You need to be logged in to post and to see the uncensored versions of these forums.
|I may have figured out what's wrong with the itembuildr code|
It’s based on average, right? So every item it produces is going to be near the middle. Making the next thing more likely to be near the middle, and so on. That’s why the tattoos all have almost the same stats, and why this lingerie is almost identical to the schoolgirl panties.
I don’t expect ET to respond to this thread in the near term (as he seems to be focusing on the gameplay improvements rather than fixing itembuildr), so maybe we can all brainstorm and figure out how the algorithm SHOULD work, rather than complaining that it doesn’t.
|Posted On: 05/19/2009 3:45PM||View MC Banhammer's Profile | #|
Astute thinking. Maybe something that compares the output to existing items, and rejects those that are within X points of something already available?
|Posted On: 05/19/2009 3:57PM||View AJ_Impy's Profile | #|
Here’s a formula we could use.
20% that it gives PP -> Hardware
20% that it gives Sexiness -> Lingerie/Glbumes/Shoes (33.3% chance each)
20% that it gives Douchebaggery -> Fetishes
20% that it gives Tears -> Piercing/Tattoo (50% chance each)
20% that it only gives Ego
45% on all items that it gives Ego additionally.
It will give Ego based on the formula below. If it only gives Ego, it will go through the formula twice.
This formula will be done once per “PP/Sexiness/Douchebaggery/Tears/Ego”, “Offense”, “Defense”, and “Other stats (Luck, Charisma, et cetera)”.
[(Lowest Offense) + (Highest Offense)]/(Random Variable between 1 and itself/(Lowest Offense)]
*Always round up.
**This is done per item type. So in the case of a lingerie item,
[(the lowest Offense of a lingerie item) + (the highest Offense of a lingerie item)]/[(Random Variable between 1 and itself)/(the lowest Offense of a lingerie item)].
[(Lowest Defense) + (Highest Defense)]/(Random Variable between 1 and itself/(Lowest Defense)]
*Always round up.
**This is done per item type. So in the case of a lingerie item,
[(the lowest Defense of a lingerie item) + (the highest Defense of a lingerie item)]/[(Random Variable between 1 and itself)/(the lowest Defense of a lingerie item)].
PP -> Hacker
It does not give Ego additionally.
PP: [(5) + (100)]/(Random Variable between 1 and 105/5) = Anywhere between 5 and 105 PP.
Offense: [(2) + (15)]/(Random Variable between 1 and 17/2) = Anywhere between 2 and 17 Offense.
Defense: [(1) + (15)]/(Random Variable between 1 and 16/1) = Anywhere between 1 and 16 Defense.
Luck: [(1) + (10)]/(Random Variable between 1 and 11/1) = Anywhere between 1 and 11.-MLF- edited this message on 05/19/2009 4:26PM
|Posted On: 05/19/2009 4:26PM||View -MLF-'s Profile | #|
Good job. Price would be proportional to stats, right? I think the very idea of automatically-generated, random stats for purchasable equipment is bad, though. Maybe itembuildr could feed a more robust item drop system.zagerblag edited this message on 05/19/2009 7:32PM
|Posted On: 05/19/2009 7:32PM||View zagerblag's Profile | #|
That’s the difficult part. The aim of itembuildr is automatically generated stats and prices, but the thing is, all stats are worth different things.
An item that gives 400 douchebaggery would cost less than an item that gives 200 Processing power.
Same thing with luck, charisma, et cetera.
An ideal formula would involve each stat point being worth a certain number of flezz, but something like:
Offense = 1,000
Defense = 1,000
Douchebaggery, Tears, Sexiness = 100
PP = 200
Those are just numbers out of my head though.
Why? Using my method, no item will never be better or worse than the current most powerful/weakest.
In addition, it would be random, unlike the current code.
Though stats could still be blown out of proportion, which shows that the above formula needs a lot more tweaking, to avoid an item that gives hardware with 5PP, 2 offense, 1 defense and 1 luck.
|Posted On: 05/19/2009 8:00PM||View -MLF-'s Profile | #|
Which is what many of us are annoyed with in the first place.
I could care less whether we have a billion items that are worse than any given item I’m using, or only one; I’m not a collector, it makes no difference to me. And that’s the crux of the matter: under your system, itembuildr will at some point become useless to every player in the game. Power creep is power creep, sure, but it’d be nice if I actually had a reason to care how the vote turned out each week.silversparrow edited this message on 05/20/2009 5:51AM
|Posted On: 05/20/2009 5:50AM||View silversparrow's Profile | #|
Yes. This is the crux of the matter. I’ve only been here 2 or 3 months, and I’ve got all the top level equipment; all I need to do is grind for the best moars (which is going to take forever with only a simple spider).
At this stage, I’m all for 2 itemBuildrs; one generating a medium or low level item, the other generating high level stuff. ARun the medium/low as per normal, but at least for the first little while, have the type of high level stuff determined by Evil Trout, based on suggestions from the community. The statistics I’ll leave to my more mathematically inclined comrades.
|Posted On: 05/20/2009 7:36AM||View Sarcasm Inc's Profile | #|
I would prefer to see an actual Build phase that precedes the Description Phase.
During this phase, we actually submit ‘designs’, based upon building points.
There are 4 ‘Energy’ Statistics: Processing Power, Sexiness, Tears & Douchebaggery
There is 1 ‘Health’ Statistic: Ego
There are 5 Attributes: Offense, Defense, Luck, Frugality, Charisma
All items in the game affect one or more of the Attributes or Statistics.
Each new item gets between 1 and 1000 Building Points.
