[KaVir]

This post got a lot longer than I'd intended...but I think it's important to 'set the scene' before describing the problem, so as to get across the angle from which I'm approaching it.



Quite a long time ago I visited The Inquisition, and Robbert gave me a demonstration of his magic system - it involved using rituals to balance yourself against the phases of the different moons before casting a spell, and I found it a very interesting concept.

I was reminded of it again last week while thinking about lockpicking, so I threw together a quick system inspired by the concept. You target a door (or anything else with a lock), hold a lockpick set in one hand, then draw a lockpick from your set with the other hand. You may then use your lockpicks to perform various commands on the lock - each of the three lockpicks has six different commands, each of which adjusts one or more of the five lock tumblers by different amounts in different directions. You can swap lockpicks with a simple command, and each pick has different styles of command, so IMO it's a reasonable way to organise the information.

Each tumbler begins in the range 1-4 or 6-9, and all five need to be at 5 for the door to open (although simpler locks will have as few as 2 tumblers). Performing a lockpicking command gives a message to the lockpicker such as "*click* *clunk* *clonk* *clunk* *clunk*", where "*clunk*" means the tumbler is under the desired setting and "*clonk*" means the tumbler is over the desired setting (these messages are also colour-coded, for quick reference in the lockpick command table). If any command would take one of the tumblers above 9 or below 1, none of the tumblers will move - the user gets a message about the lock being jammed (although this could also be replaced with a trap for the more dangerous locks).

The system itself takes a little practice and a bit of time, but isn't overly hard; it's basically a mini puzzle for players who want to play stealthy characters, yet don't want every challenge to consist of killing monsters. However one of my players - Alayla - recommended applying the same concept to crafting, so I tried a simple alchemy system based on the same approach.

Once again it utilises a 'set' - this time an alchemy set. You hold the set in one hand, then use the other hand to manipulate your potion. Unlike the lockpicks there are only two modes, representing whether or not your vial is over the flame, and the five tumblers are named 'density', 'colour', 'flavour', 'odor' and 'heat'. Each of the two modes has ten commands, although these are mostly the same, except for one mode increasing heat with each command and the other decreasing the heat.

Each potion 'tumbler' starts at either 0 or 1 (to help prevent scripting), and (currently) they all need to be set to 5 for the potion to be complete. If any tumbler drops below 0, the potion congeals and needs to be thrown away, while if any tumbler goes over 9 the potion explodes, doing an (often fatal) amount of damage to the budding alchemist. The messages are a bit more informative than the lockpicking ones, also taking into account the danger levels (0/1 or 9) - thus you might get a message like "The potion is thick, light, sweet, scentless and boiling", which is the point at which you should quickly take the potion off the flame.

The major difference between the two systems, of course, is that with lockpicking you only have one objective (open the lock), while with alchemy there are many possible types of potion that can be brewed. And that's that part where I can't seem to find an appropriate solution.

A popular suggestion is that different tumbler combinations could result in different potions, allowing players to discover potions through experimentation. My concern is that players would simply post the receipes on the web, which would quickly undermine the whole concept and effectively allow all players to brew all potions right from the start. I could get around this by randomising the receipes based on the character ID, but that would completely prevent people from teaching each other or exchanging information - what I'd really like is a middle ground.

The above could also be combined with the requirement for rare ingredients - but that would also amplify the previously mentioned problem. Experimentation is going to be even less appealing if it results in the loss of hard-to-get ingredients.

I'm currently leaning towards taking a leaf from Fable's book - in Fable (on the X-Box), you can get various different tattoos and haircuts, but first you have to find the appropriate 'card' to show the tattooist/hairdresser. Having each receipe as an item would give players additional things to find, win, collect and trade. It also means that the crafting itself would be based completely on player skill, but that the options available would still be limited based on character knowledge.

The receipes could also be stored as knowledge rather than as physical items - something that would perhaps be more realistic (although it could just as easily be said that the receipe cards are magic items and part of the crafting process). However this also removes the value of the knowledge, which (in my experience) means that people will once again soon end up with all the receipes. For this reason, if the skills were stored as knowledge, I would rather force PCs to spend something (exp, skill slots, etc) to learn them. However as crafting isn't a critical part of the game, requiring that people take valuable advancement points away from more important skills would likely result in the crafting system being mostly ignored, or delegated to crafting alts. For this reason, I think the receipe cards would be the better approach.

Various ingredients could still be used in order to limit the number of potions produced, and there could even be some special ingredients which enhance certain (or all) spells - this would allow at least some "open secrets" to be passed around between players, without giving people access to all of the receipes.

