Okay, so I know that straightforward challenges still have a slight risk of failure, but I’ve failed more of them than I’ve succeeded in over the last couple of days. It’s starting to seem a little weird that I’m having a better success rate with 65% challenges than 100% ones. Did something get tweaked that made this chance greater lately, or am I just astonishingly and profoundly unlucky?
Aha, I knew the NSA and the rats were in on it somehow!
So be it–I am fairly used to having appalling luck (I almost never, ever, do the luck challenges for that reason, unless the unlucky option gets you something as well) but it seemed to be a bit more than usual lately, so seemed worth checking. I’ll just go sacrifice a goat or two and see if it improves ;)
That thread really didn’t answer any questions, though. And it looks like Alexis further shut down discussion right when it got interesting. A lot of the programming stuff is way over my head, but the last person xKiv seems to be on to something when he says that randomisation isn’t the same as probability. There’s definitely something hinky with Fallen London’s RNG.
Logic time: if a random number generator produced consistent results, it wouldn’t be random.
And thank you, Flyte, for drawing my attention to what might be the most amusing mod responses on the forum.
Straightforward green or straightforward white? If it’s green then it’s just the RNG god playing with you. If it’s white then you need to get worried.
Oh! I always wondered why sometimes ‘straightforward’ was green here or white there. That’s an interesting tidbit of information to know.
All shall be well, and all shall be well, and all manner of thing shall be well.
Don’t think he really did. He said if you want to talk programming that’s unrelated to FL, take it somewhere other than a FL based forum. Which is fair enough, really. If you really want to continue: I suppose programming chat would work in the Off Topic boards. But really - it’s programmers arguing with programmers about a hypothetical problem that’s already been answered as ‘not a problem’. Someone on a streak of bad luck (no one complains about streaks of good luck) assumes there’s something wrong with the RNG, Alexis says they’re using industry standard. You can continue to believe there’s something wrong - but either the industry standard is broken, FB don’t know what they’re doing re RNG, or you’re wrong. Since I’m not a dev/programmer, I’m not going to automatically know that the industry standard is right, and the heartbleed thing shows that kind of thing does go wrong - but it is far more likely that you’re on a streak of bad luck, than the industry standard is broken (or FB have been blagging it this long). xKiv seems to think he might be a better programmer than Alexis and his team, or that the FailBetter team might have missed something. He might be, they might have. Unlikely, especially as they’ll have seen the posts and the follow up, and if they’d have applied would have applied them. And talked about it, in all likelihood. FB own up to mistakes, and fix things - 'tis part of why we love them.
There HAVE been instances of a tricky to find RNG roll-type bug - e.g. the Wii Flag. Wi Flag | Asheron's Call Community Wiki | Fandom
(Ah, now THERE was a game.)
But really? After him coming in and talking about what they use and explicitly stating it’s not a problem, it’s just the sheer numbers of people playing mean that statistically some people HAVE to have long streaks of bad luck (ditto good) - it’s a little presumptive to come right back and say ‘you’re the game dev, but you could be wrong and your work is broken’. You don’t think Alexis will have checked all this, given their entire game runs off rolls by the RNG? It’s in their interests to have it running smoothly.
Of course. It’s far more interesting to think there’s a conspiracy. There probably is. Now … ;-)
edited by babelfishwars on 4/11/2014
RNGs are not perfect, but they are pretty good. In the case of FL, I was collecting data on Unfinished Business, and over 300+ rolls, the actual failure rate was within 1% of the stated failure rate. [li]
Oh! I always wondered why sometimes ‘straightforward’ was green here or white there. That’s an interesting tidbit of information to know.[/quote]
Yeah, it’s a pretty recent change, and a very useful one too! Otherwise, you could always click the ? icon to see exactly what percentage you have, straightforward could be 99% and you are just unlucky. If it’s 100% success, then it’s white.
