Tesuji
Posts: 161
1/29/2013
|
With the latest patch, it appears that difficulty levels have changed slightly.
Here is what appears to be the new version:
- 1-10%: almost impossible (+4 CPs, +2 bonus CPs on success)
- 11-30%: high-risk (+3 CPs, +2 bonus CPs on success)
- 31-40%: tough (+2 CPs, +2 bonus CPs on success)
- 41-50%: very chancy (+1 CP, +2 bonus CPs on success)
- 51-60%: chancy (+1 CP, +2 bonus CPs on success)
- 61-70%: modest (+1 CP, +1 bonus CP on success)
- 71-80%: very modest (+1 CP, +1 bonus CP on success)
- 81-90%: low-risk (+1 CP, +1 bonus CP on success)
- 91-100%: straightforward (+1 CP)
This is compared with the old version:
- 10%: almost impossible (+3 CPs, +2 bonus CPs on success)
- 20-30%: high-risk (+3 CPs, +1 bonus CP on success)
- 40-50%: chancy (+2 CPs, +1 bonus CP on success)
- 60-70%: modest (+2 CPs)
- 80-90%: low-risk (+1 CP)
- 100%: straightforward (+1 CP)
This seems both positive and negative. I like that it got rid of the trap option that was the old low-risk (with the new low-risk more-or-less replacing 70% as the new sweet spot, which seems like an improvement).
On the other hand, as someone against whom the RNG appears to have a personal grudge, I will miss the guaranteed 2 CPs per attempt from Modest.
Edit: Updated for new descriptors
edited by Tesuji on 2/26/2013
--
Tesuji.
|
|
|
0
link
|
Diptych
Administrator
Posts: 3493
2/6/2013
|
The Starveling Goat!
The Starveling Goat!
Indistinguishable from a regular goat!
--
Sir Frederick, the Libertarian Esotericist. Lord Hubris, the Bloody Baron.
Juniper Brown, the Ill-Fated Orphan. Esther Ellis-Hall, the Fashionable Fabian.
|
|
|
+6
link
|
Alexis Kennedy
Posts: 1374
2/6/2013
|
Spacemarine9 wrote:
alexis please add all of my suggestions into bideo game to achieve greatness
I'd love to, but you'd sue us, innit
trufact from the Failbetter archives that I may have mentioned before. The Overgoat was conceived when I made a typo back in 2009 while entering a certain early-stage clothing item in the database.
|
|
|
+6
link
|
fareseries
Posts: 16
2/14/2013
|
I'm also incredibly unhappy with the new difficulty system, to the point where it will be a deal-breaker for me if it's not overhauled. It just doesn't scale well to higher difficulties at all, and the problem is only going to get worse as more content is developed and stat caps rise. For clarity, here's how it works (from the reference link). Let's say you have a Watchful challenge. That challenge has a difficulty number - say 20. Under the old system, that meant someone with Watchful 20 would have a 50-50 chance of passing or failing. Each extra point over 20 raises your chance of passing by 10%, to a maximum of 100% at 25. Each point under 20 lowers your chance of failing by 10%, to a minimum of 10% at 16.
Under the new system, there are three differences. The first one is that the base chance of success - with Watchful exactly equal to the difficulty of the test - is 60% rather than 50%. The second one is the important one - that instead of each additional point of Watchful raising your success chance by 10%, it raises your success chance by (60/[difficulty])% - here 3% (assuming a difficulty scaler of 60% which is default). The same goes for Watchful scores less than 20 - each point under 20 reduces your success chance by 3%. The third one is much less important - the minimum success chance is now 1% rather than 10%.
So at Watchful 20, the new system is still somewhat similar to the old one. Let's see what happens with a Watchful test at difficulty 100 - that is, a test where Watchful 100 gives you a 60% chance of passing. Here gaining a point of Watchful makes you only (60/100) = 0.6% more likely to succeed. So to reach a 100% success chance, you need Watchful 167. And someone with Watchful 60 would still have a 36% success chance. Since success chance is displayed only to the nearest 1%, gaining one point of Watchful - something that generally takes 20-50 actions - may have no visible effect on your chances at all. In summary, the problems with the new system are as follows:
- At high levels, gaining stat points becomes almost meaningless. Under the old system, after gaining a point in a stat was always a big deal - it meant that the storylets you were doing became significantly easier. Under the new system, there may literally be no visible difference. It takes a lot of the sense of progress when tests don't become noticeably easier as you level up.
- As a corollary, stat-boosting items become almost worthless. Paying 450 echoes for a Bengal Tigress makes a certain amount of sense when +4 Dangerous translates to a 40% higher chance of success. It makes no sense at all when +4 Dangerous translates to a 1.8% higher chance of success (at difficulty 130).
- Many, many storylets and opportunity cards currently vanish when the relevant stat gets too high (more than 10 points over the unlock threshold seems to be the norm for storylets, about 20 points for cards). When having 10 points more in the stat made the storylet trivial, that made sense. Now it means that (at difficulty 50) it's impossible to ever get the storylet above a 74% success chance.