For the sake of consistency, whichever approach I take should be applied to the crafting of other items - weapons, armour, and so on. This is another point where I start having doubts about a 'receipe card'...does that really make sense for forging a breastplate? On the other hand, perhaps the armour cards might only apply to magical breastplates, allowing the beginner to practice smithing with ordinary gear (all forms of non-magical equipment will be very easy to get hold of, so this won't be a problem from a balance perspective). This would still maintain consistency with the potions, as potions are just a form of magical food.

Any better solutions out there? Your thoughts are welcome

[Rhuarc]

You could still randomize the recipe, but allow for similarities. IE: recipes within a certain range would have the same effect, possibly greater or lesser than a particular characters 'ideal' recipe value.

This would let people post the recipes on the web, but the posted recipe may not work for any given individual, although it would get them close to right. In fact, the posted recipe may work, but not work nearly as well, since the recipe value may be far from the individuals ideal. (This could be taken further, maybe the recipe does something else entirely for this individual, or maybe it's off far enough to be harmful -- or maybe -- with yet another skill, the alchemist could more easily adapt this other persons recipe to himself?)

This would require some sort of valuation to the recipe, and a way to apply those values to specific affects, but that should be trivial. Especially if each 'tumbler' of the recipe has it's own value (range). Also, it would require determining the ideal value per character per combination (not that all characters would require a unique ideal value), which would be more difficult. Some sort of hash value might be in order.

[Estarra]

I really like the lockpicking system--nifty!

For potions, I'm wondering what if each action (each card?) results in a random one of two results. Say you add salt and the potion could either turn milky or clear. Add sulfur to milky and it could turn sweet or sour. Add sulfur to clear and it could turn either bitter or bland. Your potion for health recipe may result in a milky, sweet potion, and your potion for fire resistance recipe may result in a clear, bitter potion. Thus, although players may still trade or pass recipes, making the potions still retains being somewhat of a mini-puzzle game, adding the twist that while an alchemist is trying to make a health potion the results may be unpredictable so he'll have to think on his feet and turn it into a fire resistance potion in order to salvage his work.

If there's any skill level involved, perhaps the greater skilled the alchemist, the greater chance of getting the results he wants, i.e., when an alchemist with a high skill does ADD SALT FOR HEALTH, it may have a 60% chance of turning milky rather than a 50% chance that a lesser skilled alchemist may have.

[shadowfyr]

I considered an alchemist guild too. But for mine I considered going astrologist on them. Basically, simple formula may work as expected from one player to another, with some variance, but more complex things would start to take into account the numerology of their character name, the phases of the moon, time of day, when the character was created, etc. The same formula for some truely complex one would thus only come close, and even for some people, might produce completely opposite results. Never got too far into it though, it would have been a balance nightmare, not even considering the nightmare of making it work in the first place, with enough variation to not just be, "there are 20 potions you can make total, that is all."

I figure for a system to work really well, you would need to make the entire mud with a magic/alchemical system built straight into the code for everything. So the wings from a bat would automatically have properties in alchemy, instead of having some player add 'bat wings' as an item some place, then being useless because the alchemy system was made a year ealier and doesn't know about them. The reason the magic system would be tied in too, is it makes for a similar option to create spells and provides consitent effects between potions, armor, weapons, spells, etc. Same rules governing the effects, but different rules for how those effects come about. A spell might let you enchant armor, then the next spell, if the right type, would imbed in the armor. An alchemist would maybe make a single potion with both inbedding and prep, or one that allowed the next potion to seep into the armor. Same result, different approach. But again, the problem then become balance, which means taking a 'pure' system of rules and imposing artificial limits on what the system 'can' produce.

Kind of the opposite from normal guilds. Normal = special rules for every thing that 'will' exist, limited actual options, artificial variation. Fluid and adaptable = special rules to try to limit what is made, but unlimited options within those boundaries. Either one is bounds to be complicated if you don't make the result identical for everyone, which is then just boring.

But heh, I am totally nuts. Your simpler system is probably a lot more sane. lol

[KaVir]

Originally Posted by (Rhuarc @ Mar. 02 2005,17:21) You could still randomize the recipe,but allow for similarities. IE: recipes within a certain range would have the same effect, possibly greater or lesser than a particular characters 'ideal' recipe value.

An interesting idea, but I think it would still run into the experimentation problems - if it's too easy to adapt your recipes to match someone elses, then everyone will copy them off the web instead of experimenting, while if it's too difficult then recipe trading won't be worthwhile.

Originally Posted by (Estarra @ Mar. 02 2005,19:10) For potions, I'm wondering what if each action (each card?) results in a random one of two results.

I think that would make it far too difficult - most of the testers have already blown themselves up several times even with the way the system stands now, but if you can't even be sure which result a command will have I don't think many people would ever finish :p  Imagine trying to complete a rubix cube, but every time you tried to perform a move you had to roll a d6, and on a 1-3 you instead moved it in a different direction.  Now imagine that, on certain combinations of that cube, it blew up in your face...

Originally Posted by If there's any skill level involved, perhaps the greater skilled the alchemist, the greater chance of getting the results he wants, i.e., when an alchemist with a high skill does ADD SALT FOR HEALTH, it may have a 60% chance of turning milky rather than a 50% chance that a lesser skilled alchemist may have.

Reduce the failure percentage a lot further, and that might work well - perhaps every command has a 5-10% chance of moving a randomly selected tumbler one point in a random direction (or, as you suggest, base it on your Alchemy skill and the difficulty of the potion).  That would add an unknown risk factor, making it more comparible with activities such as combat.

Originally Posted by (shadowfyr @ Mar. 02 2005,21:22) I figure for a system to work really well, you would need to make the entire mud with a magic/alchemical system built straight into the code for everything.

I think you're right, and what you describe would be excellent for a mud where alchemy was one of the key aspects of gameplay, but that's not something I'm trying to achieve - this isn't designed to be a core feature of the game, but rather a sideline activity which rewards smart players for solving puzzles and encourages interaction through trading.

[Robbert]

Hey KaVir -

I linked this post to Bazarov, the author of the magic system on The Inquisition.  He wanted to reply to your post, but the account system TMS uses takes several days for approval apaprently, so he asked me to post the below:

Originally Posted by On The Inquisition, I had some of the same concerns, of course. I wanted player skill to matter, but I also needed character skill to factor in strongly. Here's how those little pieces fitted in: The number of attempts that the player had to 'balance' the situation wasn't always the same. Depending on how 'out of balance' the situation was (divergence from balance as an absolute value) there was a test against the player's skill level. And the spells themselves had initial divergences built in; a low level spell would be like +1 0 0 +1 0 and a high-level spell like +4 -3 +5 -2 -5 or something. This would be your 'lock difficulty' or 'intended potion difficulty' (or perhaps based off the ingredients involved - salt has little effects, mandrake has big effects). If the player failed that 'test' (the divergence was more than they could handle) the process would move forward quicker, making it even more difficult to manage. This also then forced less-skilled characters to use the most-balancing ritual they could use first, then fine-tuning after they had a 'handle' on the spell, wheras a more skilled practioner could use a simpler or more efficient route. Secondly, the actual rituals themselves were (like your recipes) not all known to each character. They started with only a basic set and had to learn the others by being taught or seeking them out. Master Mages would teach others some of the rituals, but would also keep certain ones to themselves, just like a crafter might. I might suggest something slightly different for your system. For each skill level they have, (or gradation of  ax_rank/numberoftechniques) the player can 'pick' a technique that they can learn. So a player who has half-mastered Alchemy knows half the techniques, and two different half-masters could be closely similar or completely different. So, if you really wanted to make fire-resistance potions, you could pick those techniques that most facilitate that, maybe even see the 'online' formulas for reference. But by cherry-picking those, you would cripple yourself in other areas for the time being. Players could then argue about what sequence of techniques is the 'best' - even in character. By techniques, I mean balancing-tools, not recipes! Clever and creative players could warp and twist what techniques they have to get to the same result as someone else who had a straighter path. It might take you two steps to make a potion of healing but take me 6. They could find out what they need to make a healing salve, then try to figure out just how THEY can make it with their particular set of tools. You have your body of knowledge as well as your individuality. BAZAROV

[Valg]

Something someone mentioned above is rather intriguing:

Let's say the game has a certain set of numbers denoting a 'recipe' for a potion. For each player, that set of numbers gets encrypted by a unique ID attached to the player file. (Upon character creation, one is randomly assigned.) Now, each player has to (essentially, not in an IC way) figure out the answer to that puzzle.

Let's say Alice comes along, and figures out through clever experimentation that he can make a certain potion using objects X, Y, and Z. If Bob came along, and tried the same ingredients, it wouldn't work- Bob has a different ID number, and his items get encrypted differently.

This solves the issue of characters subverting IC communication with OOC information sources, but it creates a quandry as KaVir pointed out... why can't Alice teach Bob ICly?

However, the ID number also solves this. Let's say you create a command called teach. Alice can do:

teach Bob potion_name objectX objectY objectZ

Because the recipes are encrypted in a predictable fashion (Alice can always make that potion the same way... it's not determined by random number at the moment of creation), the code can 'translate' from one character to the other. The code says, "Alice uses X, Y, and Z. I've checked that it correctly would generate potion_name, so I'll give accurate information to Bob instead of misleading results. Using Bob's ID number, I'll tell Bob that he needs U, V, and W."

Now:
1) Alice can teach Bob in-game. Players are rewarded for in-game collaboration.
2) Alice can't teach Bob out-of-game. Web sites aren't very useful. People have to play your game, not fire up Google.
3) ICly, you'll need justification for the different ingredients (alchemy linked to the user's soul, you're balancing out your own magical auras, etc.), but that's fairly easy to do.
4) Alice can lie, or simply be wrong. If she inputs bad ingredients, Bob gets misleading results. The roleplay dynamics of finding a knowledgeable mentor that Bob can trust are interesting.
5) Alice can make a living doing this, thus driving an in-game economy. If Alice is the first person to figure out how to make something, she can charge an arm and a leg for it. Of course, she might also have to swear her new student to secrecy...
6) It's compatible with your game's metric of ability, be it level, skill, etc. The teach command may malfunction if the teacher doesn't fully grasp the Potion Making skill, or if the potion is too high level, or anything else. This helps keep Alice in business- while she might be able to teach Bob how to make his potion, Bob might have to practice quite a bit before he's ready to teach Cindy. In the meantime, Alice is out seeing what price her talents can get.
7) Once Alice dies, the system still works. Her next character has to start as the student again. No one "beats" your game and moves on to other things- each time they play, it's a new challenge, based on interaction in two ways:

A) With other people: She has to convince new master she's a good student.
B) With the game's areas: Knowing you need mandrake root to make PotionA is all well and good, but you still need to go out and get some. And Alice's last character probably never had to bother finding mandrake root, since it wasn't in her recipe.

Thoughts? Problems? Donuts?

[Eagleon]

Hmm, the thing that really bothers me about most alchemy systems out there is they rarely use authentic alchemical ingredients. The most obvious thing to replace the 'cards', in this case, would be a philosopher's stone. If you haven't played Darklands, an old RPG based in medieval Germany, I much recommend it...

Anyway, the stone... You might make it upgradable through short bursts of 'insight', while still keeping away from adding to a skill system. The way I'm thinking of it, the philosopher's stone represents the alchemist's affinity for his ingredients; It's extremely cheap to produce if you know how, but it's useless if you haven't used it for a long time, and gained some intimate knowledge of each stone's intricacies.

The way you learn a stone is pure experimentation, and that's where you could make things more or less expensive for the player, and where the risk might be; Too strong of an acid, and you might ruin your stone, at least until you figure out a way to neutralize it.

Just a few ideas.

[KaVir]

Originally Posted by (Robbert @ Mar. 03 2005,20:59) The number of attempts that the player had to 'balance' the situation wasn't always the same. Depending on how 'out of balance' the situation was (divergence from balance as an absolute value) there was a test against the player's skill level. And the spells themselves had initial divergences built in; a low level spell would be like +1 0 0 +1 0 and a high-level spell like +4 -3 +5 -2 -5 or something. This would be your 'lock difficulty' or 'intended potion difficulty' (or perhaps based off the ingredients involved - salt has little effects, mandrake has big effects).

I considered having a limit on the number of attempts - the problem being that, as it stands, it takes quite a lot of fiddling back and forth to get the potion to the right point (we're talking two or three minutes work). When you're talking about such a large number of attempts, I'm not sure that it really makes sense to have a limit - plus it doesn't really help with locks, as you should be able to spend as long as you like working on them (although in some particular scenarios you might have external time limits, such as an approaching guard).

I also don't want to reduce the number of attempts (at least too much), as the puzzles are supposed to be full game elements in their own right. With the added danger factor of traps and exploding potions, they should be fairly comparible with combat in terms of both risk and duration.

I suppose I could add a bit of a safety net for more skilled characters - the chance of a potion exploding or a trap going off might be reduced, and/or additional warnings about a tumbler getting too close to the danger level. But player skill should also cover these factors fairly well.

Originally Posted by I might suggest something slightly different for your system. For each skill level they have, (or gradation of ax_rank/numberoftechniques) the player can 'pick' a technique that they can learn. So a player who has half-mastered Alchemy knows half the techniques, and two different half-masters could be closely similar or completely different. So, if you really wanted to make fire-resistance potions, you could pick those techniques that most facilitate that, maybe even see the 'online' formulas for reference. But by cherry-picking those, you would cripple yourself in other areas for the time being. Players could then argue about what sequence of techniques is the 'best' - even in character.

An interesting idea, but my concern is that if I add too many it will become too quick and easy to complete the puzzles.

[KaVir]

Originally Posted by (Valg @ Mar. 03 2005,22:57) Let's say the game has a certain set of numbers denoting a 'recipe' for a potion. For each player, that set of numbers gets encrypted by a unique ID attached to the player file. (Upon character creation, one is randomly assigned.) Now, each player has to (essentially, not in an IC way) figure out the answer to that puzzle.

Originally Posted by This solves the issue of characters subverting IC communication with OOC information sources, but it creates a quandry as KaVir pointed out... why can't Alice teach Bob ICly? However, the ID number also solves this. Let's say you create a command called teach.

One concern I have is that of ingredients rarity. If the ingredients are fixed then it's an easy matter to say "mandrake root is used for the more powerful potions, therefore it should be hard to get hold of" - but if the ingredients vary from character to character you can no longer do that. You lose the supply and demand aspects of ingredients trading.

Having said that, I suppose you could get around the problem by dividing the ingredients into rarity levels, and then saying (for example) that a healing potion requires two common ingredients and one slightly uncommon, while a potion of invulnerability requires three rare ingredients, etc.

Another way might be to say that the ingredients are fixed, but that you need to learn the recipe IC before you can use it. You might use google to discover that you need a snail shell, a piece of steelwood bark and a pinch of powdered iron can be combined to create a barkskin potion, but unless your character also knows, you won't be able to create the potion. This would allow the recipes to be made more fitting to the potion type, but the major drawback is that you could no longer experiment. It also raises the question of how anyone would learn a recipe in the first place (although that might fit well with an additional research-based puzzle system I've been thinking of, based on searching through books rather than mixing potions, with the options based on the library you're using - the 'find and trade' aspects I wanted to handle with recipe cards could instead be handled through the discovery of lost tomes of magic, with the more basic recipes being discoverable by spending time searching through city libraries).

A third possibility would be to say that each ingredient has two values, with the first being used to determine the potion type and the second being adjusted by the character ID to generate a side affect. You might look up on the web that snail shell, steelwood bark and powdered iron can be used for the barkskin spell, but then discover that replacing steelwood bark with silverwillow bark provides a longer duration, while replacing the snail shell with a tortoise shell gives a stronger defensive bonus and a minor speed penalty. Having said that, I don't think this system would be a solution in its own right, as the websites would just list all possible ingredients and then just note that their bonuses varied.


The recipe cards could also be used as an optional alternative to teaching - Alice could teach Bob (at which point he has to use his own ingredients), or she could just give him a recipe card (allowing him to use hers - but then preventing him from teaching it to anyone else, although he could just sell the card). Perhaps Alice could even set a limit on the number of potions each card could be used for, allowing her to hold a monopoly on certain potion types without having to spend all her waking hours spawning potions. Perhaps 'recipe card' wouldn't be the best name for it, but that's really just a cosmetic issue.

[KaVir]

Originally Posted by (Eagleon @ Mar. 04 2005,06:14) Hmm, the thing that really bothers me about most alchemy systems out there is they rarely use authentic alchemical ingredients. The most obvious thing to replace the 'cards', in this case, would be a philosopher's stone.

I think it's difficult to define 'authentic' when you're talking about something fictional like magic. I could certainly rename the cards to something better sounding, but as I mentioned previously that's really a cosmetic issue and doesn't affect the core design. The philosopher's stone is more associated with transmuting base metals into gold, but I suppose it could also be used in this context.

Originally Posted by Anyway, the stone... You might make it upgradable through short bursts of 'insight', while still keeping away from adding to a skill system. The way I'm thinking of it, the philosopher's stone represents the alchemist's affinity for his ingredients; It's extremely cheap to produce if you know how, but it's useless if you haven't used it for a long time, and gained some intimate knowledge of each stone's intricacies.

Hmmm interesting - a sort of magical item which can be attuned to the user? Would this attunement involve a separate puzzle system (you spend time 'training' your stone)? Would each alchemist use just one stone, or would they attune different stones for different tasks?

