# Unreliable Magic - Is it worth it?

In my world, magic is unavoidably unreliable, for example:

You have a puncture. You cast a spell to mend it and with equal probability either it is mended or another tire is punctured.

You are trying to save an injured person. You cast a spell. Either they are saved or you kill them.

Question

It seems to me that such a magic is completely worthless. Can anyone prove me wrong by suggesting a case where both the intended result of a spell and its opposite would both be advantageous?

Note

There is no way to make magic more reliable. It is just a fact.

EDIT

I have been asked to clarify the extent of the unreliability. It does not happen with mathematical precision but there has to be some proportionality. For example if I try to light a fire I won't accidentally freeze the entire continent.

Examples of how things might go wrong:

I try to heat up my dinner. I end up with a frozen meal.

I try to repair a broken vase. Something else of a similar size falls off a shelf and breaks.

Explanation

The supreme being has tasked sprites with keeping magic balanced. There must be an equal amount of desired results as undesired. The sprites however have limited intelligence and a short attention span. Therefore to make their job easy they apply this 'balance' to each spell as it happens. They do their best either to comply or to do what they perceive to be the opposite. They are not mathematicians, they just do their best to maintain balance. Sprites flit around in the spirit dimension so you are very unlikely to get the same one for two spells in a row.

Magic is fairly small scale. It relates to everyday life. You can't use it to make the Sun disappear or even make a person disappear. The spell has to be 'plausible' and within the capabilities of a sprite.

EDIT 2

Some are trying to get me to redefine 'unreliable' as 'predictably wrong'. However those are quite distinct concepts. Sprites have autonomy and can use a certain amount of discretion to suit the circumstances. Sprite A might think "opposite" of "cast fireball at enemy" is "cast ice ball at enemy" but Sprite B might think the opposite is "cast fireball at spellcaster."

Most sprites would treat a simple coin-toss as resulting in heads or tails. In probabilistic terms your spell won't make a difference. However life in general isn't that simple so they have to use discretion in the limited time they have. Sometimes they will make a snap decision that is roughly opposite in their estimation.

As a guide, imagine that you are a sprite. You have no vested interest in or sympathy about the result - what do you choose in a given situation to be the desired outcome or the opposite? Pick the most obvious and move on.

• Comments are not for extended discussion; this conversation has been moved to chat.
– L.Dutch
Commented Mar 30, 2019 at 6:39
• If only I could remember the Pratchett quote about using magic being similar to trying to build a house of cards with cards that are razor-sharp, I'd post it here. Perhaps @Valorum has access to it? Commented Mar 30, 2019 at 11:02
• Now I'm wondering if this supreme being purposefully created flawed sprites, or maybe this being isn't quite so supreme... Commented Feb 20, 2020 at 8:24

## Sure, it's worth it in many cases.

For instance, if your car is stranded in the middle of the wilderness due to a tire puncture, using magic to try and mend it would be reasonable. In this situation, the car having two punctured tires isn't any worse than with only one. You'll still have to make your way to civilisation on foot either way. On the other hand, in the best case scenario (50% chance, so it's not even unlikely), it'll be repaired, and you can complete your journey within the comfort of your automobile.

Many similar examples can be thought of in the case of an injured person. Let's say one of your battle compatriots has been injured and you're about the be surrounded by the enemy who you know will show you no mercy. Well, then it makes perfect sense for you to attempt to cure him. Either he will recover and perhaps help your group stand a better chance against your foes, or he will die by magic instead of by the hacking and piercing of steel.

There are all sorts of times where a partly broken object or person is no more useful than a very broken object or person. It's times like these that magic comes in handy.

• Beat me to it by a minute or so ;) Commented Mar 27, 2019 at 21:23
• And me as well. Ah, well. Great minds think alike? (nervous grin)
– MacA
Commented Mar 27, 2019 at 21:25
• It's also the case that it hardly matters, think of all the millions of individuals and companies who have invested significant resources in IT 'time saving' devices and etc and then spent all the time they might have saved learning how to use it or fixing problems with it, when a type writer and filing system would have cost less...yet they persevered into bankruptcy.. Society benefits from those early adopters, but not necessarily the early adopters themselves. There's always plenty of people willing to try the shortcut or 'great leap forward'...or lottery ticket Commented Mar 27, 2019 at 22:34
• And if you have two punctures you can cast it again. You either have fixed car or 4 punctured tires. This time probability favors you. Commented Mar 28, 2019 at 0:15
• Not even counting combat related magic... if my fireball has a 50% chance to do something random, there's a decent chance it'll still be harmful. Commented Mar 29, 2019 at 12:28

### If repeated castings are possible, then the odds just got a lot better.

There are all sorts of times where a partly broken object or person is no more useful than a very broken object or person. It's times like these that magic comes in handy.

