In my world, fortune tellers and psychics aren't con men but actually possess supernatural abilities. People born into the Romani lifestyle can make repeated and accurate predictions via tarot cards, palm reading, dowsing, astrology, abacomancy, crystal balls, and other forms of divination. Fortune tellers cannot scry more than one year in the future. A future prediction can also be wrong if a new event occurs (like hearing a fortune) and people change their actions. Of course, there is also a chance that the prediction included other people's predictions and created a self-fulfilling prophecy. Fortune tellers initially use their abilities to great effect on the stock market, casinos, lotteries, sporting events, foreign policy, etc.

So if you are an entrepreneur who runs a lottery or owns a casino, and you're tired of fortune tellers winning all the time, what do you do? Banning fortune tellers wouldn't work because others would simply hire them. You could hire fortune tellers of your own and try to change the predictions but that could create a self-fulfilling prophecy. Random number generation won't make a difference because the fortune teller knows what the final result is going to be.

  • 3
    $\begingroup$ Brainstorming is prohibited on this site. So is asking questions about the decisions of characters within your world. $\endgroup$
    – sphennings
    Commented Sep 23, 2022 at 0:52
  • 4
    $\begingroup$ Easy: only play games where the gains are split among the winners, minus taxes and casino commission. (so, roulette and lottery, but no blackjack). If everyone gets it right, there's no benefits to pay a fortune teller. Some balance happens. I would imagine that only low enough prizes remain. $\endgroup$
    – njzk2
    Commented Sep 23, 2022 at 20:47
  • 2
    $\begingroup$ Err, given human nature, why does a group of people who can predict the future continue choose to live the Romani lifestyle? (And, no, it's not going to be some belief system. There's always that one guy.) $\endgroup$
    – Ouroborus
    Commented Sep 24, 2022 at 0:32
  • 1
    $\begingroup$ Frame challenge: what makes you think any number of real clairvoyants is going to share this gift with outsiders? Especially when the Romani people are already particularly insular? $\endgroup$
    – user8827
    Commented Sep 24, 2022 at 23:20
  • 3
    $\begingroup$ If clairvoyants are routinely (correctly) predicting future stock prices, then stocks will be priced correctly and there will be basically no money for anyone to make on the stock market, besides a basic interest rate return on invested capital. $\endgroup$
    – Jedediah
    Commented Sep 25, 2022 at 1:46

15 Answers 15


Maximize the impact of the fortune-tellings on the outcome of the draw

Don't use a normal random-number generator to draw the number. Instead use a cryptographic hash function (like SHA-1), to generate a random result from the data describing all the bets.

The beauty of this is that anyone can predict the outcome, but if they then try to use this prediction to place another bet, they would change the outcome.

A future prediction can also be wrong if a new event occurs (like hearing a fortune) and people change their actions.

Thus, any predictions about the drawing would be no better than a random guess if someone hearing a fortune uses this information to change whether or not they place a bet (or how large their bet is).

EDIT: Example

I initially didn't want to bog down the answer with too much details, but I'll try to give an example, as it might make the idea more clear:

Say the lottery needs to draw six random numbers between 0 and 99, allowing duplicates. The lottery allows bets of any amount on combinations of numbers (and somehow pays out based on the number of correct guesses, but that's not important to the drawing itself).

As @JohnO points out, a random string of letters can be added to the end of the list to avoid non-fortune-tellers to easily compute the result ahead of the announcement. Obviously, a fortune teller would be able to predict this random number, and the outcome of the draw.

The list of bets would look something like this:

John Smith   $10   10,10,20,30,50,80
Jane Clark   $99   5,45,89,90,95,99
Jack Locke   $11   4,8,15,16,23,42
Adam Baker   $200  1,2,4,8,16,32

During the draw, this list (as a text-file) would be passed through a hash-algorithm, to obtain a number such as


This hash is in fact just a number written in hexadecimal, so we can easily get our 6 numbers by converting to base 10 and taking the last 12 digits, giving us the numbers 90, 85, 31, 49, 78, and 17.

Now, a fortune teller could predict all of this, and know the numbers ahead of time. However any change to the bets based on this newly devined information would completely change the hashed value, and thus the outcome of the draw.