- Narratively, the game to this point has been built around the idea that a Watchful 130 character can run rings around a Watchful 70 character intellectually. High level characters get to do awesome things that low level characters wouldn't be capable of. A Watchful 130 character has a 60% chance of passing a test with difficulty 130. A Watchful 70 character has a 32% chance. This is not a large difference, and it completely breaks the idea that a Watchful 130 character is significantly brighter than a Watchful 70 one.
- As stat caps get higher, test difficulties will get higher and so stats will have even less impact.
- As other people have pointed out, an inability to be certain of success means far more menaces and far more gear-swapping, both of which are annoying to deal with.
I understand the idea that lower-level content shouldn't be completely trivial for higher-level players, and high-level content shouldn't be completely impossible for lower-level players, and I'm not against it, but this specific implementation is badly broken and tuning it won't help. No matter what you set the difficulty scaling to, as long as you're using the broad system as you've defined it, stats are either going to matter far too much at low levels or far too little at high levels. Fortunately, there are better ways of doing it. I am a mathematician (or at least a PhD student) IRL and will be happy to look at the problem in more detail if you like, but as a first attempt, why not use a bell curve? When someone takes a test, generate a normally distributed random variable X with mean whatever the test difficulty is and standard deviation 10, and say a player passes it if their quality is greater than or equal to X. This would give success chances looking like the following at difficulty 100:
Watchful 70 -> 0.13% chance
Watchful 80 -> 2.28% chance
Watchful 90 -> 15.9% chance
Watchful 95 -> 30.9% chance
Watchful 99 -> 46.0% chance
Watchful 100 -> 50% chance
Watchful 101 -> 54.0% chance
Watchful 105 -> 69.1% chance
Watchful 110 -> 84.1% chance
Watchful 120 -> 97.8% chance
Watchful 130 -> 99.87% chance
And likewise at difficulty 50:
Watchful 20 -> 0.13% chance
...
Watchful 49 -> 46.0% chance
Watchful 50 -> 50% chance
Watchful 51 -> 54.0% chance
...
Watchful 80 -> 99.87% chance
As you can see, the result is that individual stat points matter far more if you're close to the difficulty level than if you're a long way away from it, and stats don't become less important as the game progresses. You can lower the standard deviation to make stat changes close to the difficulty level matter more and raise it to make them matter less - for example, with standard deviation 15 rather than 10, success chances at difficulty 100 become:
Watchful 55 -> 0.13% chance
Watchful 70 -> 2.28% chance
Watchful 85 -> 15.9% chance
Watchful 90 -> 25.3% chance
Watchful 95 -> 37.9% chance
Watchful 99 -> 47.3% chance
Watchful 100 -> 50% chance
Watchful 101 -> 52.7% chance
Watchful 105 -> 63.1% chance
Watchful 110 -> 74.8% chance
Watchful 115 -> 84.1% chance
Watchful 130 -> 97.8% chance
Watchful 145 -> 99.9% chance
This approach would mean that low level players do have a chance at higher-level content, and failure would be a possibility for higher level players, but individual stat points gained or lost would still give a meaningful boost to success chances in high difficulty tests and a player with higher stats is dramatically better - not just slightly better - than someone with lower stats. It would also mitigate the need for gear swapping, since the further someone is over the difficulty level the less individual stat points matter. You'd still need to make low-level storylets stay visible for longer and make either reducing menaces or getting second chances easier, but it would be a dramatic improvement.
If you don't have much mathematical background, this approach might sound complicated, but it's actually very easy (read: a few lines of code) to do using standard library functions. All you need is the ability to generate random numbers, which you already do, and the ability to calculate erf, which almost any language has as standard in its libraries. (And if it doesn't it's not that hard to write your own.) I'll be very happy to give more details if you'd like. I'd also be happy to work up something different if you feel this approach leaves something to be desired - personally I'd rather have stats matter even more when close to the test difficulty without quite such a sharp drop-off as you get further away. I enjoy the game a great deal, I just really do feel this difficulty scaling is disastrous as it stands.
God I hope someone reads this. 
|
|
|
+5
link
|
Tesuji
Posts: 161
2/14/2013
|
San wrote:
Not sure I'm entirely thrilled about the switch to Broad difficulty (I'll have to look up how that works). It will definitely lead to me doing a lot more min-maxing when it comes to my equipment and qualities, which can get annoying sometimes.
Under the old system, you had a storylet of Difficulty Level X. If your quality was X, you had a 50% chance for success; at X+5, you had a 100% chance for success; at X-4 or less, you had a 10% chance for success.
Now, the chance for success is (Your Quality)/(Difficulty Level)*0.70
This means that under the old system, you needed Difficulty Level + 5 for a guaranteed success. Now, you need Difficulty Level / 0.7.
For example, a DL 130 opportunity card required 135 for 100% chance of success before, and now requires 185 to avoid any chance of failure, though the rewards haven't changed.
I think there are some obvious issues here, as there are a lot of storylets that seem to have the assumption that, unless you're a gambler, you just flat-out shouldn't be playing them until you hit the 100% success point: multi-action storylets, carousels that give half-rewards if you fail the final storylet, storylets that have significant negative consequences on failure.
Some stories really aren't worthwhile to play for material gain any longer, because they were balanced under the idea that you need to gain X of something in N actions (e.g., Newspaper, Pickpocketing, Polythreme). Throwing in random failures now adds an element outside your control.