The corollary to this is that if you can keep trying repeatedly, and you're not significantly worse off with a failure, you'll eventually succeed. You went from one puncture to two? Cast it again. You have four punctured tires? Cast a spell to repair them all. Your car now has a giant hole in the middle? Cast it again. Eventually, you'll get your car back in perfect condition.

Obviously, this won't work if the "bad" outcome is worse than magic can handle (such as killing your friend, if resurrection is beyond magical means), but in that case, you're back to where we started with AngelPray's quote.

• I was going to post something like this but with a counter spell/shield idea. Cast a shield before the spell, you have a 25% chance that the original spell and the shield will both fail at the same time. Repeat for even better chances. Commented Mar 27, 2019 at 23:17
• It might be worth adding reference to en.wikipedia.org/wiki/Martingale_(betting_system) Commented Mar 28, 2019 at 8:20
• Be careful though with such impredictible magic. If you fail to repair one tire, you wreck another. If you fail to repair both, maybe you break 4. The next logical step is 8 punctured tires, and one has to wonder where and how these tires will manifest. It would be a shame if the next car that comes around loses control due to flat tires and runs you over. Eventually the probabilities will favour you, but you need to survive until then. Commented Mar 28, 2019 at 15:22
• "if resurrection is beyond magical means" I would guess resurrection spells would be rare anyway if the backfiring were killing the caster. Commented Mar 28, 2019 at 17:38
• I would picture it as sort of a state machine. Trying to take a used tire and make it new again might shift it to "new" or to "punctured". Trying to take a punctured tire and switch it to whole again make take it to "whole" or "disintegrated". Trying to go from no tire at all to having a tire might take it to "tire" or "your head explodes". Maybe if the sprites can't think of a reasonable worse state, they just blow your head up as a default. Commented Mar 29, 2019 at 22:55

It depends on how unreliable your magic really is. If sometimes your magic does what you want, and sometimes does the opposite, then you can make it do what you want 100% of the time (for certain spells)!

1. Either way it goes, it still gets the job done. There are a lot of ways of hurting people. You could conjure an enormous fireball, adding an incredible amount of energy to a localized area... or you could conjure a blast of ice, removing the same amount of energy. But burned or frozen, the enemy is still just as dead.

2. Ignore the bad results. Sometimes when you try to make a pebble glow brightly, it instead creates darkness. No worries; if you cast the same spell a bunch of times, you can throw away the "darkness stones," and hang on to the glow stones you made. Or, in the case of a punctured tire, don't fix just one - try to fix a whole pile, and sell the ones you fixed. And when you need to inflate it, a spell that removes all the air from the tire is easy to revert until you get the spell that fills it up.

3. Make magical items. Sure, sometimes your spell makes a potion of healing and sometimes it makes a potion of harming, but you can test the result, and people will buy both.

4. Use spells that either work or fizzle. Trying to make something do something it's not doing - a person fly, a dog speak English, a cat obey - will either work, or do the opposite... which is nothing at all. Sure, you're going to have to cast spells more than once, but you'll eventually get the job done.

If you want to cause damage, that's easy; if you want to be constructive, that's a little harder, but carefully choosing what your spell does will ensure that after enough attempts, your spell will eventually work as intended.

• So, basically, just p-value fishing :)
– vsz
Commented Mar 28, 2019 at 8:55
• Is "freezing my enemy" the opposite of "burning my enemy", or is it "freezing myself"? Commented Mar 28, 2019 at 15:11
• @AmiralPatate I guess it depends on exactly how opposite it is. "Sometimes has the opposite effect", or "sometimes does the literal exact opposite"? Is "over there"'s opposite "over here", or "the same distance on the other side of me"? Either way, of course, a little experimentation will show the right spells to use. Commented Mar 28, 2019 at 19:20
• +1 for point 2 - the majority of the population are not trying to fireball anyone. Forget about glowing pebbles though - think about turning rocks into mobile phones, laptops, solar cells etc. Just make sure that your quality control testing equipment is manufactured the mundane way. Any failures can be re-used. Only risk is if a failure mode is for the caster trying to turn a rock into a mobile phone to be turned into a mobile phone themselves... Commented Mar 29, 2019 at 8:58

Such a magic would be useful for those situations where you are not too concerned with the immediate outcome but rather the consequences of those actions.

