playing, went extra P2W and made me wonder how much I was getting out of any games I had on my phone.

That P2W bridge too far was the Ice Tunnel Dungeon, specifically the Winter’s Crusher custom card made for it. The dungeons are PvE events with soft caps based on how much money you have spent. As you go higher, the enemy heroes get stronger abilities, and their decks tier up to fully upgraded premium cards. Premium cards are a separate set only available with the cash shop currency, and they are stronger than their base set equivalents and keep getting stronger through the usual power creep that keeps people spending money. But then the events have custom cards that are even stronger than those, with no real thought of balance because they will never be in player hands.

For folks who don’t know this game, let me give you a quick sketch. This is a fully upgraded base set (F2P) card. It has 7 attack, 16 hit points, potentially buffs itself to 11 attack, ignores 4 layers of shield, and vampirically heals up to 3 hit points per round. Winter’s Crusher has 4 attack, 48 hit points, adds 5 layers of shielding to itself and its allies, heals itself and its allies for 4 hit points per round, and weakens all opposing attackers by 5; due to the current battleground effect, it also self-heals for 12 hit points per round, even if frozen. It effectively counters 26 points of damage against itself and 14 against each ally, versus the attacking frog that can buff itself up to 11 attack. That is one card in the enemy deck.

Someone did that on purpose. There was probably a meeting, and it had to pass review from a half-dozen people. Loss of faith in the developers’ judgment? I quit the game so hard it deleted all the games off my phone.

: Zubon

Boom Beach I was just bored with after sitting at the soft cap for a couple of months.

Thousand-Dollar Decks

Spellstone is still my go-to online CCG, because I always seem to have one going. I have been playing long enough that I now have a top-thousand deck, and I am increasingly meeting the players who pay for the servers to stay up. I used to mentally refer to them as “hundred-dollar decks,” but looking at some of these top decks I run into, no, these are definitely thousand-dollar decks.

Spellstone always has a couple of premium card “boxes” running, and buying one out completely costs $300-$400, depending on how you buy your RMT currency. The reason to buy one out is to get 4 copies of each premium legendary, because you upgrade and combine them into one quad legendary. You can get 4-6 quads out of a full box, limited to 4 copies of the same legendary. I see decks that have three copies of the same quad premium legendary, meaning these folks had to buy out at least three boxes, and those are for recent cards so these folks are paying ~$1000 for that deck and will need to do so every two months or so to keep up with the P2W curve and shifting environment.

I know in principle that people spend this much on virtual card games, but I did not really get it until I saw the fight in the P2W ranks of the game. I should know this because I played Magic: the Gathering when it first came out, but it is still surprising to see, and I wonder about things like stolen credit cards.

twitter screenshot referring to the Magic the Gathering 'power nine' as part of an investment portfolio

: Zubon

Incentives, Peaks, Valleys

Events add new incentives to games, or sometimes they are just additions to existing rewards like double experience points or gold. For regular players, these are nice boosts, and more players log in for the increased rewards.

And then the weekend is over. The visiting friends are gone. You can run that dungeon again for half the reward without the holiday event buffs.

You are exactly where you were the day before the event, but it sure feels a bit lower.

: Zubon

The Room 2

I previously enjoyed The Room, so I picked up The Room 2 on the Steam Halloween sale. It is what you want from a sequel: more of the same but better. In this case, we are less confined to a puzzle box and instead get to see a variety of puzzle rooms (or is it the same room across several dimensions?). More puzzles, more steampunkish mechanical boxes that whirl, better graphics and art, creepy settings for Halloween. Recommended.

The biggest improvement from the original is the expanded scope. You get puzzle boxes, but you get a seance room, a laboratory, a fallen temple, and more. It’s good. The puzzles are mostly fun, and there is joy in watching it all come together, watching the completed device whir. The use of the eyepiece is significantly improved, with visual indicators that you might want to try it out rather than randomly needing to double-check everything, and the places where you need the eyepiece are more intuitive. The shifting camera does a good job of directing your attention towards the next step.

The flaws of the game remain the same, more or less the flaws of this type of game. Sometimes that shifting attention feels like a trail of breadcrumbs you are following, a very visual non-novel, particularly at the beginning when there are fewer moving pieces (less so by the final room). Sometimes the game operates on old school adventure game logic, i.e. arbitrariness that might have made sense to the designer. Why does putting X in Y make Z open? Why is there a fuse in there? Just keep playing the game, and the illogic is less severe and less common than in the previous game, although some of that might be having learned how the developers think after playing through the first game. The settings of the various rooms lack a unifying theme other than “A.S. was here,” but they are rather nice rooms with variations on “creepy.”

The puzzles are better, the settings are better, the graphics are better. If you liked The Room, you will like The Room 2. If you almost liked The Room, you will probably actually like The Room 2. The game is short, but it is longer than a movie, more interactive, and far cheaper, even before the Halloween sale.

: Zubon


I occasionally use the term “elegant” in reference to design, generally meaning that complexity arises from the interaction of a few simple systems. The Civ V Krepost is one of my favorite examples, one bonus on one building creates three effects that yield Russia.

Definitions are as much about what is excluded as what is included, and our friend Wilhelm has effectively pointed to the opposite of elegance:

Paradox games are deep… as in fall in, founder, and quickly drown levels of deep. There are always lots of moving parts that influence each other that you have to keep track of so that the initial experience for all of there games seems to be getting totally lost in a morass of details thrown at you in rapid succession that quickly leads to overload, exiting the game, and rarely, if ever, returning to it.

This also highlights a reason why elegance in design is desirable. Elegance yields “easy to learn, hard to master.” Inelegance yield confusion and “eh, I’ll get back to it” … and not getting back to it. I never finished the tutorial for Banished because the first one involved multiple actions that each went several layers deep in menus, with little explanation of why would want to do those but a clear implication that these were basic actions that would frequently be necessary. User-oriented design: good interfaces naturally guide you to the action you want to take.

: Zubon

Are Your Eyes Part of a Botnet Attack?

No, not yet, but let’s pair two stories:

  1. The Internet of Things participates in DDoS attachs
  2. Let’s add more things to the Internet of Things

The idea of the Internet of Things is to let all the dumb things in your house talk to each other, thereby collectively creating a smart world. (If that sounds dumb, remember that all the cells in your body are individually dumb, thereby collectively creating you.) Lots of devices now have some degree of connectivity, like DVRs, security cameras, thermostats you can control via app, keyless locks, or your garage door opener. Now think for a moment: when was the last time you changed the password on those? If you have one of those fancy, programmable toasters that makes a picture on bread, do you even know how to access security on it? Maybe they do not even build security into a toaster.

I remember mocking a headline about hackers “using your blender against you,” because the worst case scenario of a hacked blender seems like a lousy smoothie. So far, the apparent worst case scenario is contributing to a DDoS attack. If we now have millions of insecure, dumb devices online, and many of them can make some online requests, you can now flood anything with requests from millions of vectors. And those devices will continue to be vulnerable until they are out of service. Do you even know how many connections you have to the Internet of Things? If you know every device, can you access its firmware and update its security settings?

Adding in that second article: and now we have proof of concept for letting non-powered devices contribute to the Internet of Things using signals already in the air. “The goal is having billions of disposable devices start communicating,” he says, adding to the millions of insecure devices already doing so. The example in the article is contact lenses, hence my headline.

If we are not building security into the Internet of Things, we are building a world where you could attack a computer literally just by looking at it.

: Zubon

Unrelatedly, did you know that you can damage computers by shouting at them? Sound waves are physical vibrations, and computers can be very sensitive.