[Robbert]

I like the concept of different ingredients having similar affects when mixed. I think this is one of the lanes of approach to your overall dillema, as well:

A recipe (or a users recipe card) could make the potion, but the longevity and strength of that potion with that recipe could be based on the users individual attributes. Perhaps interrelated with how much experience they have with a particular reagent - the more knowledge they have on it (the more they've used it in various ways), the greater the effects they can wring from it (to a point, more below). Thus a potion of invisibility from Clarice in Podunk could last quite a long time, because she made it with willowbark, chamomile and rat innards; the same potion made with the same ingredients by Fred - who has little experience iwth willowbark, and only heard of the recipe through word of mouth - leaves you translucent and lasts only a few moments.

Later it could be discovered that Fred can make a much stronger potion with fried bat guano, boiled chamomile and a mush of rat innards, dissolved into mineral water, with the same affect (invisibility) as Clarice's.

So you have recipes which can be bandied about, and will provide the users with quick, sure-fire solutions to what they want to make, but may not necessarily be the best for them. Even continued use of that formula won't make it their most effective solution, because they may not be using the ingredients which are most effective to their character.

The balance you're looking for is hard to achieve, and even moreso when what you're doing is making a sidebar to the true game dynamics. You want it to be effective and utilitarian, but not so simple as to be something everyone has.

[Eagleon]

Originally Posted by I think it's difficult to define 'authentic' when you're talking about something fictional like magic. I could certainly rename the cards to something better sounding, but as I mentioned previously that's really a cosmetic issue and doesn't affect the core design.

Of course. It just bothers me when people call it alchemy and then make it herbology. Incidentally, http://levity.com/alchemy/ has a lot of information if anyone wants to start making their alchemy more 'alchemy'-like.

Originally Posted by Hmmm interesting - a sort of magical item which can be attuned to the user? Would this attunement involve a separate puzzle system (you spend time 'training' your stone)? Would each alchemist use just one stone, or would they attune different stones for different tasks?

I was thinking one stone, and that you could use the stone in different recipies to potentially make them more powerful. The tricky part would be keeping your stone undamaged and usable, but random, pre-determined conditions in experimentation for each stone would let you learn more about it.

You might also intentionally make a stone upgrade only after hazarding potentially dangerous mixes; You have a potion of explosion, a stone, and you think by process of elimination that this recipe is the one which will teach you more about your stone. Hmm...

Also, I suppose you could make some variance in the potential of each stone, and make each one better or worse for different tasks so that it's to the alchemist's advantage to learn more of them, but I'd keep them focused on only one or two stones.

[shadowfyr]

Originally Posted by (Valg @ Mar. 03 2005,16:57) Something someone mentioned above is rather intriguing: Let's say the game has a certain set of numbers denoting a 'recipe' for a potion.  For each player, that set of numbers gets encrypted by a unique ID attached to the player file.   (Upon character creation, one is randomly assigned.)  Now, each player has to (essentially, not in an IC way) figure out the answer to that puzzle.   ... Thoughts?  Problems?  Donuts?

Hmm. Actually that kind of solves some issues and would be a lot easier that having to code for the name, etc. I would still throw a wrench into things to some degree by adding in stuff like moon phases or the like. Its not unheard of for some things to only work on full moons, dark of the moon, midnight, winter instead of summer, etc. Though, seasons are more problematic on must games.

Only comment I could make is that 'simple' alchemy should be just that, simple. I would consider making the level of the potion determine how far off the ingredients are. A truly high level one might use completely different ones, a low level one might just be a slight difference in amounts and a 1st level cantrip type alchemy might be identical for everyone, with little or any difference. Or even more interesting, their could be powerful 'primary' igredients, which your spirit, etc. would cause interference with. The true alchemy would be combining other 'balancers' that differ from person to person. You might know that a pheonix feather could either create a fireball like effect or let you empower something to give you permanent flight, but without first learning what ingredience counter the effect you don't want, you might get the wrong thing or end up with a piece of armor that literally makes you fly and fry at the same time. Needless to say, this would prove to be a major 'Oops!', unless by some coincidence you also accidentally got the heat to radiate towards 'apponent' instead of 'self'.

In a truely complex enough system you could end up with something that exploded the moment you used it to fly or any number of other possible problems, not the least being accidentally producing the exact opposite of the effect, which itself would be bad. Armor that freezes the wearer? A bracelet that nullified flight in those that normally have it? A detailed enough system could give even the best alchemist nightmares over what might go wrong.