Rough example. There is a war. You don't particulary care who wins but you want the ceaseless fighting and killing to end. So you cast your magic on one side to win. Either they win or they lose...but the war is over! (And if the magic did nothing...try again)

You have to trick the magic into giving you the real end result you want.

• Relevant username... Commented Mar 27, 2019 at 22:30
• This post contains roughly 3976 helpful bits Commented Mar 28, 2019 at 13:45
• Or the sprites might make a pyrrhic victory instead of a true win. If the sprites are looking to undermine intent 50% of the time then you can’t win this way.
– SRM
Commented Mar 29, 2019 at 8:56
• @SRM, the answer was written before the lengthly edit with details about the sprites unreliability. Commented Mar 29, 2019 at 9:52

I suppose this depends if your 'opposite effect' is predictable. In combat you may, for example, try to fireball an enemy, or maybe throw them into the air; you may not necessarily be concerned if they end up frozen or pinned to the ground instead. But you may not want to risk casting the spells if other unpredictable outcomes are possible.

• This exactly. Reminds me of the 'teleport' spell in the original Diablo 1 game. The spell was guaranteed to "get you out of there", but you had no idea where that would be. In theory it could take you to a far worse spot than where you started from. It also reminds me of those sweets in the harry potter universe, I forget what they're called. Seems to me people enjoyed them regardless of the 50% risk you could be tasting earwax next. Commented Mar 28, 2019 at 15:41
• @TasosPapastylianou Every Flavour Beans? Commented Mar 28, 2019 at 17:30

Use it as a weapon. Your mages go into battle and heal the enemy. Fix their high blood pressure. Congenitive heart defect or heart murmur? Cure their gout. If it succeeds the enemy isn't really better off, you aren't reviving a fallen warrior. You may also be able to cast it again until you do cause catastrophic or deadly injuries. If it fails, one less enemy. With a fifty percent success rate your mages will be racking up kills fairly quickly.

• yep If you can cast them in quick succession, just do so until the person has something horrible happen to him - it doesn't matter that you cured his acne on the way to exploding his cranium :-)
– user45032
Commented Mar 28, 2019 at 17:52

Statistically if you run the spell enough times you would eventually get what you want.

1. So you cast a spell to heal someone, but they die.
2. Then you cast a spell to bring them back to life but their body denigrates.
3. Then you cast a spell to turn their dust into the original person.
4. And so on.

Of course the down side is that each time you fail you may need a better and better spell (and possibly more energy) to fix the accumulated effects of the failed spells. So in the long run those who can cast better spells and who have more stamina have a better chance of getting what they want.

There is also the case of casting a spell that enables you to do something that was already very improbable.

So I cast a spell to print out the lotto numbers on a piece of paper... There is a 50% chance that It doesn't work out and the paper disintegrates. Well that's way better than the one in a billion chance I had before.

• I thought of the lottery one myself, however I'm thinking the "opposite" of this is that the magic finds a way of putting you millions of euros in debt... Commented Mar 29, 2019 at 8:39

Sometimes a simple corollary is the most efficient way to answer a question:

People undergo surgery to fix something. Sometimes it kills them.

Surgery is a \$21 billion dollar industry

• true, but the chances aren't 50:50 while the question specifically states that both outcomes are equally likely.
– Tom
Commented Mar 28, 2019 at 5:28
• @Tom : but there are many illnesses (and surgeries) where the chance to survive is less than 50%. In such cases the spell will improve the odds.
– vsz
Commented Mar 28, 2019 at 6:31
• To add to this, I recently refreshed on CPR training. One of the things they repeat, to drive the point home is "Don't worry about breaking bones or hurting them. you can't make things worse" While you could argue that a stopped heart + broken ribs is worse than just the former, that is not the concern if they'll be dead either way. Better to try and keep them alive, instead of worrying about something so unimportant as an injury. Commented Mar 28, 2019 at 9:21
• @Baldrickk well if your course was well done, you understand that it's "a stopped heart someone kept mechanically beating to prevent the advancement of the clock of death and brain damage + broken ribs vs a stopped heart, brain and neurological damage as time goes by"..... so it's not really the same here, imho.... (but I still get your point :) ) Commented Mar 28, 2019 at 12:24
• @Patrice yes, they did go through that, they also went over the stats of how many heart attack victims who were alive when the ambulance was called, are not when it arrives - without CPR, it's a tiny number. At this point, we may as well just compare "dead" to "maybe not dead" Commented Mar 28, 2019 at 16:22

When it comes to effects whose opposites are also useful, it strongly depends on what "opposite" means when it comes to an effect. Hopefully it's not some kind of pseudo-intelligent monkey paw type of opposite.

For example, if I cast a spell to induce current in a wire, is the opposite effect to induce an opposite current from what I want? If so, we just add a rectifier to the circuit, so the output current is always the same.

We could also generate power pneumatically by casting a spell to double the pressure inside an air tank. If this tank has one-way check valves (i.e. a rectifier), we can control the flow of air and use that to spin turbines.

It's even possible to build rotational rectifiers, so you could spin something with magic and have the output always be the same direction. Or magically kick a pendulum hooked up to an escapement.

If any of these work, congratulations, you have infinite free energy and with that you can develop technology that is distinguishable from magic by being actually reliable.

• The idea of "what is the opposite" is ofcourse completely contextual. If I have 1 apple and use a spell to add 1 other of that same apple, the opposite is removing one. If I double the amount of 1 apple, the opposite is halving the apple... A simple change in wording the same effect can already change what the opposite is! :) Commented Mar 29, 2019 at 7:38

Your magic is statistically useless on the individual level.

If I have a band of soldiers, all wounded badly, then I don't have a fighting unit. If I do nothing, then over time they will either heal or die. In fact, until the invention of antibiotics and other modern medicine, the chance to die from a serious wound was pretty much 50:50 anyway.

Your magic basically just accelerates this process. But that means I get a fighting unit back! It is now half the original size, but that half is healed and ready to fight. Very useful!

So on the level of groups or collections, your magic is far from useless.

• I don't see why your example is useless to the individual. Every wounded soldier can recover completely or die quickly, either of which is probably preferable to a slow, painful death from a severe battle wound. There will be individuals with a <50% chance of survival, so the magic is clearly of benefit to them as it improves their individual odds of survival. Commented Mar 28, 2019 at 13:08
• Opposite of “Heal soldier” could be “Hurt another soldier”. Then the magic is categorically useless.
– SRM
Commented Mar 29, 2019 at 9:07

I'll not add more of the "use chance calculations" answer but provide a surprisingly still unique view on this subject.

Your magic is what I think of as "aware magic". If you try to mend a hole in your Tire the magic (or the user of the magic) is aware of what a hole is, what a wheel is and how it needs to be to be "repaired", or how to punch a hole in another wheel instead of in the car door, the asphalt or the person casting the spell. So if you are aware of what the magic or the user thinks is the opposite you can use it.

Example: you fill two large canisters with gas or liquid and connect them, put a dynamo in the middle and a release valve. Make the gas/liquid something valuable.

Now you cast a spell on the contents of one canister to double its contents, increasing the pressure. If it succeeds then the pressure drives the dynamo as the pressure pushes the opposing piston. If it failes it will halve the contents and the opposing canister will now have its contents flow to the other side. No matter which outcome you get you get electricity, and probability says that because the gas creates an equilibrium, you are more likely to create more gas in the long run.

If you get more gas and the pressure gets too high the release valve can be used and now you have created both energy and a valuable material, which you can store, sell or use. Sometimes chance will make you halve the gas content so much that you have to refill it, but with the stored gas from the times you generated too much you should be able to do this virtually forever.

Another option: nuclear waste disposal. You cast a spell to rejuvinate the waste into useable nuclear fuel. If successful you need to buy less of the expensive nuclear fuel or get rid of it.

If unsuccessful it depends on the outcome:

• It disappears. Yey! No disposal and a lot of concerns put to rest.

• It doubles. Yey! Double the nuclear waste means double the chance to purify it (even better if your spell can target the heap and have 50% chance success each time). This means absolute infinite energy for everyone, assuming magic avoids the Newtons law pitfall (which it usually does).

• It becomes extremely radioactive. Well either you can use that for the nuclear reactor, or if he was alive Ghandi would know how to use that.

https://youtu.be/lQBV3-kwh5k

Although gambling has come up, the answers mentioning it aren't going where the money is.

Gambling houses make much money off people betting on the outcomes of (apparently) random events. Don't use your magic to alter the outcome of a game. Go into business as a casino specializing in games based on magic misfire chances. Then you're not playing a slightly tilted gambler's ruin. Related: sports betting. Related: stock markets. The players are playing a weakly negative sum game because the house takes a commission from every bet. It's much better to be the house in that story.

This form of magic can completely overwhelm the first law of thermodynamics. Apparently, your magic can create or destroy energy (randomly choosing one or the other). Either way, without much effort, this can be transformed into a heat engine. "Magic, make this cylinder 100 K hotter." Maybe it's hotter. Maybe it's colder. Maybe its temperature is the same, but the other cylinder stored near it has warmed or cooled. If less entropy is produced by casting the spell than is removed by its effects, this is a win.