[li]
[quote=babelfishwars]
Don’t think he really did. He said if you want to talk programming that’s unrelated to FL, take it somewhere other than a FL based forum. Which is fair enough, really. If you really want to continue: I suppose programming chat would work in the Off Topic boards. But really - it’s programmers arguing with programmers about a hypothetical problem that’s already been answered as ‘not a problem’. Someone on a streak of bad luck (no one complains about streaks of good luck) assumes there’s something wrong with the RNG, Alexis says they’re using industry standard. You can continue to believe there’s something wrong - but either the industry standard is broken, FB don’t know what they’re doing re RNG, or you’re wrong. Since I’m not a dev/programmer, I’m not going to automatically know that the industry standard is right, and the heartbleed thing shows that kind of thing does go wrong - but it is far more likely that you’re on a streak of bad luck, than the industry standard is broken (or FB have been blagging it this long). xKiv seems to think he might be a better programmer than Alexis and his team, or that the FailBetter team might have missed something. He might be, they might have. Unlikely, especially as they’ll have seen the posts and the follow up, and if they’d have applied would have applied them. And talked about it, in all likelihood. FB own up to mistakes, and fix things - 'tis part of why we love them.
There HAVE been instances of a tricky to find RNG roll-type bug - e.g. the Wii Flag. Wi Flag | Asheron's Call Community Wiki | Fandom
(Ah, now THERE was a game.)
But really? After him coming in and talking about what they use and explicitly stating it’s not a problem, it’s just the sheer numbers of people playing mean that statistically some people HAVE to have long streaks of bad luck (ditto good) - it’s a little presumptive to come right back and say ‘you’re the game dev, but you could be wrong and your work is broken’. You don’t think Alexis will have checked all this, given their entire game runs off rolls by the RNG? It’s in their interests to have it running smoothly.
Of course. It’s far more interesting to think there’s a conspiracy. There probably is. Now … ;-)
edited by babelfishwars on 4/11/2014[/quote]
[li]
I was getting interested in the nitty gritty between people who knew progamming and statistics talking about the game design – when it was asked that discussion stop. Just a personal observation. It felt like xKiv was brushing on a topic that I was also thinking of but couldn’t really flesh out as I’m not mathy enough, so I wished the discussion had played out.
Why was I interested when normally I duck numbers like they’re made of death? I really dislike the game design in FL around the RNG and wanted to learn more about it.
Mostly, it doesn’t make narrative sense. I’m a Master Thief and I’m losing a Very Modest skill check literally 10 times in a row? This happens often enough that it’s frustrating. I’m losing out on precious action points to explore more of the game!
Yes, they may be using an industry standard RNG, but I feel there could be more sophisticated filtering/framework/design built around it to produce a statistically-sound system which has
- less frustrating, more intuitive gameplay
- integrated narrative
That’s why I was intrigued when xKiv was talking about the different reset duration options, such as during the course of 1 session or globally across servers (totally mangling the actual terms, but I’m writing out the gist).
Because if you’re flipping a coin heads or tails, the odds are 50/50 but in an infinite timeline it’s statistically "correct" for the coin to come up heads 1 million times in a row as long as it ends up tails 1 million times somewhere in the tail end of infinity. It just would be a hell of a lot less fun playing with this time table.[/li][li]
[/li][li]And regarding this line: [ [color=rgb(119, 119, 170)]it’s a little presumptive to come right back and say [/color][color=rgb(119, 119, 170)]‘you’re the game dev, but you could be wrong and your work is broken’ ][/color][/li][li]
[/li][li]Yes…? And we are the game players, who like the game enough to discuss it among like-minded fans, both the parts we like and the parts we dislike, with constructive comments. No one is infallible, as far I’m aware?[/li][li]
[/li][li]And the more I think about it, I really don’t think there NEEDS to be people who have such long (good or bad) luck streaks. There could be tweaks made to alleviate the issue that obviously a lot of people are having.[/li][li]
edited by cinderfallen on 4/12/2014
[quote=cinderfallen] Why was I interested when normally I duck numbers like they’re made of death? I really dislike the game design in FL around the RNG and wanted to learn more about it.
Mostly, it doesn’t make narrative sense. I’m a Master Thief and I’m losing a Very Modest skill check literally 10 times in a row? This happens often enough that it’s frustrating. I’m losing out on precious action points to explore more of the game!
Yes, they may be using an industry standard RNG, but I feel there could be more sophisticated filtering/framework/design built around it to produce a statistically-sound system which has
- less frustrating, more intuitive gameplay
- integrated narrative
That’s why I was intrigued when xKiv was talking about the different reset duration options, such as during the course of 1 session or globally across servers (totally mangling the actual terms, but I’m writing out the gist).