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$
    – L.Dutch
    Commented Sep 24, 2022 at 17:08
  • 3
    $\begingroup$ This is a cool suggestion, but it could probably be more accessible if it didn't require something as complicated as a hash function to introduce the dependency on the bets. For example, one might have a standard lottery of drawing tokens from a container, but have the number of shakes/stirs/etc of the container depend on the guesses. $\endgroup$
    – Josiah
    Commented Sep 25, 2022 at 17:30
  • 4
    $\begingroup$ Of all the stack exchanges, I didn't expect to have to point out on this one that SHA-1 is cryptographically broken and should not be used. That said, if you used the more famously broken MD5 algorithm then you could make a plot point of this kind of attack. $\endgroup$
    – kaya3
    Commented Sep 26, 2022 at 3:25
  • $\begingroup$ @Josiah hash functions aren't complicated in any way. Especially if you're just using them it's as easy as any line of code since you're just using a library for it. Any bet has to be registered before drawing anyway (to prevent cheating) so I see absolutely no problem in implementing such a system $\endgroup$
    – Hobbamok
    Commented Sep 26, 2022 at 9:18
  • 1
    $\begingroup$ Ooooh, very elegant solution. $\endgroup$
    – brice
    Commented Sep 28, 2022 at 8:55

Don't bother with small change

Any effective and accurate fortune tellers would be far too rich (and busy) with their demanding jobs forecasting for banks and insurance firms and the Weather Bureau and the military and the police and investment houses and a hundred other vital organizations to bother with piddly little casino heists. They simply won't have the free time, and the payoff is too small to bother.

A few clairvoyants might own the casinos as a hobby projects. They would be rich enough. They would obviously know when a heist gang --likely using an inferior clairvoyant who couldn't hold a better job-- was going to work their shed. Then it's clairvoyant vs. clairvoyant for relatively low stakes, and you can have it come out any way you like. Be sure to end it with a big dance number.


Place bets, then predict:

Lotteries will be easy, and even a casino could still be made to work. There are two factors that allow this. Prediction can alter prediction, and the prediction can't be more than a year in advance.

So fundamentally, you can't predict the outcome of a prediction.

As a result, your Fortune tellers would be making predictions of results that will happen in a year. Lottery players will place their bets, and after all bets are in, the fortune teller will let you know what number will be generated - in a year. Any attempt to predict beyond a year will fail (so pre-predictions of next year's lotto will fail) and any attempt to predict the prediction will fail (because you can't predict the outcome of a prediction, and if you could, then a chain of predictions could potentially see far into the future as you predicted the predicting of predictions, thus violating the year limit).

Casinos might be odd affairs. The wheels and slots are cranking out result before people's eyes that are already known, because they were gambled on LAST YEAR. While a lottery would require just one fortune teller, a casino would require as many as the games they offered.

Anyone who could see even a few extra days into the future would have a huge advantage. But someone's always gaming the system.

Because sports involves so many people and so much statistical prediction already, it would be virtually unaffected - as long as there was a prediction about the game publicly posted. The randomizing effects of so many people on the teams and in the stands would mean tiny perturbations in reality would re-randomize the effects of individual performance. As it is, predictions tell us the likely outcomes of sports events, but they can't guarantee random effects don't change it. Any coach who is told they will lose a game will change their strategy. Even attendance at the stadium can affect the outcomes, and if you think your team will win or lose, it may affect if you go or not.

The other solution is to have games and slot machines with very modest payouts. It allows the fun of the gambling, but any fortune teller who could predict the outcomes would make a lot more money predicting stocks or working for the casinos. So you can still fleece the little guy, and the big fish can play at the (now even more exclusive) high roller tables run by fortune tellers.

  • $\begingroup$ … So, you take bets, close betting, get your Resident Psychic, the Great Rhymhousi, to predict the numbers. No one guessed right, so the House Wins. 12 months — and lots of money — later, right before the first numbers should be rolled, you and your non-psychic friend "Rhymhousi" skip town, before anyone can realise that you were just secretly passing him a list of numbers no one had bet on as part of your scam. $\endgroup$ Commented Sep 26, 2022 at 1:07
  • $\begingroup$ @Chronocidal There would be loads of money paid to these psychics, so while fraud is possible (people are people), the lotto one would need to be unimpeachable. Casinos are always and still are kind of suspect. Someone will always cheat, but it at least theoretically allows a fair game. $\endgroup$
    – DWKraus
    Commented Sep 26, 2022 at 1:14

"Reverse Observer Effect"

Get a couple of good, usually reliable fortune tellers. More than two, for safety. Put them into different cabins and have them predict the lottery outcome. Compare the predictions.

