I'm a tinkerer, working in my garage on a time machine. I had some success.

Saying "I built the time machine" would be too much. I built the receiver. It can provide exactly one bit of information, 0 or 1; building the unit cost me about \$10k, it takes a lot of electricity, about \$100/day worth, and it needs to be continuously powered at least until the value of the bit of information is set (in the future) using the "transmitter" module - which I have a good hunch I can build, but still no good clue, how. It's gonna be expensive, considerably more so than the receiver, though I still don't know how much... ten times maybe?

What's worse, if the receiver is powered off or breaks before the value is set, or the device breaks, the value I had read today is just a fluke of quantum uncertainty, worthless. Still, I'm willing to take the risk, and exploit what I have right now, to create the transmitter.

I have a common, stable job (university staff, non-teaching), no debts, earning roughly $30k/year and a month worth of savings; can set aside half of my salary, have no special connections, partners, resources.

I can build more receiver units providing I have the money to build and run them. One unit will take about a month to build in my garage, considerably less (and cost maybe half as much) if done "factory style", although setting this up would cost an arm and a leg. Each unit is another single bit of data, sent from unknown future, needing to be maintained until transmitter is created.

I also want to retain monopoly on the device. Never mind, until I have a solid proof it works, nobody's gonna believe it. My theory is sound, but the moment I publish it, bye-bye monopoly, never mind some idiot creating a hard paradox and breaking the universe (which my theory foresees; I absolutely don't intend to create the hard paradox; meanwhile soft paradox is fully allowable: I can use data from the future, which I obtained thanks to that very data.)

Once I have the transmitter, it's gonna be a breeze, a singularity. Put the unit in a hotel room above a casino with roulette, 0 black, 1 red, always bet 1/35th on 0, roughly 10 minutes to nearly double my capital (always setting some aside in case of a fluke), then invest in more units and I can receive whole books from the future in no time, build automated systems that game the market, win wars and discover immortality.

But the problem is now and here. I need to decide what the data means, how to interpret it, to maximum advantage, so that my future self will be able to set it to the value I read today. And I need to develop a strategy to achieve the success before I grow old.

How can I succeed?

Edit: since the way the machine works causes a lot of confusion, let's take a simple example.

I already have a transmitter+receiver unit. I want to determine the result of a coin toss I will perform.

Scenario 1:

  • I decide "Heads=0, Tails=1."
  • I power the receiver unit on, then press "read". A LED by "1" lights up.
  • I toss the coin. It lands tails up.
  • I press the button labeled "1" on the transmitter, because that was Tails, and because I don't want to break the universe.
  • I can now power the unit off, or reset it and perform another experiment.

Scenario 2: (never happens)

  • I decide "Heads=0, Tails=1."
  • I power the receiver unit on, then press "read". A LED by "1" lights up.
  • I toss the coin. It lands heads up.
  • I power the device off, without touching the transmitter, because I don't want to break the universe ;) (I might also press 0 on the transmitter instead).

Thing is, if I definitely intend to press the button corresponding to how the coin lands, and carry out my intention, it will always land the way the readout indicated. I'm never compelled by "fate" to go with scenario 2, the readout will be correct unless, say, power outage happens or things other than me not wanting to break the universe force me to enter wrong value or power the device off early. In other words, if I press "1" on the transmitter, then most definitely the receiver had lit "1" before, and unless I intentionally try to create a paradox and change my prior intention, I'm never forced to create it:

Scenario 3: (can happen)

  • I decide "Heads=0, Tails=1."
  • I power the receiver unit on, then press "read". A LED by "1" lights up.
  • I toss the coin. It lands heads up.
  • I only get to mutter "I wonder what will..." when a lightning hits the nearby building and power goes out in the whole block.

Edit2: I can't get anything else out of device during the reception-transmission period. I can only either wait (delay transmission) or power it off (make reception random/worthless.) I can't enter or transmit any extra bit over the same device. Think of it as a single bit of computer memory (RAM), only write is performed after read, instead of before.