To increase an ‘Energy’ Statistic by +25 costs 1 Building point. For each additional +25 , it costs 1 more Building point than the last level. In other words, the second level costs 2 Building Points, the third level costs 3, and so on. So, to add +200 Tears to an item costs 10 Building Points.
The same applies to Attributes. One Building Point for the first +1, two Building Points for the second +2, and so on.
Ego would increase at a rate of +10, but would be otherwise identical to the Energy Statistic build.
Each type of item would have an upper limit on each Statistic or Attribute, to keep them in line with the existing items. To design an item, the first step is to select an item type (Special Items are excluded from this step, and would have to be manually activate when ET decides it’s time to let one into the system).
For example, if you choose to design a Troll Fetish, then you would have the following Maximum Values for each Attribute/Statistic:
Douchebaggery Max +700, Ego Max +100, Offense Max +15, Defense Max +40, Luck Max +7, Frugality Max +7, Charisma Max +7.
Even with 1000 points, you couldn’t max out all of these values. So, you have to actually design the item.
If we have 500 points to build our new item, we can design a few different possible Fetishes.
Design A: Douchebaggery +400, Ego +40, Offense +15, Defense +21, Luck +2
Design B: Douchebaggery +200, Ego +100, Offense +10, Defense +23, Luck +7, Frugality +7, Charisma +7
To control the overall power level of an item entering the game, all ET has to do in this situation is determine the number of building points to allow for each item. To keep the right balance, it’s simply a matter of weighting each Attribute or Statistic properly. Maybe Frugality needs to have a base cost of 2 Building Points instead of 1. Or maybe Charisma costs more for Fetishes than it does for Lingerie.
The point here really is that being able to actually design an item instead of relying upon a random program is sure to produce more innovative ideas & items. And to me, that’s the whole point of ItemBuildr.
|Posted On: 05/22/2009 11:52PM||View Sergeant Cid's Profile | #|
Also a terrible idea. How long do you think it’ll be before each clbum has an item maxed out in one or even two stats in each category? Even if ET sets the caps low, we’ll get three items tops — Off/Def maxed, Cha/luck maxed, and Frugality maxed — and we’re back to not caring about itembuildr.
Additionally, ET would be figuring out a new cap every other day. It’s not like he has nothing to do, folks.
|Posted On: 05/23/2009 5:26AM||View silversparrow's Profile | #|
That depends on the caps and and building points.
We can already see how itembuildr is generating items at the high end and the low end. Even with this new system, the same could be done by letting people design items with a varying number of builder points (and possibly varying caps). By varying the number of points to force the creation of items for use through the entire level scale, it will take that much longer for us to fill all desired niches.
Will we eventually have created every combination of stats that we really want with some time? Sure. But, we’ll do that regardless of how it’s done. Is it better to have a run of items people want created or is it better to have a huge pile of items people are completely disinterested in because they are more expensive with smaller bonuses than items tht already exist and an occasional good item?
|Posted On: 05/23/2009 10:42AM||View guest42's Profile | #|
I like the idea of a random number of points per week.
|Posted On: 05/23/2009 11:16AM||View Sarcasm Inc's Profile | #|
As much as I hate point-buy systems, this could work incredibly well, given a decent interface for designing items. Furthermore, with the amount of build points being generated every run, you could be forced to make a low-end item or a high-end item, which would be good for players across the board.
|Posted On: 05/23/2009 11:57AM||View Invariel's Profile | #|
Sergeant Cid Posted:
The problem, however, is that any build of an item that doesn’t completely max a stat is going to get downvoted by a majority of the population as “useless”. As mentioned, the winners will inevitably be the ones that spend all the points possible to maximize Offense (and perhaps defense) or Charisma (and perhaps luck) or Frugality. Odds are low that we’d end up with any items the community liked better.
|Posted On: 05/23/2009 11:58AM||View Samildanach's Profile | #|
Why is this a problem? If all of the items were items that the community liked, then the items coming out of itembuildr would all be items that people wanted, as opposed to items that are designed to no apparent purpose and cost more than the simpler items with higher stat modifiers.
|Posted On: 05/23/2009 7:53PM||View guest42's Profile | #|
What does the current algorithm actually look like? It’s not as though we could exploit it, and ET has historically shared code.
|Posted On: 05/23/2009 8:14PM||View markchd's Profile | #|
Because, knowing this, if the amount of points available exceeded or even were equal to existing equipment, it would lead to item inflation. If the points weren’t enough to exceed existing equipment when you just dumped them all in one stat, then people would **** as much about having “useless items” as they already do, except with an extra step in the chain.
In other words, either it leads to escalating EQ values, or the same complaints as we currently get.
|Posted On: 05/24/2009 2:36AM||View Samildanach's Profile | #|
Just take the current formula and add a random factor that either adds up to either +/- 20% in each stat from the original forum, you’ll still get random items without completely breaking the game.
|Posted On: 05/24/2009 4:42AM||View MINION9000's Profile | #|
... until you have the random item chosen from one of the +20% items, creating a 44% stronger item, which will in turn be the basis for a 72% stronger item that can then make a 107% stronger item than we had at the beginning.
|Posted On: 05/24/2009 5:39AM||View Samildanach's Profile | #|
Which is actually in the formula at present at 5%; we just haven’t gotten to its equivalent of 105% yet.
|Posted On: 05/26/2009 2:15AM||View silversparrow's Profile | #|
Do one of the things that will cause the inflation, but put a cap slightly above our current best item (That won’t change)
|Posted On: 05/26/2009 2:43AM||View Lord Shplane's Profile | #|