Basically, this change ramps up the randomness across the board, except for (ironically) luck challenges. I can see the value in this sort of system for a game designed around that paradigm, but it's a huge shift for a game designed around different assumptions.
Another issue is that, as has been noted, this strongly incentivizes micromanaging equipment each time you switch to making checks with a different stat, despite equipment management being something the game does particularly poorly.
I'm okay with the idea of failing because my character just isn't good enough at something, but failing because it's too much of a hassle to switch back and forth while playing a dual-stat storyline doesn't really strike me as fun.
edited by Tesuji on 2/14/2013
--
Tesuji.
|
|
|
+3
link
|
Diptych
Administrator
Posts: 3493
2/26/2013
|
Can we tell the difference between Modest and Very Modest from the necklines?
--
Sir Frederick, the Libertarian Esotericist. Lord Hubris, the Bloody Baron.
Juniper Brown, the Ill-Fated Orphan. Esther Ellis-Hall, the Fashionable Fabian.
|
|
|
+3
link
|
Urthdigger
Posts: 939
2/27/2013
|
As a fan of social actions (Including garnering second-chances), this makes me sad to hear.
--
Looking for second chances to maximize your loot output from those troublesome storylets? Check out our handy gang of volunteers in this thread, or even volunteer yourself!
@Urthdigger on twitter
|
|
|
+3
link
|
Alexis Kennedy
Posts: 1374
2/15/2013
|
Hi all
I can't respond point by point, but here's a giant post to cover the basics.
Firstly, we're in transition. The situation now is not what it will be in a week or six months. We'll be
- making second chance use optional
- adding new items
- adding advanced Professions and other character-build choices
- increasing the bonuses of existing items
- or allowing upgrade options as with the ravens
- changing existing content
- adding new content
- introducing higher-tier rewards.
And this will go on for as long as FL continues - years, I hope. This change gives us room to grow. It allows us to make rewards meaningful again, it allows us to extend content lifespan, it allows us to build minisystems like the Heist content without resorting to Luck checks. The content lifespan in particular is a really big deal, given my own sharply limited resource and the near-infinite appetite for content.
Second, thank you for your feedback, and not just in a passive-aggressive fsck-you corporate way. It's not a surprise that some people are upset - that always happens with rebalances (some people may remember the house catching fire when we introduced variable cp rewards for challenges way back when). But it is useful to realise that we may need to tweak the challenge band descriptions so people can see the needle move better, and it is useful for people to point out bits of content where we may not have accounted for consequences (so e.g. now I have to go check on the Boatman).
Thirdly, the mechanic itself. It has to be intuitively comprehensible to people who aren't especially mathematically sophisticated, because those people are not only most of our audience, but especially most of our StoryNexus creators. There are people who are both fine writers and good mathematicians, but they're thin on the ground, and SN was intended to support creators who are primarily writers, not coders. Honestly, this one pushes the envelope on that requirement as it is.
Some other points.
"Many, many storylets and opportunity cards currently vanish when the relevant stat gets too high" <-- You may have noticed we're gradually removing this: as we rework content, it'll gradually disappear. It's a leftover from a bygone age, and it was never the best solution.
Default difficulty scaler: actually, to ease transition pain, this is set higher than normal, so everyone's chances of success are better than 60% at parity right now.
"It's a step in the FarmVille direction, focusing on the players as a source of income, rather than a source of loyal devotion and artistic feedback--or to put it another way, it's a step toward making money rather than making story." (1) You know I think someone has accused us of being Zynga, and called us crassly commercial, every time we've made a major change. And yet here we are, four years on, less like Farmville than ever, and still a tiny indie company with, I believe, the largest quantity of free content in terms of number of words available for, perhaps, any game ever. (2) Loyal devotion plus artistic feedback plus £2.50 will get you a cup of coffee. If I had relied on loyal devotion and artistic feedback, the game would have ceased to exist c. March 2010. I am glad and grateful that people love our content enough to pay for it, and I'm not embarrassed about asking people to pay.
"there are a lot of storylets that seem to have the assumption that, unless you're a gambler, you just flat-out shouldn't be playing them until you hit the 100% success point." If you mean by 'seem to have the assumption' that this was our assumption when writing content, no this isn't the case. Particularly risk-averse players (who are, obviously, the people upset by this) do the sums and do that, but the most popular choice of balance on the way up is to start repeated action at a 60-70% chance of success.
"I really, REALLY hope this doesn't apply to challenges using other qualities, like casing..." No, it doesn't, and I have no plans to make that happen.
"Narratively, the game to this point has been built around the idea that a Watchful 130 character can run rings around a Watchful 70 character intellectually. " As the guy who built the game, I have to say, this was not what I built it around, narratively. :-) It's just an unfortunate artefact of an early design decision when I never knew if the game would last a whole year. Fallen London characters are flawed legends who achieve wonders but can always fail; an FL veteran is scarred, canny and experienced, but they can never relax.
--
tl;dr: Trust us a little longer. We've been at this for almost four years now and we never, ever do a mechanics change lightly - we don't like upsetting players, but even without that, it takes substantial time and effort away from other things. This is short-term pain for long-term gain.
