It is exactly what meta-gaming is, tho. You have outside information that gives you advantage over who doesn’t have it.
Regardless of how you define meta-gaming, this information can be exploited to gain an unfair advantage without detection so hiding it is poor policy.
1 Like
Not my definition, though. Well, sure, you can try to gather this info by yourself, but the results would be far from the exact rates being given. Repeating: this is bad, baaad. Limits claim space, gives you specific people to target based on how likely it is for their claim to spawn and this goes downhill to a game with even less good deceit. Idk why this need to have control of info like that lol this removes the fun of the game almost entirely.
gives you specific people to target based on how likely it is for their claim to spawn
We already have this to a certain degree. And you have not addressed the problem that players who put in the necessary effort can discover this information for themselves to get an advantage over others.
The more games played the more accurate the data will be, how can you say “far from exact”? If someone records data from their next 30 games they will have a good idea of the spawn rates. You probably only need 20 games to have enough information to the point that it’s “unfair”.
2 Likes
I feel that sheriff is, without a doubt, the best Special/Investigative claim to make, since if there are not 2 sheriffs, you are fine. If there are 2 sheriffs, play scorned, or fool. Since you will appear NS for some nights, you could play it of as a Fool. Claiming a common role (like sheriff/phys) will often be a fool claim as well, and sheriff is one of the most common claims for scorned.
Princess will guarantee your death when the king dies, and Obs requires both luck and a lot of work to compile everybody’s logs together. If you are asked for logs too soon as an obs claim, you will likely make a mistake.
1 Like
Correlation is not causation! Inq spawn could be 5% and you still get an Inq game in 8/10 games. Tell me how accurate your data would be.
Well kind of how statistics work… generally the more data points (IE games played) the closer your percentages will come to reality. Sure it’s possible 1% scenerio’s happen to hit 30 times in a row… but it’s very unlikely. If 3 people come together with 200ish games played each and combine their data, you’ll most likely find something nearly identical to the sets (unless the dev’s ninja tweak the settings every few weeks which is also fully plausible, that’s the biggest challange in self collecting data is not knowing when the odds change)
1 Like
Yeah, but I don’t get why this would be so much of a bad thing to justify devs giving us the exact information.
It is a bad thing, because developers not giving us this exact information gives players who go to the trouble of finding out those spawn rates a big advantage over players who do not. You can simply record the frequency of specific classes spawning and basic machine learning algorithms can approximate reasonably well the actual frequency or class rules that are in place. This makes the information assymmetric, which is a way worse situation than symmetric. If everyone knows exact class spawn rates, it is way more fair than only experienced players. It will not widen the gap between newbies and veterans in that way.
1 Like
So I messed around with a binomial confidence interval calculator and apparently I’d need 70 matches to get spawn rates +/- 10 percentage points with 90% confidence.
Quite a bit more than my estimate of 20 or 30 matches lol
Anyone want to help record spawn rates? I recorded endgame screen from 4 matches so far
Honestly I’d say it’s probably less of a factor than it’s worth noting at a cursory glance it seems like things are close enough that it’s hard to imagine them being significantly different enough to actually be different enough. IE a hypothetical in which the last 2 possible evils are a knight and a CW claim. One having a 2% higher odds of having started in the game is probably relatively unlikely to be as much of a factor. Doubly so if you have to explain your reasoning to get bd to pull the trigger on it.
That being said I also do agree in general security by obscurity isn’t the best of systems. Town of salem went through a huge bug at one point, in which classes could more or less be identified by what order they showed up on in a trial. There were timeframes in that bug… about a year where it was most likely only known by a handful of players that shot to the top of the rankings with their knowledge… and a year later when someone posted it publicly and then every game everyone knew of it… the bug of course wasn’t fixed until a week after it became public knowledge.
More like 10 IQ.
But if you’re wondering whether to fake-claim Sheriff or Knight as Mastermind in jail N1, knowing the approximate spawn rates of Sheriffs and Knights gives you a significant advantage over those who don’t.
right, but I’m saying those are probably close enough that it’s a pretty insignificant effect on your winrate. (well minus the part that we do know… of which there is 1 guaranteed sheriff in every game, which automatically makes it so that unless there’s some insane reduction to the odds of a class being sheriff sheriff is most likely twice as likely to 2 more.
Okay I see your point. Which I think is basically it’s not worth recording data from 100 matches to figure out spawn rates to get a significant advantage?
My point is the significance of the advantage is probably very very low. It’s likely the difference between turning a 50/50 into a 55/45 off of statistics alone, and quite likely trying for it is more likely to cause you to miss something in the logs/chat etc… that could have actually solved it. I don’t think published or unpublished it would have a significant advantage in winrates in either direction.
That being said, I probably lean more towards transparency over secrecy with regards to those rates. Just on a general principle that if someone cares they will find out and it creates more of an image of fairness for it to be known, even if the information is worthless strategically.
1 Like