  • If most/all predictions are the same, do something physical to the random number machine (re-sorting the initial position of the balls, rotating a drum with the tickets, etc.) and repeat the fortune telling to check if that was enough. That's reacting to hearing a fortune to make it less likely.
  • If the predictions differ, the draw is safe. Hurry, draw the numbers.
  • $\begingroup$ If they use a randomizer like on TV (selecting a digit by pulling a random ball from a hopper) than removing one of the commonly guessed digits from the hopper which provides that digit to the pick will ensure the future cannot happen. If it does, then perhaps the winner cheated. $\endgroup$
    – hszmv
    Commented Sep 23, 2022 at 19:27
  • $\begingroup$ @hszmv, that would penalize people who picked that number without aid of precondition. A form of non-randomizing the draw. $\endgroup$
    – o.m.
    Commented Sep 24, 2022 at 4:21
  • $\begingroup$ @hszmv it has nothing to do with removing the ball. in OM's idea, you very simply, say, "shake the machine a bit". In OM's idea, the authorities have "reacted to the prediction" and thus nulled the prediction. $\endgroup$
    – Fattie
    Commented Sep 24, 2022 at 17:03

What's the difference between prefect fortune tellers and counting cards?

People have been cheating casinos since time immemorial. Casinos have been working to keep the cheating in-house1 for that same amount of time. When all is said and done, your fortune tellers are no different than counting cards with a really good memory or using a communication device/computer to keep track of the statistics to bend the odds in the player's favor.

What does the casino do? The same thing they're already doing.

They run real-time statistical analyses against the winners to see if any of them are winning measurably outside the bounds of the (*ahem*) "random chance" the casino has built into their games. When they find someone who is winning too often, one of two things will occur.

  • In those casinos where corruption is kept remarkably in check, the cheaters are either invited to the door with a warning to never come back or are turned over to the law where they are stripped of their dignity and thrown in jail for cheating.

  • In those casinos still enjoying the freedom to run their businesses without too much government oversight, the cheater is hauled out back to have their knee caps broken.

Problem solved.

Like all modern casino cheats, the really good fortune tellers will keep their winning to a practical minimum so that they come out ahead but don't get their knee caps broken. Greed may be good when it comes to business, but not so much when you're cheating a casino run by the local mob.

But in the end, your godlike fortune tellers will lose to the simple reality that casino owners go out of their way to understand statistics and probabilities.

It would do you a lot of good to research what modern casinos do to detect cheaters. It's not a small science — it's a massively involved science that involves everything from serious mathematicians to tech gurus to top-shelf psychologists. All that in an effort to understand the patterns of chance and behaviors of cheaters so well that people can't benefit from the kinds of skills you're suggesting. Similar analyses occur in stock and commodity markets to detect people with too much inside information. The simple truth is, casinos know exactly what random chance looks like and are very good at detecting when it isn't happening. It would be surprising if they didn't understand your fortune tellers better than the fortune tellers do themselves.

1All due respect to casinos. They're businesses in an entertainment industry. They're looking to make a consistent and predictable profit and they know very well how to do that: by controlling not-as-random-as-you-think chance to balance the thrill of winning with the common sense that there must be a thousand better places to spend your money. If you believe there's a significant amount of random chance in casinos (not individual games, but casinos as a whole), you need to do some research into casinos.

  • $\begingroup$ The difference is that counting cards gives a small edge which can be exploited over a large number of games (which casinos can then analyze and counter). But if I can predict the outcome, I can just place one single large bet on eg a specific number in roulette and make millions. The casino can't counter that, as there is nothing to analyze; it's just one bet. $\endgroup$
    – tim
    Commented Sep 24, 2022 at 19:09
  • $\begingroup$ And if you have multiple fortune tellers placing just one bet, the casino will go bankrupt quickly. For this to work, you'd need to limit what can be won on one single bet (blackjack would be fine, roulette not so much; at least, you'd need a cap on the bet amount). $\endgroup$
    – tim
    Commented Sep 24, 2022 at 19:10
  • $\begingroup$ @tim That's just an issue of quality and it doesn't change my point at all. Casinos watch for winners breaking their statistical expectations. And unless you're assuming fortune tellers didn't exist an hour ago, you're wrong about bankruptcy. Casinos would rapidly advance laws (or break kneecaps) to ensure that fortune tellers couldn't use their talents (or would be severely punished for doing so). Only the assumption of perfect prediction in a world that can't do anything about it causes a problem. The fortune tellers are cheating. Do you really think that would last very long? $\endgroup$
    – JBH
    Commented Sep 24, 2022 at 22:24

Just make your money off the location

Let the gamblers deal with the risk of cheats.