EDIT: moved a misplaced sentence and acted to soothe the bad-words filter.
edited by Alexis on 2/15/2013
|
|
|
+3
link
|
Alexis Kennedy
Posts: 1374
2/6/2013
|
So, seriously, a lot of this shuffling about is in preparation for the tremendous impact which is the incorporation of Fallen London into the main StoryNexus UI. If you've seen Melancholia, it'll be sort of like that. But after that, it'll be much easier to tidy up everything at once (and roll out our currently embryonic responsive mobile design).
The first 10 people to click this link will receive an Undergoat: http://bit.ly/UndergoatAccessCode
|
|
|
+3
link
|
Alexis Kennedy
Posts: 1374
2/6/2013
|
Saharan wrote:
what's with the -3 Persuasive on the Starveling Cats? Were they intended to be equippable Companions once upon a time?
It's a gag. The Cat is not even useless.
|
|
|
+3
link
|
Spacemarine9
Posts: 2234
2/7/2013
|
My issue is that not enough UI features can be led in dance in this manner.
man idk what the deal is with these new boxes they seem pretty energetic. They're just so excited! http://youtu.be/J2UV9AL5aP8
edited by Spacemarine9 on 2/7/2013
--
my rats will blot out the sun
Ratgames
FL lore/mechanics questions and answers
#FallenLondon IRC (irc.synirc.net) Channel! Click to join via Mibbit.
#SunlessSea IRC channel! Like the above, but zee-ier.
|
|
|
+3
link
|
Spacemarine9
Posts: 2234
2/6/2013
|
I would also like to point out that my inventory does not even have any semblance of being ordered by rows any more.
maybe we should just make a new thread "Yell Things At Alexis And Use Rats As A Means Of Bribery" is a nice snappy title, I think.
Otherwise he'll probably put on his an emptiness jumper and edit all our posts to say "i love ui. i love box"
edited by Spacemarine9 on 2/6/2013
--
my rats will blot out the sun
Ratgames
FL lore/mechanics questions and answers
#FallenLondon IRC (irc.synirc.net) Channel! Click to join via Mibbit.
#SunlessSea IRC channel! Like the above, but zee-ier.
|
|
|
+2
link
|
Nigel Overstreet
Posts: 1220
2/7/2013
|
philmorley wrote:
I agree, it's like having a neon sign in the corner. Also seems very 'modern' styling for old london
I still miss the big, red "Go" button.
I wholeheartedly agree with Woogawoman. Contrast the new FL screen with the mantelpiece screen. Everything about the mantelpiece screams Victorian horror mystery.
The newer UI seems to say "Insert your text into this generic template." I understand they want to make the UI of Storynexus games similar, but where is the virtue in that? Isn't having all of your games look alike a drawback rather than a feature? Especially when so many SN games are set out of our modern times and the new display seems not just modern, but video game like. It makes me very aware I'm playing a game rather than reading an interactive story.
Moreover, I don't see the point of the change. Were people complaining about the look of this or any game? Was there a flood of hate mail about FL not looking enough like other SN games? Am I missing the pages upon pages of suggestions on the feedback forums asking Failbetter to change the challange italic font?
I'm not opposed to change, but I'd rather the interface look more like the mantelpiece screen or have a similar Victorian aesthetic. It seems the most recent changes have been made for the sake of change rather than to fix a problem.
--
The Romantic Egotist: Most Hedonistic Man in All of Fallen London
Are you or someone you know Overgoated? Please, let me know!
Cider Club
|
|
|
+2
link
|
Tesuji
Posts: 161
2/6/2013
|
Not a fan of the latest format change to the way difficulty levels are shown.
I thought the previous way (with the percentage change appearing in the alt text) worked. I find the new version, with its dark box and expanding/contracting text to be garish, obtrusive and distracting.
--
Tesuji.
|
|
|
+2
link
|
Sara Hysaro
Moderator
Posts: 4514
2/26/2013
|
Nice side effect of optional Second Chances - no more failure lag. Unless you happen to like frowning at a loading screen, in which case it's an unfortunate side effect. I think I might take a little while to recover from my conditioned, "Yes, success!" when there is no lag after taking an action. XD
--
http://fallenlondon.storynexus.com/Profile/Sara%20Hysaro
Please do not send SMEN, cat boxes, or Affluent Reporter requests. All other social actions are welcome.
Are you a Scarlet Saint? Send a message my way to be added to the list.
|
|
|
+2
link
|
Guy Scrum
Posts: 197
2/26/2013
|
Because I'm kind of a nerd, here's the math for expected echoes:
Let R be the reward for a successful action, and p be the probability of success (without second chances). I'm going to assume that R*p is more or less a constant for a given character level, regardless of challenge level (that is, R is proportional to the challenge level). Let M be the number of actions it takes to reduce the menace from an action, and let S be the number of actions it takes to acquire a second chance.
Without a second chance, we have:
Expected reward = R*p
Expected actions = 1 + (1-p) * M
With a second chance under the new system (using up the second chance even on success), we have:
Expected reward = R*p * [1 + (1-p)]
Expected actions = 1 + S + M*(1-p)^2
With a second chance under the old system:
Expected reward = R*p * [1 + (1-p)]
Expected actions = 1 + S*(1-p) + M*(1-p)^2
Plugging in some reasonable values for M and S (M = 1, S = 0.5), the rewards per action tops out at:
R*p at p = 1 for no second chances
about 1.05 R*p at p = 0.77 for the old system of second chances
about 0.85 R*p at p = 0.40 for the new system of second chances
Under the new system, the most lucrative thing to do is to grind away at straightforward challenges, although this changes if you make M or S very small. For the values of M and S that I picked, you'll actually get a lower return on echos per action if you use second chances when the probability is higher than about 0.75. For tougher challenges, it makes sense to use second-chances.
Make of all that what you will. Not really sure what the gameplay implications are.
--
http://fallenlondon.storynexus.com/Profile/Guy%20Scrum
Interactive fidgeting writer simulation
|
|
|
+2
link
|
Spacemarine9
Posts: 2234
2/6/2013
|
Remember when this topic was about difficulty levels? I don't!
i much prefer goatcoat chat rather than ui design chat anyway
i just hope the FL storynexus integration doesn't end up making me die out of fear of change, though. That would be grisly.
--
my rats will blot out the sun
Ratgames
FL lore/mechanics questions and answers
#FallenLondon IRC (irc.synirc.net) Channel! Click to join via Mibbit.
#SunlessSea IRC channel! Like the above, but zee-ier.
|
|
|
+2
link
|
Tesuji
Posts: 161
2/15/2013
|
fareseries wrote:
Why not use a bell curve? When someone takes a test, generate a normally distributed random variable X with mean whatever the test difficulty is and standard deviation 10, and say a player passes it if their quality is greater than or equal to X.
A bell curve has its advantages, but I'm not even sure that the best distribution for this sort of thing is a symmetrical one.
Take, for example, a DL 130 storylet. Under the old system, this storylet is largely irrelevant to anyone with a modified stat < 130 (if we want to be generous, maybe 128). There's also no chance of failure with a modified stat of 135+.
For a DL 130 storylet, the new system takes the 50% to 60% chance that used to cover 130-131 and stretches it over 93-130 (a stat of 55 is now equivalent to an old-system 128 for this storylet). On the other hand, you now need to reach 185 to get the same chance that a 135 stat character used to have.
I can see how the first issue is a problem, since it means that content for higher-level players is irrelevant to lower-level ones. Stretching the 0 to 60% curve over more levels increases the potential audience, which is probably a good thing. (To avoid creating other problems, you need to add in floors to unlock content, but the game largely already has them in place, so not much extra work.)
I'm not sure the second issue is a problem, or that "solving" it makes the game better. I guess this comes down to designer intent. If the designers think that a significant problem with the game is that players don't suffer enough random failures and that there aren't enough incentives to micromanage equipment, well, then I guess this takes care of that problem. That's not my perception, but then it's also not my game.
Personally, I think the better solution would be to use a hybrid system: If Character's Quality is less than the Difficulty Level, have a slow increase from 0 to 60% over X levels; if Character's Quality is greater than the Difficulty Level, have a rapid increase from 60 to 100% over Y levels. Seems like the best of both worlds to me.
--
Tesuji.
|
|
|
+2
link
|
Sara Hysaro
Moderator
Posts: 4514
3/30/2013
|
Maybe the color of Straightforward could change? Use the color it is currently for stuff that isn't 100% (but in the range of Straightforward), and a different shade of green if it is.
--
http://fallenlondon.storynexus.com/Profile/Sara%20Hysaro
Please do not send SMEN, cat boxes, or Affluent Reporter requests. All other social actions are welcome.
Are you a Scarlet Saint? Send a message my way to be added to the list.
|
|
|
+2
link
|
Zeedee
Posts: 276
4/7/2013
|
Public Service Announcement!
Folks, Broad calculators are out now. Get one, two or all three here for the bargain price of 1 click:
http://fallenlondon.wikia.com/wiki/Broad_difficulty
There is also a redirect page which is somewhat easier to remember:
http://fallenlondon.wikia.com/wiki/Broad
They were made by Esterhazy.
--
Please do not send me monstrous invitations tinged with the inks of the undernight or Boxed Cats. (I rotate my Starveling list, so it might take me a while to reach your name. I haven't forgotten anyone!)
|
|
|
+2
link
|
Alexis Kennedy
Posts: 1374
2/26/2013
|
Just popping in to say
- Modest has been broken into Very Modest and Modest, Chancy into Chancy and Very Chancy. This is a direct response to the earlier feedback on this thread about how it was harder to see the effects of changes in the middle ranges. We may tweak further, but we're going to let it settle for a good while.
- The optional Second Chances functionality was a key reason for the rather striking difficulty block that some people didn't like: we wanted to make space in the UI, and we wanted to draw attention from people who might not otherwise notice the change.
- Watch the skies for more changes.
|
|
|
+2
link
|
Sara Hysaro
Moderator
Posts: 4514
2/27/2013
|
rykarmalkus wrote:
Best example is a friend I currently recruited to the game. He's not going to see PoSI content for... god knows how long. Months, at least. Though the training professions WILL help a good bit, I imagine. Though, with the amount of story that's already been created, it seems impossible (in this current state of affairs) that he'll ever "catch up" to the end-game content, except maybe in one quality. At best.
Is this really so bad? It's really fun having content left to do, and holidays aside it isn't as it'll be going anywhere. He'll get to everything eventually, and it'll be as fun then as it was when it was released. If anyone wants a faster rate of leveling they can always take advantage of the time-delay stat boosts. I think Dangerous is the only one unavailable to higher level players.
--
http://fallenlondon.storynexus.com/Profile/Sara%20Hysaro
Please do not send SMEN, cat boxes, or Affluent Reporter requests. All other social actions are welcome.
Are you a Scarlet Saint? Send a message my way to be added to the list.
|
|
|
+1
link
|
fareseries
Posts: 16
2/15/2013
|
Thank you for the detailed response - a few points:
- Making items dramatically more powerful and adding advanced professions (which would presumably have some form of stat training) will somewhat mitigate the problem of progress. But ultimately you'll still run into the same problem - unless you're really shaking things up, the primary form of advancement will still be through experience. And advancement through experience will always have the same problem - at high levels, it takes 25-50 actions to gain a level in a quality. As the game goes on, test difficulties get higher, and a single level becomes progressively less meaningful, that fact doesn't change at all. So by level 100, where under Narrow you had to put in 25-50 actions to get an extra point in the quality and a +10% chance of success boost, under Broad you need to put in 425-850 actions to get an extra 17 points in the quality for that same 10% boost! Even if you are shaking things up that much, so that the primary source of advancement does become new gear, it would be a huge change in tone narratively - from doing awesome things because you've learned and grown to doing awesome things because you bought some awesome gear.
- As for higher-level characters versus lower-level characters, I wasn't trying to say that high-level character shouldn't be able to fail at low-level tasks. What I was trying to get at was this: a character with Persuasive 60 has only just entered high society. A character with Persuasive 110 may feel that they have "done" the Empress' Court. Narratively, the former character might be capable of doing anything the latter character can do given sufficient luck, but there's a vast gulf of experience separating them - it might be possible, but it certainly shouldn't be likely. So when the mechanical difference between the two, in a difficulty 100 skill check, comes down to one having a 36% chance at passing to the other's 66% chance, it feels incredibly wrong.
- If mathematical sophistication is a concern, you could get a similar effect without using any scary phrases like "normal distribution" by giving people the ability to patch Narrow distributions together. Something like this: if your quality is equal to the difficulty, you have a 50% success chance. Each additional point gives you 5% extra success chance until you hit 90%. From there, each additional point gives you 0.5% extra until you hit 100. That would give you a clear 56 points of difference between 0% and 100% success chance, while still making individual stat increases meaningful. I'd even call it simpler than Broad difficulty in terms of visualising how it works. (In any case, I'm not advocating removing Broad difficulty from StoryNexus altogether - only introducing a new Focused difficulty for use with Fallen London.)
- I genuinely don't think this can be compared to past design-change drama. The worst case when introducing variable cp rewards was accidentally turning the game into a grindfest (which didn't happen). This, on the other hand, is gameplay that impacts on story. Not the text of the storylets, but the entire wider story of a character over months or years of gameplay. Think about the Seeking Mr. Eaten's Name content - there's a reason so few people keep pursuing it, and that reason is that people value their characters. They don't want their stories to end with... whatever happens to people who find the Name, and that's because they care about the wider story. And the worst case scenario for Broad difficulty is that it could change that.
|
|
|
+1
link
|
Corran
Posts: 401
2/16/2013
|
I have to say that I don't like these changes either.
After reading Alexis' post I'm curious to see what will come in the near future but personally I'm much happier with a system where leveling gives more of a boost and where you can wait till you get a 100% success chance.
I'm no fan of gear swapping (like most here, it seems) and the lag when clicking an item doesn't help either.
edited by Corran on 2/16/2013
--
My Fallen London profile
|
|
|
+1
link
|
Fhoenix
Posts: 602
2/16/2013
|
Ah, I feel like this is not going to end well. Alexis is trying to tailor his game to a majority and get as many people to play for as long as he can. I get his reasons, I understand his reasoning. Problem is, I am not part of said majority. Nor are the other posters here part of the majority, judging by their replies.
No matter how you look at it, this change is a nerf. It breaks the silent agreement that a character with a 100 quality can generate income at 1 echo per action (on average and when taking challenges of appropriate difficulty), a character with 150 will generate 1.5 and so on. We will get less now. Whoever grinded in the past are lucky, new players are not.
More importantly to all of us, we could live with a nerf, I think, but we are very much bothered with the sheer amount of newly introduced randomness.
Particularly this snippet botheres me.
Alexis Kennedy wrote:
Particularly risk-averse players (who are, obviously, the people upset by this) do the sums and do that, but the most popular choice of balance on the way up is to start repeated action at a 60-70% chance of success.
Okay, so for people who took chances before, everything is fine. What is a risk-averse player who can do his sums supposed to do? I know that with a 30-40% chance I will not only waste an action, I am going to generate menace. I am fine with getting it in small quantities, but I am not going to do repeated actions that generate menace. Because that does not make sense in the long run.
Also Alexis, can you adress the problem with frequent item swapping? Because a lot of people mentioned this, but it was missing in your reply.
I just tested it, and swapping a whole set of items to max one of my stats takes almost a minute of my time (time to load the page, time to find the correct items in the ever growing list of items, time to reload the page, time to get back to the action I was supposed to do in the first place). That's on a PC. On a phone I fear it will take much more. Is FB planning to do anything about this? You just said we are going to get even more items.
I mean ideally, I'd like either to have 4 templates, that I can feel in myself and swap at a press of a button, or have the item swap done automatically by the game for me. This is a very simple min-maxing problem, that a computer can solve in a blink of an eye. I understand this is not going to happen for the reasons of immersion. But immersion can only go so far. At his point swapping items is becoming a real chore.
edited by Fhoenix on 2/16/2013
edited by Fhoenix on 2/16/2013
--
My Character
|
|
|
+1
link
|
MittenO
Posts: 9
2/16/2013
|
I understand what you are trying to do here, and it is a good idea, but you really should be using the normal distribution.
It would be fairly simple to implement, would make increase in stats meaningful and still keep content relevant for a longer time.
In fact it would give you a single parameter (the variance) that would allow you to tune the relevance width elegantly and arbitrarily.
It will give you everything you want.
As for "It has to be intuitively comprehensible"
The normal distribution IS intuitively comprehensible. Sure, the formula might make little sense to most people,
but it is one of the most common and most natural probability distributions in the world. Every human being has an intuitive feel
for it. They may not have a name for it, but they do know it. Any other more "comprehensible" distribution, except the uniform one,
will be less intuitive. People don't usually comprehend things through their mathematical properties, so having a lot of them
will not in itself make it harder to grok.
And anyway you can just show somebody a single picture and they will see what's going on,
to a sufficient extent. Hmm.. log-normal might work well too, to give you a bit of unsymmetry.
About that one I know nothing, but a single picture in wikipedia has made me see it would work nicely
for making a challenge get easier faster before the half-point, than after.
(Might be a tad harder to program, though)
Even if you really feel that you must keep it simpler than this, please get rid of the effect that
skill points mean less and less the higher you get. That is just bad.
|
|
|
+1
link
|
Joscelin
Posts: 18
2/16/2013
|
My biggest issue is still the diminishing returns. Stat growth is linear past 50 and echos per stat in the bazaar grow exponentially. That was fine when difficulty was linear, but now it's not. Doing the math on difficulty is pretty silly now. It's a hyperbolic graph; as difficulty rises the value of one stat point approaches zero. With new items you can band-aid it through 200, maybe 300, but in the long run its going to get silly. Someone who just got PoSI is already at the old minimum success chance for a 600 difficulty check. It will takes them years to get the next 900 points to make that 100%. We're not nearly there yet, but it worries me when design gets weird at the edges, especially edges we're slowly moving toward.
|
|
|
+1
link
|
Joscelin
Posts: 18
2/16/2013
|
What tweaks are people hoping for, within the current system? For me, the ideal endgame would be something where the cap on difficulty is equal to the stat cap (seems reasonable) but with enough work you can boost the odds to 90-100%. If we assume a cap of 200, that means 100-133 points worth of stat boosting items. It would allow you to get to a fully straightforward endgame if you wished, but keep it non-trivial for anyone without maxed items, and you could take them off to tune it yourself. I kind of doubt we'll see that much of a boost, though. And it puts a full half of your character's potential into their stuff and not themselves. :-/
|
|
|
+1
link
|
Alexis Kennedy
Posts: 1374
2/18/2013
|
>Another consequence which I just realized: ignoring menaces, all actions now have almost exactly the same echo expectation value. The expectation value is E = (reward value)*(probability) = (reward value)*(char level)/(challenge level) * 0.6. If the reward value is pretty much the same as...
>Except if you use second chances. Then the high-level actions will always be more valuable (if not taking in the cost of the chance itself.)
Ah, I meant to address this too.
Yes: this is part of what I mean by 'keeping content relevant'. I want Fallen London to feel like a city that you grow to know full of places that you return to, and under the old system only the most diehard non-optimising roleplayer would return to (e.g) Wolfstack Docks once they're past the relevant Dangerous band. I want to make it supportable to go back there if you decide you prefer the header illo and the ambience, or you're looking for the specific rewards you find there.
But, yes, use of second chances makes the option of higher-level activities more desirable (which is why optional second chances are right around the corner). Higher-level activities also give some items you may not be able to get at lower levels. So now there's another currency implicated that you get in a different way.
The change also makes more sense of something we attempted early on and realised was a waste of time: alternate stat branches on primary content. Sure, you could try a Watchful 70 test instead of a Shadowy 70, but you're in the middle of grinding your Shadowy, not your Watchful, so your Watchful is either far too low or so high that using it is a non-decision (unless you're desperate for Shadowy CP). If there are more incremental differences between them, it becomes supportable (nb supportable, not optimal) to try out a different route if it makes more sense in RP or flavour-seeking terms.
|
|
|
+1
link
|
Sara Hysaro
Moderator
Posts: 4514
2/19/2013
|
I think that those who have put a lot of work and/or real money into the game will be unlikely to leave even if they are upset with the change. At least, not permanently. At some point down the line when they are no longer angry they will think about the game again, and return to see how things have changed. Chances are by the time they've returned from their irate departure there will be new content to distract them from the things they took issue with, and the metaphorical house will have its roof. Other players more like me will grumble a bit while continuing to play as normal, and get used to the rain.