Because if you’re flipping a coin heads or tails, the odds are 50/50 but in an infinite timeline it’s statistically "correct" for the coin to come up heads 1 million times in a row as long as it ends up tails 1 million times somewhere in the tail end of infinity. It just would be a hell of a lot less fun playing with this time table.
And regarding this line: [ [color=rgb(119, 119, 170)]it’s a little presumptive to come right back and say [/color][color=rgb(119, 119, 170)]‘you’re the game dev, but you could be wrong and your work is broken’ ][/color] Yes…? And we are the game players, who like the game enough to discuss it among like-minded fans, both the parts we like and the parts we dislike, with constructive comments. No one is infallible, as far I’m aware?
And the more I think about it, I really don’t think there NEEDS to be people who have such long (good or bad) luck streaks. There could be tweaks made to alleviate the issue that obviously a lot of people are having.[/quote]
Then what you need is to offer Feedback, which you can do here: http://community.failbettergames.com/topic7596-feedback.aspx
However, be aware that what you’re asking for is to have both good and bad luck (both, obviously, for balance and ‘realism’ {hygiene quotes as good/bad luck streaks happen in real life}) streaks tamed. I suspect that without insight into the minds of gamers it would be difficult to tell whether this would upset or please the majority, but if a comparison with gamblers could be made, reducing risk may actually reduce enjoyment. Depends on the type of player is in the majority of FL players, I suspect. Those who’ll have come from board games will be used to evil dice streaks. True, doesn’t mean they’ll tolerate them, though.
I also suspect it would be a deal of work and increase load, as to prevent streaks with certainty you’d have to track each individual’s rolls for x number of rolls each time, and not just compare the number value of each outcome, but also compare to the gameplay result (some results might be ‘lucky’ if low, some high - SN means you can put nasty things on a lucky roll if you wish) to ensure that the positive outcome goes with breaking an unlucky streak and so on.
Do offer feedback, but do so based on the knowledge that the current system is deliberate rather than broken, so it’s clear you’re talking about design rather than a bug - but if they haven’t already looked at how to optimise enjoyment for the majority of their players, they’re fools - and they’ve never come across as such. So I’m not suggesting you hold out much hope! :-)
If you want to continue talking programming/design - do PM the guys in the thread linked to above, or start a thread here: http://community.failbettergames.com/forum11-off-topic-the-surface.aspx but keep it not-FL related/general, as that’s the Off Topic forum - as you’ve already had a response from FB: they are aware of how the game works, and it’s working as designed. They’ve planned the streaks in.
edited by babelfishwars on 4/12/2014
Confirmation bias seems to really enjoy the company of RNGs. Though it is not everyone’s cup of tea, I suggest tracking your results over a very long trial.
RNGs are streaky, it’s true. Good streaks don’t tend to stick in one’s craw, and likewise don’t generate many forum threads.
I should probably resist, but I was invoked by name …
I don’t think I am a better programmer. I might have better understaning of some of the underpinning theorethical issues (that’s largely the essense of what I studied at college), but there’s just too much to know. One person, or even an usefuly-sized team, just can’t know enough to cover anything even marginally comparable to everything known to mankind. No matter how good you are, there’s always somebody who’s better than you at something.
For example, I know effectively nothing about fallen london’s server architecture, what kind of software runs it (and on it) and how that software handles Random.Next (or an equivalent) call when serving a page. And I don’t know that last thing even on the application I am developing right now (where it’s completely irrelevant).
What I know is that without knowing the things in the last paragraph, I can’t guarantee anything.
Hey David, confirmation bias doesn’t come into play here at all. (Assuming you were replying to my comment!) I’m absolutely acknowledging that the RNG isn’t "buggy" and produces the correct stats over a long enough trial. The craw in my cuppa is the "long enough" – how long is long enough? I feel like it’s too long.
The design as it is now works great on narrative advancement choices – Will I get this? Will I fail? What response will I get? It’s exciting. It provides anticipation for the next page in the story.
But it’s absolutely super frustrating when grinding – first, that you have to grind in the first place, and second of all, though you’ve gone through all this effort to raise your skill, it’s still mostly luck-based where you have to accept that yes, you may fail 10 times in a row? And you have 5,000 more items to collect? A lot of my friends (who I successfully got into playing the game!! Was so excited!) have dropped out of the game at this point – "Way too much grinding. Gotten boring. Have better things to do with my time." (Wahhh, including the bf. Goddammit, where’s the loyalty!)
I understand that FG instituted this as a stopgap while they produce more content, but it’s also a block for their more casual audience.
I wouldn’t submit actual feedback to FG as I assume this model works for them. Probably increases Nex purchases from impatient people like me. Increases the amount of "busy work" from actively hooked players ( = less complaints).
Not knowing anything about programming, I’ll still stick my foot in it and say I don’t think you’d have to track every individual’s rolls, but address, again, the span of time the RNG is governing / how often the RNG is resetting itself. Length of time is obviously not infinity because no one has complained that they have spent ALL of their action points on bad rolls. So FG is obviously already in manual control of this.
How often the RNG resets itself may be the first step in figuring this out.
But in any case, here I am again in a niche hobby that no one in RL will join me on. =(
[li]
There have been tweaks to RNG streaks in other games, notably World of Warcraft. It’s not easy, but it’s possible. Relevant link:
http://www.shacknews.com/article/62807/sid-meier-and-rob-pardo
[quote=cinderfallen]Hey David, confirmation bias doesn’t come into play here at all. (Assuming you were replying to my comment
[/quote]
Hi - I was not replying to anybody in particular. I get frustrated when, say, I need a quick 250 candles and 3 actions gets me 112. Happened today. But in a game like FL, I figure pain and agony are a feature, not a bug. I don’t think this is a casual game. It’s a little, outré, no?
edited by mblamar on 4/13/2014
[quote=Theus]There have been tweaks to RNG streaks in other games, notably World of Warcraft. It’s not easy, but it’s possible. Relevant link:
http://www.shacknews.com/article/62807/sid-meier-and-rob-pardo[/quote]
"Purists might dislike Sid Meier changing the match to appease players, but the game became more fun. Blizzard opts to package systems as a reward over penalties and takes a little randomness out of the equation."
[li]Oh, perfectly relevant. Thanks. Basically, highly refined game design. (Really addictive junkfood gameplay / Gotta have more gameplay.)
[/li][li]
edited by cinderfallen on 4/13/2014
While I do share the sentiment, to a point, that the extent to which luck factors into certain actions regardless of progress (e.g. Master Thieves failing low level Shadowy-action-type lucky checks) is a bit counter-intuitive at times, I do tend to heavily favor responses of the kind that the ever-Spockean Babelfishwars has been offering.
There is most certainly nothing wrong with being absurdly lucky or unlucky on the programming end of things—and even if there is something wrong with the RNG program at the theoretical level, the point is still moot as it relates to the topic at hand; since your “good luck” or “ill luck” in reality overlaps 1-to-1 and onto with the results returned by the game, I don’t see how it matters much whether the RNG literally simulates the condition of absolute randomosity in reality. As long as it’s a comparable simulation, the task of the RNG is being carried out effectively.
An additional point: the fact that confirmation bias is in play here should not be debatable: case in point, as has been stated, no one seriously complains about being unusually lucky in practice. And even if you’re feeling inclined to do so, it’s still not a counterargument to the above.
Additionally additionally: I particularly want to condemn this anti-feedback-submission attitude: the argument that the model ‘works for them’, and therefore they’re not likely to change it is exactly the reason you shouldsubmit feedback—if you’ve got something to say about the game, for good or ill, you ought to—this is how things get changed. It’s really the whole point of the feedback system in the first place. Too, the insinuation that they wouldn’t change it because it makes them money, in addition to being fantastically unlikely, seems unnecessarily cynical to the point of being offensive.
A final word: I don’t want to come off as either condescending or brown-nosed here—I’m not in FBG’s corner by default (though if one had to throw their lot in with a company on faith, they probably couldn’t do much better than FBG, given their tremendous commitment to ethical business practice): it does bother me, though, to see dubious arguments given such weight. Not to be antagonistic to xKiv, since he hardly deserves it, but e.g. the point about ‘there’s always someone better’ is literally a self-defeating non-argument. I’m also not sure how the impossibility of human omniscience is relevant to anything.
<3