You just provide a location, some singers, food and booze, all at healthy markups, and have a 2% spread on the chips, that is, people pay 1.02 dollars to buy 1.00 dollars worth of chips.


Just change how the games work.

Fortune tellers would not be a big deal for lotteries for example. They would win the big pots and fleece other players with you... It is all about how payouts are set. For example pick 6 numbers. 20% would go to everyone who got all 6 rights, and lesser cuts from down there so that you end up distributing 50-70%. Still leaving healthy margin for the runner. This however means there would not be any guaranteed prizes.

Now casino and other games should be run like poker. People betting against each other or into a pool and the casino only taking a rake. That is cut from the pot. Zero risk of losing money from gambling part. Parimutuel betting for sports would be an option.

Now, getting enough players might not be possible. As most people would lose too many times.


To me the key is this:

A future prediction can also be wrong if a new event occurs (like hearing a fortune) and people change their actions.

If this is true, then making a lottery immune to fortune tellers seems easy enough. After all the tickets have been sold, a prediction is made of winning number. That result is removed from the possible set of results before making the draw.

While not applicable to all games of chance, a similar method could work for many.

The larger question, of course, is how valuable are predictions if the only way they come true is if noone acts on them?


fortune telling only works on predictable events

there is a fundamental difference between things that someone does not know and things that can't be known. for example, humans don't currently know how many planets orbit our nearest star, but with the right instruments we will eventually figure it out. in contrast, no one can know the position and velocity of any electron. that knowledge is unknowable. similarly, in your fictional world fortune tellers could predict things about the future magically that could be known with enough research, but cannot predict anything fundamentally unpredictable. so having someone guess how many candies are in a jar won't work. but having them guess the polarity of the next photon detected by your photon detector would work fine since it's an unknowable property. realistically, you don't need fancy equipment to do this. quantum uncertainty can be the dominant factor in any sufficiently chaotic physical system. honestly, a standard air-blown lottery ball machine would probably work just fine, since the quantum fluctuations in the air stream would compound fairly rapidly.

further reading: https://en.wikipedia.org/wiki/Uncertainty_principle


Don't make the future predictable to 100%. Prediction works like quantum mechanics, just probabilities. So if anyone would be able to predict to 100% accuracy, this represents a huge anomaly even in a world where everyone is able to predict the future. Prediction works with symbols, not numbers ... except the number is a symbol.


Bribe the police, and break the knees of any fortune tellers.

If anyone cheats in your casino, send some thugs to their house and beat the crap out of them, or their families. If they keep cheating, kill them and their families and bury them somewhere remote.

That has been the standard practice for crime ridden casinos. You can detect cheating with cameras and psychology and statistics models, along with sharing data with other casinos about anyone trustworthy.


Another perfect opportunity for quantum handwaving about the nature of consciousness!

I (i.e. the person writing this, me, fattie) can in fact predict the future. I'm incredibly good at it. (And indeed, I know others who are even better.) I'm often paid professionally to guess whether, for example, a startup will fail or succeed. (And again, there are others better than me at this.)

There are others who are incredibly good at predicting (say) the winners of certain sports events, and on average they have "magical" records statistically.

Note though that all these things depend on consciousness being involved.

If you toss a normal fair coin, I have absolutely no clue what the outcome will be.

You can see that in a more magical world, there would be fortunetellers (much better than me) who could predict really accurately a vast array of things.

But just as in the Schroedinger's Cat thought experiment, their skills would be not relevant to purely physical random outcomes.

(So, in your world, sports betting or horse racing would no longer exists - fortunetellers would ace it every time. But as you ask, gambling (dice games, etc) would behave normally.)


Games of chance would not work

Things like gambling and other games of chance only work when it's impossible for anybody to reliably engineer a win.

If everyone can reliably engineer a win, the people who run the betting will never be able to make a profit, and so they will close up shop and sell mustache wax instead.

Perhaps book-makers might experiment with a scorched-earth anti-scrying policy, such as:

Any game whose winner placed their bet based on fortune-telling will not pay out, and said winner will be permanently banned from this establishment.

So, imagine that you wait for the ponies to run, and then you know who the winners were, and then you consult your own fortune-teller to find out whether any of those people placed their bets based on a fortune-teller's advice; every person who did gets their money confiscated and is banned.

If that works, great! But if cheaters get more sophisticated, the book-makers will figure it out pretty quick just based on their payout rates, and if they can't invent new preventative measures they will again just have to get out of the business.