--
http://fallenlondon.storynexus.com/Profile/Sara%20Hysaro
Please do not send SMEN, cat boxes, or Affluent Reporter requests. All other social actions are welcome.
Are you a Scarlet Saint? Send a message my way to be added to the list.
|
|
|
+1
link
|
lady ciel
Posts: 2548
2/6/2013
|
That dark green box really detracts from the other text. Not good!
--
ciel
Sorry RL means I am not a very active player at the moment. No social actions unless you are prepared to wait and definitely no sparring or other mult-action things.
No Calling Cards or boxed cats please. Will take dupes on the affluent photographers. Other social invitations welcome. Parabolan Kittens usually available, send me an in-game social action saying you want one and I will get one to you as soon as possible.
storynexus name - reveurciel
|
|
|
+1
link
|
Fhoenix
Posts: 602
2/6/2013
|
The fanbase never forgets! The fanbase never forgives! Fear the Unpleasable fanbase!
Anybody remembers that feature, that allowed us to browse through actions even at zero candles? That was so cool. Those were the days.
--
My Character
|
|
|
+1
link
|
Diptych
Administrator
Posts: 3493
2/6/2013
|
Wow! An Undergoat of my very own! Wait, if my Undergoat meets an Overgoat, will I end up with just a Goat?
--
Sir Frederick, the Libertarian Esotericist. Lord Hubris, the Bloody Baron.
Juniper Brown, the Ill-Fated Orphan. Esther Ellis-Hall, the Fashionable Fabian.
|
|
|
+1
link
|
Lady Red
Posts: 517
2/6/2013
|
I hate you, Alexis.
Sir Fred - I expect that if an undergoat meets an overgoat, baby goats may be the result.
--
http://fallenlondon.storynexus.com/Profile/Lady~Red
|
|
|
+1
link
|
Lady Red
Posts: 517
2/6/2013
|
In light of this revelation, I demand an overcoat for my overgoat.
Also, an evil chicken would not go amiss.
--
http://fallenlondon.storynexus.com/Profile/Lady~Red
|
|
|
+1
link
|
Fhoenix
Posts: 602
2/7/2013
|
Having the same UI makes it easier to maintain and improve it. So it's probably done for FB convinience, not because people are complaining.
You do raise a valid point. If every game looks the same, they all lose something in flavor and atmosphere. And people will complain about any interface change, because they are used to the old design and don't want to relearn. So the change should better be functional and actually add something to the game, not just make the fanbase irritable. So far, I think the consensus in this thread is, that showing probabilities on mouse over was a nice change, but adding this big black box was not.
--
My Character
|
|
|
+1
link
|
Lady Red
Posts: 517
2/7/2013
|
Is your issue that you have too much time on your hands?
--
http://fallenlondon.storynexus.com/Profile/Lady~Red
|
|
|
+1
link
|
Woogawoman
Posts: 80
2/6/2013
|
As a UI designer/interaction designer, the thing that struck me the most about the new difficulties styling is that it really screams "I'm the most important thing on this page!" More important than the storylet and option titles. Than the action buttons. Than the actual text that is the reason we play FL.
It's a general degregation over time of the visual information hierarchy of the screens, to the point that it may start impacting both brand and usability.
That said, we know that change is inevitable. I would just prefer to feel that adequate thought had been given to fitting all these changes into a consistent aesthetic.
--
----------
Eleanor Smythe's Profile
|
|
|
+1
link
|
Alexis Kennedy
Posts: 1374
2/14/2013
|
Aximillio wrote:
With second chances, the chances are *much* greater, since you get 2 attempts.
Sneak preview: we'll finally be making second chances optional in the near future. (The UI requirements for this are the main reason for the difficulty block UI tweak.)
|
|
|
+1
link
|
Fhoenix
Posts: 602
2/14/2013
|
Alexis Kennedy wrote:
Fhoenix wrote:
Great game mechanic here.
edited by Fhoenix on 2/14/2013
Thanks! For a moment I thought you hadn't read the forum guidelines and were being rude and negative, but I'm really glad you like it.
Sorry if that came off as rude, I apologise. I am negative here, yes. If what I am saying is somehow against the rules, do tell. If you want some positive thoughts, please check the Feast thread, there I am all smiles.
--
My Character
|
|
|
+1
link
|
Ben Cardwell
Posts: 27
2/14/2013
|
Fhoenix wrote:
Catherine Raymond wrote:
I've been assuming that it's incomplete--that they I've been trying to craft a box like the ones for quality checks, and it hasn't quite come off yet.
No, it's not that it is incomplete, it really looks like a bug. The Investigating the Big Rat card now looks like this
 I've been having this same problem with a lot of storylets since day before yesterday. Most inconvenient, I must say.
|
|
|
+1
link
|