The field of risk management (or more generally, the study of risk) is the study of when you would want to use this type of magic. If the negative outcomes are "cheap" and the positive outcomes are "fabulous wealth", i.e., if the potential benefits times their likelihoods outweigh the potential negatives times their likelihoods, it is rational to use this kind of magic. See also reversible versus irreversible decisions.

Unreliable Magic worth it as long as the risk, production and cost is under control within a sustainable amount.

The sustainability can be subjective. Putting those environmental drawback aside. Nuclear power having risk of leakage and it is costly compare to gas and oil, but the power produced is far greater in 10x.

Another example with unreliable magic in some other culture, Strait_Jacket treats Unreliable Magic as a kind of pollution.

Due to an invisible contaminant called the "malediction", or simply the "curse", people who use magic too often are at risk in transforming into "Demons," or horrific, malevolent abominations of nature that become immune to ordinary weapons.

This is sustainable as long as the damage comes under control with a special mess cleaning up organization. One of the scene, the surgeon having magic overdosed in the operation theatre while using magic like x-ray. Potentially the x-ray benefit produces a lot more than the risk of having the additional bit of pollution.

• Welcome Kelvin Ng, whilst you've given an interesting starting point, this doesn't answer the question. You can edit your answer to change it though. Commented Mar 28, 2019 at 11:39

It could be tremendously valuable in some scenarios, especially with a small amount of planning ahead. If you intentionally provide circumstances to make it easy for sprites to see a negative outcome, you can potentially mitigate the negative result.

With the vase example: If you have a vase that you want repaired, precariously place several objects of similar size but less value on a shelf. If the spell fails, you break a worthless object. I wouldn't anticipate the sprites to have the time or intelligence to assess the value of an object, and value is arbitrary anyway, so your negative result could very easily be mitigated.

You could even use it in your favor. If while making a cake you accidentally burn it, you can whip up some more batter and try to bake it using magic. If it succeeds, you have a perfectly baked cake. If it fails, your burnt cake is un-baked, giving you the ingredients to try again.

In the Coldfire trilogy, by Celia S. Friedman, magic is very wild. It can manifest just by a person untamed thoughts/fears, yet it can be controlled by very talented people.

https://en.wikipedia.org/wiki/Coldfire_Trilogy

One of these people is Gerald Tarrant, who is capable of controlling it beyond what anyone believes is possible. He can do it because of his very powerful mind, among other reasons.

This isn't an exact 1:1 relationship with your world, but there's a lot of similarities. With about 1000 pages to each book, you should run into a lot of the same problems as your world. Even if not, it's a very good series and might give you some ideas anyway.

# Twilight City, gonna set my soul,

gonna set my soul on fire

Roulette has 38 slots. Betting on the correct one is 35 to one pay out. Except this magic just made the odds 2 to 1. Start with ten dollars. Bet one dollar on 25 and cast your spell. Spell works, rake in you money and make a moderately larger bet. Spell fails, and it lands opposite 25, or some other number, or even you accidentally bet on 24 instead. Whatever the case, you try again. By steadily increasing you bet with each win, while keeping up a reserve against losses, you can quickly make a huge sum (ain't exponential growth grand).

Roulette not interactive enough? Try craps. Those both too slow? Most progressive slot machines start at about a 20K to one payout, and increase over time.

Of course, cast some spells to keep the casino from noticing first. And now all you have to worry about is another caster at the same game. Or bankrupting the entire city.

## Statistics are on your side here.

There's a gambling algorithm that guarantees indefinite gain given:

• There's no limit to how much debt you can be in
• You can choose when to stop playing, and the house can't.

Here's how it works:

1. You bet 1 dollar.
2. If you win, take your winnings and repeat from the beginning. If you lose, increase the amount you bet so that it is equal to 1 + the amount you owe. 3.Bet again. If you win, you recoup your losses and gain a dollar. If you lose, repeat step 2. Keep alternating between 2 and 3 until you win, then go back to 1. Continue playing until you have the desired amount.

This is how your magic works. If you puncture a tire, try magic. If it fails, you have 2 flat tires. Try again with those 2. If you fail again, you have 4 flats. Try again. Eventually, the odds that you keep failing are so low that as long as each new spell fully reverts the damage done by the previous mistakes, you will eventually solve your problem.

The only problem is when the stakes get too high and you do end up freezing a continent, or kill yourself so you can't keep going, or run out of energy: you don't have unlimited debt. But you can choose when to leave the table, which gives you a solid advantage.