Regardless of what actual measures are taken by book-makers, game theory tells us that the continued practice of gambling will always be in grave danger because just one bad actor could end the whole enterprise by deciding to not abide by the rules.

The saying "this is why we can't have nice things" often comes up in contexts like this.

  • 3
    $\begingroup$ "when it's impossible for anybody to reliably engineer a win." is not the opposite of "If everyone can reliably engineer a win". It's not an either-or - you could have some people reliably engineering a win. And that can be controlled to still allow a business to operate. You can limit the bets and/or winnings, so even if, say, one person always bets and wins, the business still has a profit. You can adjust the system to different number of people. Now, the gambling business would still not like getting less profits but it's not a sink or swim situation. $\endgroup$
    – VLAZ
    Commented Sep 23, 2022 at 6:43

There are two possibilities: either the universe is purely deterministic and you can make reliable, accurate predictions of the future or the universe is not purely deterministic and your predictions are unreliable. There is no middle ground to be found here.

A future prediction can also be wrong if a new event occurs (like hearing a fortune) and people change their actions.

In a deterministic universe that can't happen. Everything that will happen is already set in stone and nothing you or anyone else can or will do is going to change the results.

In a non-deterministic universe - not pure chaos, but one in which future events are merely probabilistic - then sure, the introduction of the prediction can change the outcome. Which means that your predictions aren't particularly useful, since there are plenty of other apparently non-deterministic things that could also change the outcome.

Let's consider the weather for a moment. While it might not make much difference to you whether or not it rains on my car today, it will have a minor effect on me. It so happens that we have a leaky roof at work, so rain on my car means water getting into my office. That leak happens to be above the power distribution box, so if it does rain then we're all going home for the rest of the day. (Don't worry, we've got someone coming to address that, hopefully before the next thunderstorm hits.) If that happens a whole bunch of things will be delayed, our clients will have to wait for stuff to happen, some of them will be OK without but at least a few will have to put back their own plans and so on. Eventually the wave of after-effects will spread out to change the future path of any number of things in the world... all because of the weather over my office. If the weather is non-deterministic (which appears to be the case) then the effects of rain on my office are rooted in the non-deterministic and cannot be predicted until, at best, after it does or doesn't rain.

Predicting lottery numbers is like that, only worse. Billions of apparently insignificant things can affect the outcome, and many of those things are directly or indirectly non-deterministic. Roulette? Pfft. The direct data required to accurately predict one fall is insane. Not only do you need to know the precise amount of force imparted to the wheel and the ball, you need complete information about the microscopic imperfections in the ball and the race, the precise consistency of the grease on the axle (which is affected by the exact temperature among other things), the exact structure of the bearings and how they are going to wear during the spin, the exact nature of the air currents, how the ball was positioned before the snap and which part of it is going to impact the pins... and on, and on. Every one of those factors can affect the drop and place the ball in a different slot, so any one of those that is directly or indirectly non-deterministic ruins your predictions.

But what about a simple game like Blackjack? Best case scenario, you have 6 sealed decks that are exactly on the nominal production specs, an evenly worn baize on the table and the dealer is slightly inexperienced. Now all you have to worry about is humidity affecting the friction coefficient of the table during the spread, tiny changes in the dealer's muscle tension and fingertip moisture during the riffle, etc. Oh, and let's hope that the dealer put exactly the right number of grains in her morning coffee, didn't get rained on while waiting for the bus to work, that her bus was exactly on time and so on. Because any tiny deviation from the situation you predicted is going to result in radically different outcomes.

Long story short (yeah, I know, too late) you can't have it both ways. Either you can make accurate predictions or you can't. Even really, really good predictions aren't going to cut it when large amounts of randomness are involved, be it dice, cards, spinning balls or lottery numbers. You might be hell on wheels when it comes to figuring out when your mark customer is going to meet her soulmate, or when a particular stock is going to dip, but if you want to throw money away there are thousands of casinos out there ready, willing and able to take it away from you.


This problem is self-solving if more than two fortune tellers try to predict the same event.

  1. Ball-skryer tries to predict an outcome of a game.
  2. Astrologist predicts the actions of the skryer.
  3. The Skryer predicts the actions of the Astrologist predicting his actions. 2+3 enters a recursive feedback loop, which the original cannot win in any way, other than forefeiting the heist and calling it a draw.

In fact, Fortune-Teller battles where BOTH fortune tellers are trying to predict the original outcome and battle each other through "I Know that you know that I know" stratagems would be a betting event of its own. Nobody sane would even try to use prescience to predict a battle of prescients recursively reading one-another's future, they would go insane.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .