Tobold's Blog
Thursday, March 10, 2011
Discovery crafting in Guild Wars 2

A reader alerted me to a description of crafting for Guild Wars 2:
When the correct items for crafting an item are added to the interface, the resulting item can be crafted. If you haven’t previously crafted that item, you discover the recipe for that item, allowing you to easily view the correct combination to recreate the item. Some basic recipes are automatically learned by characters, but the recipes for most items must be discovered by the crafter.
We decided to go with a discovery system for learning recipes so as to allow crafters to distinguish themselves. When there is a static list of recipes learned from a trainer, every character with that crafting discipline is the same. However, with a discovery system, players that spend time and effort on their crafting disciplines have a way to distinguish themselves.
Unbeknownst to ArenaNet somebody already developed a sneaky technology which is designed to thwart this sort of game design. It is called "the internet". The kind of puzzle that is finding all valid combinations of a limited number of resources is best solved by brute force, and is done fastest by crowdsourcing, many players working together on a Wiki. Before Guild Wars 2 is even released, all possible recipes will be already available on some such site. Thus every character with that crafting discipline will be the same. Objective failed!

So what could ArenaNet do instead? One system that can't be killed by some internet database is recipes being discovered randomly while crafting other items. Every time you craft an item of level n, you have an X% chance to randomly discover a recipe for an item of level n+1, and so on. If you want every crafter to be unique, you have a cap of how many recipes of every level can be discovered, while having at least twice that many recipes in the game. Thus every crafter has a different random set of recipes in the end. Of course that means that some people will be luckier than others, and get more desirable recipes, but that is the price you pay for not every crafter being the same.

Theoretically one could make a game in which random recipes that are different for every crafter are combined with the kind of combination style puzzle ArenaNet is planning. Two people using the same combination would get different results. But given a limited number of recipes that would also mean that two people crafting the same item would use different resources, and thus have different costs. Not a good basis for a player-run economy.

The game which solved crafter differentiation by far the best is Star Wars Galaxies. Everybody had the same recipes, but the quality of the items produced depended on the quality of the resources used. The difficulty was thus in finding the most high quality resources, the location of which randomly changed every week.

But sorry ArenaNet, a system which is based on combination puzzles to uncover hidden information doesn't work well when players can exchange that hidden information. You have to do better than that.
I was toying with a similar system to this for a project I've been designing (with no serious plans of implementing, despite everything I say to the contrary.) In my case, the recipes themselves would reset every month or so, wiping all crafters' knowledge clean, making quick, early discovery and properly protecting your secret recipes the path to having valuable/powerful crafting.

This system likely would not fit well with a persistent world, however, as players with a particularly nice combination would mass-produce it before the skill wipe, so they could sell immediately on day one of the reset. Scratch that, any and every combination would be stockpiled beforehand. In my case, I already planned to reset player items periodically, resetting the crafting came later.
Dev Comment on Guru:
"Our goal is for the crafting system to be enjoyable and rewarding for players that enjoy crafting, without requiring players who don’t like it to have to spend a lot of time crafting, or feel that they need to learn a crafting discipline. These external resources can be used by players to bypass the discovery system, but if the player doesn’t actually enjoy crafting or discovering recipes on their own, we feel that this is a good thing."

so, yes, the system is Horadric Cube style and so - thotbottable. i'm not sure i like that or subscribe to their reasoning (although it's in the general spirit of the game - a next-gen MMO-lite)

personally i enjoyed the minecraft crafting, although i used the wiki for the recipes. lets wait and see for this one, but i'm not looking forward to it too much
They already did this in Final Fantasy XIV.

They already failed at this in Final Fantasy XIV.

Then again... you can hardly blame discerning gamers for not knowing what FFXIV is doing. Discerning gamers don't play it.
I agree. To ignore the
Information Curse
is a bloody beginner's mistake. Are you sure there's nothing more to it, Tobold? Seems too bad for a company like ArenaNet.
In order to go around the sharing on the internet of the combination, why wouldn't you switch some of the ingredients around for each individual character? A given elixir would require component A & B for one player, but B&C for another, finally A&C for the 3rd. If you had let's say 5 ingredients per "level" of ingredients, you could just play permutation. Would still keep the spirit of the discovery, but sharing the information would be useless, thus nullifying the "information curse".
Unless you make the recipes character-dependent, this will fail in the way you described. Making recipes character-dependent would make crafting a complete nightmare, so they'll not do it. Anyway, I'm surprised they are so naive as to expect this to work. Either this is not the full story, or I guess that the game will be a huge letdown after all the hype....

Ryzom did a lot of things right with crafting, of course it did also a lot bad things.... which is why it's dead....
This is a pretty weird idea but what if you had enough degrees of freedom (material combinations, positions, settings etc) You could design a really complex system (from the design POV) that could create items with similar but different recipes per character.

So the sword of awesome for my char could be 2 ultra Irons, heated for a long time, hammered 14 times and cooled fast.

And the same sword for your character would need 3 simple irons plus 1 coal, heated for a medium time, hammered 20 times and cooled slowly.

To minimize frustration a hint system is 100% necessary.
Helistar beat me to it:D
I don't understand how everyone thinks this is an automatic /fail on their part.

Actually, I like it a lot. More so than TOR's system, where you'll essentially have only a handful of recipies and grind grind grind for those rare mats to make something worthwhile.

You're saying that crafting will not be interesting because you can look it up on the internet. If this spoils it for you.... don't look it up? At least you have the option, compared to a static list of things you can make, all available from a trainer. And if you're just itching to make your loot ASAP and are willing to sacrifice the minigame part of crafting, then you probably weren't all that interested in the mechanics in the first place.

I have a little java game on my phone called Alchemy Classic that is essentially the same thing - combine elements together to form new ones, until you've discovered all possible combinations. I could probably have looked them up on the internet by now, but what's the frickin' point?
I think if you wanted to limit the scope of recipes that each crafter has so that not everyone is the same it would be better to limit the number of recipes one can learn at each 'tier' of crafting - and scrap the discovery system. That way each character has to choose to specialize in a specific niche of the entire market by default.

Also - as Tobold pointed out - you can't have a specific item require a different combinations of inputs to be made on different characters (randomly determined) because one will be optimal (by cost) and so crafting becomes luck based but in a very bad way because it's permanently tied to your toon.
The problem of "guess-a-recipe" systems is twofold:
- unless it provides feedback which allows a gradual improvement it becomes synonymous of "try a lot of random stuff and hope for the best", which usually translates in "farm hell" for components.
- unless it's dynamic (= new recipes/components) are added over the time (and maybe old ones are phased out), it gets old really fast.

My ideal crafting system would be Ryzom's with a couple changes.

Going into detail: Ryzom's system works by having static recipes which can be realized using different materials. Example: a sword is a 3x"blade slot" + "handgrip slot" + "pommel". Multiple materials are available for each slot, and depending on the combination of the materials the stats of the final item change. Where it gets interesting is that the determination of the final stats is not just the sum of the stats from each component, but there's a "global fiddling" of values which derives from the combination used. If you think that for every slot there are 4-5 materials and each material comes in 5 grades (from basic to supreme) that's a lot of combinations. The interesting thing is also that using the highest grade materials does not necessarily provide the best result, since at times putting in some lower-grade stuff can improve the "global fiddling" of stats.

The net result is that creating normal-use items is easy, good items requires some work, and optimal items requires a lot of work. Getting the right materials is part of the work, as they can be drops from mobs/bosses or simply gathered.

Where the system failed is that in addition to the stats, the item has a level, which is capped by your skill, and the only way to raise you skill is to craft thousands of crappy items. Translation = farm-grinding nightmare. Additionally, the staticness of component meant that a new player joining in would be faced an army of year-old crafters able to produce superior items.

Phasing out old components in favor of new ones (i.e. new recipes to research all the time), and getting rid of the skill leveling would have made the thing a lot more interesting. An approach where the level of the item is simply your character's level, or where skills are advanced by how good the stuff you create is, would be a lot better.
With your SWG comment on crafting, you reminded me that a lot of today's game has a fundamental theme of loot. What if loot was scaled down to a lesser role with stats and player abilities having a greater percentage of power? What if loot was made to break for good? And, what if the 'good' loot itself was truly rare?

Combine this potential corpse loss and you bring back the risk vs. reward of days past. New mentality? You decide to wear your secondary sword and armor sets to a zone that had high risk. You wear your best sword/armor when you're in a group with relatively low risk.

Let's make a comparison to RL: why would you bring your epic mount (Porsche) and all of your epic gear (platinum watch, suit and top hat) and drive down to the projects for the chance of 1) someone smashing your car with a bat and 2) robbing you of your clothese/accessories? Yet we condone this in today's MMORPG with the minimal risk vs. reward.

Say the above person did get robbed and their Porsche was smashed. Yes, they would be set back some, but next time they want to visit the projects, they can bring the proper gear (sweats) and mount (Honda Accord) and explore the projects with less risk vs. reward.
Perhaps crafting isn't meant to be more than what it is. Sure, if the main point (or one of the main points) of the game is Crafting, then this is terrible news. But GW2 is much more than it's crafting system. I am reserving judgement until I lay a few levels. I like the idea of crafting in an MMO, but it's not the make or break feature for me.
There is no crafting (as of yet) in DCUO either. Personally, I am enjoying one less thing to worry about. I'm enjoying just being able to play the game when I log in and not having to worry about a leveling game within a leveling game.
Depending on the player market that plays the game you could end up with the majority of players not reading up on the game outside if it. It is easy to say not every player reads forums and wikis, but what percent do?

There are still people out there they have no idea why WoW goes down every Tues afternoon in the USA. We don’t see them, because they aren’t on forums and blogs. But on patch day or when a holiday event starts you sure see them in /trade and /general. So yes, if every player looked at the list they could make every item and everyone would look the same. However, since that isn’t the case, the players that don’t look the information up will probably play the game longer experimenting on their own, than they would if there was just a list from the trainer.

So, for educated players it won’t make a difference, to the uneducated it should help Anet retain them a bit longer. All without having to try to create and new or novel crafting implementation.

Companies aren’t trying to be innovative for the sake of being innovative. They are trying to capture and keep the largest part of the market they can. I am still of the belief that most MMO subscribers don’t do much of any research for the games they pay-to-play. So even if 80% of the people online at any given time do research, there are probably >50x that many that only log of for 2-5 hours a week that don’t do much if any research, but still pay the same.

AKA if you polled people in game and asked if they do research you would get much higher numbers than if you polled subscribers – the more you play the game the greater the chance you will research it – Many subscribers don’t really play that much at all.

Another example would be StarCraft II. There will always be much much more people that have bought the game and never played an online match than those whom have played online matches. If you look at the online community for Star Craft it focuses mostly just on online play - you will find much more information on player vs player strategies then strategies to complete the campaign. So just reading online you would think the majority of StarCraft purchasers must have bought the game to play online, which isn’t the case. The players with the higher play times per week often do focus on online, they just don’t represent the majority of purchasers – majority of active players maybe… but not purchasers. A game company cares more about the 5 people that will buy the game play it for a week and stop, and be egger to buy the next expansion, then the 1 person that bought the game and plays it 40+ hours per week.
Look its 1999 and the AC spell system. Wonder how this is going to turn out.
All that's needed for a discovery system to provide pleasure and satisfaction for players that feel they would enjoy such a system is for said players to exercise self-restraint. As anyone who enjoys crossword puzzles knows, just because the answers are in the back of the book doesn't mean you have to look at them.

Taking Rift as an example, the game I'm currently playing, tomorrow I'll have been playing every day for two weeks. As yet I haven't looked up a single piece of information on any website about any aspect of the game. I'll only do so when I find myself in a situation where not to do so would seem more likely to diminish my enjoyment of Rift than to do so.

I would expect this to happen very rarely in any well-constructed MMO and I'm confident that GW2 will be fully enjoyable without recourse to external sources of information, so I imagine a discovery system for crafting should work quite well.
Why all the focus on limiting recipes from the players? Just like in the real world, once a recipe is known, it should be known to everyone else. It doesn't make sense to do x%-based discovery. I don't grind the same thing over and over in the real world and suddenly have a chance at learning an entirely new recipe.

Instead, why not on top of what they're suggesting, add a skill-based layer that actually requires the crafters to be good at what they do? Have some sort of skill checks that require the best crafters be actually good at what they do, not just a bunch of people who stood in front of an anvil all day making crappy daggers and such.
If I recall correctly, EQ2 had a similar system of skill-based crafting, but I don't have any experience with that (never played it). Anyone else here have any such experience with it?
Ravious of Kill Ten Rats has the right of it. We are looking at crafting in GW2 and applying other MMO standards to it. GW2 is not a typical MMO not because it has an instanced storyline, or because it eschewed the holy trinity (although those are contributing factors), but because at its core it seeks to drive players differently than other MMOs.

Even if the hardcores and the minmaxers get to the recipes first - so what? The fact is that if people are driven to rule the in-game world with their uber loot and their uber recipes, (rather than their uber skill of being able to take on a dungeon with three thiefs instead of a 5-man, for example), they will find that the world goes on largely without need of them. GW2 is simply not that kind of game.
This comment has been removed by the author.
I gotta think there is more to it than just 'hidden recipe list'.

I'm not sure the objection to a 'unique character recipe' system is something that can't be answered by clever spreadsheet usage.

If crafting materials are sellable to vendors, then they'll have a set value. It shouldn't be impossible to keep a given recipe's random components within a reasonable variance.
@Grimmjakk: I think that the idea that different methods of crafting the same thing could be made equivalent through a clever spreadsheet is pretty far-fetched. We're talking about an economy of hundreds of thousands of people, no one has every successfully planned or predicted one of those.

I think people are also overestimating the hidden system. Just because the information will be harder to express doesn't mean that it won't be available on the internet. Put a system like that in a game and within six months you'll have people putting up websites showing the optimal path to level crafting professions for alts to determine whether they got the cheapest version of certain desired items and droves of people will be re-rolling when they find out it's not them.

People figure out extremely complex systems by sharing information and analyzing it. If the game is popular enough that people care, everything will be known.
Variable recipes that have the same manufacture costs is achievable if the variable ingredients are the same. For example to craft X, you must have the normal ingredients and 5 stones. The stones come in lots of different shapes, but have the same cost. (This can be achieved by having the stones available from a vendor or crafted from the same material - though this could led to some price flucuation unless all shapes can be crafted with equal ease) The play must then work out what combination of shapes works for them.

Of course, such a system is essentially a time sink as a player has to brute force the combination.

Additionally, once you have the recipe it is exactly the same.

My suggestion to make crafting more interesting is to get rid of it entirely! What I mean by that is that choose an affiliation with a guild eg blacksmith, tailoring etc. However, rather than learning recipies and buying mats and clicking a button to craft, guild npc give you a quest. The quest can complexity and difficulty determines the quality of the item. Quests can be as simple as killing 10 rats to something along the lines of Verigan's Fist questline to the extreme shadowmourne questline. Of course the items are BOA and repeatable. For added complexity, (though it would be difficult to balance) you can have the stats variable, depending how well you complete the quest eg. the faster you complete the quest the better quality of item.
u guys forget the most important; if someone is lucky enough to create a really cool equipment that none else can make , i dont think the first thing he do is go to a wiki and post the recipe .
more like he is goin to mass product and sell it to make profit (in game ofc)
so still more like some basic recipes are going to be posted in such forums with the exception of guild forums (i dont think many will keep it secret from their guild ^^)
Post a Comment

<< Home
Newer›  ‹Older

  Powered by Blogger   Free Page Rank Tool