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$ Apr 5, 2017 at 3:14

8 Answers 8


There is no sure way to success, because the information you get is unreliable.

Strategy that might work: You borrow \$100k, use your receiver to make a successful bet and double your money amount, build another receiver to make even more money, repeat until you are a billionaire. Now you can build enough receivers to obtain full instructions on how to build the transmitter. Build it and send back all the info you've used before.

But it also might end like this: The info you received is wrong, you are in heavy debt, can't pay $100 per day to keep receiver on and turn it off. That's why the information is wrong.

Are there safer way?

Yes. Just don't use the receiver. Start saving money for building of the transmitter. If it costs ten times more than receiver, you'll need about seven years. Hope in that time you'll find out how to build it.

And what if I can't build the receiver without information from future?

Then there is still no sure way to success. Even if you can get enough money and build a lot of receivers they can all receive garbage. In this case the only information you've got is that you'll be stopped from sending back the data by some obstacle (like impossibility of building transmitter).

Why do you think I might not make it?

That’s because the only proof that you'll do is "a good hunch". Of course, if you've met yourself from future, which gave you the complete drawing of the receiver and said that that’s the only thing he can give you right now, but you'll manage - that's entirely different matter. In this case you could choose any strategy you want. You are guaranteed to succeed.

  • $\begingroup$ Since I'm pretty sure it works, the $100k loan way would work. How could I manage to obtain such a loan though? I don't need guarantee this will work, and I'm willing to risk everything, including getting ears deep in debt I'm unable to pay back IF I'm wrong. $\endgroup$
    – SF.
    Mar 31, 2017 at 17:17

The Problem

The real problem is that you can't even run your coin test to prove that it works without a functioning transmitter.

You could risk that you will, in the future be, able to build a transmitter and it function how you think it should; take a number of loans or other risky methods of getting cash, fire up the receiver and do the roulette or other 1 bit information gambles and use the return money to construct the transmitter to tell you which way to bet to win, but you really are just betting that you can build the transmitter.

Verification Bits

The safer bet would be to build a second (or more) receivers and use the first receivers to send random test bits to verify that it actually works before you mortgage the house to bet on black. The more receivers you construct and more test bits you have that align with whatever random phenomena you choose (coin toss etc.) the more likely you are to actually be able to build the transmitters and send the data back in time.

The Odds

Just using the one receiver you have to gamble, you have 50/50 odds on the gamble with no way of knowing if the machine actually works.

Using one test bit, you power on the machine and flip a coin or use an otherwise one bit random method you predetermined and if it aligns with what your receiver says, your machine could work, or it might just be random chance that it matched your test. The more tests you run (with more receivers) the more likely the machines actually work, with the down side of the costs involved having to build new receivers and keep them running indefinitely until you can build the transmitters.

The odds go up for your test signal with each unit. One test receiver gives 50% odds that the machine works vs. random chance, two receivers move it to 75%, three to 87.5%, four to 93.75%, etc. P=(1-.5^n). You would want to balance the odds against the increased cost vs. the decreased risk.

This would however be able to disprove that it functions on only one missed bit, if you ever set up a test where your machine says 0 and you flip tails, you would instantly know that it doesn't work, either you can't build the transmitter or it otherwise doesn't work how you think it should and you should cut your losses and find a cheaper hobby.

Other Considerations

Also based on the high cost of maintaining a receiver powered you wouldn't want to actually turn any of them on until you are ready to make your bet and start building the transmitters as fast as possible so you could turn them off and use them for another transmission.

Also single bet odds like red/black or odd/even in roulette have very low payouts, usually 1 to 1 (i.e. bet 1 dollar get 2 dollars [a dollar plus your dollar bet back]). Doubling your money is good, but it would likely be worth it to build a few more receivers to allow other higher return bets to increase your money pool to cover the costs involved while you're building the transmitters (cost of transmitters plus electricity cost of receivers).

100 dollars a day in electricity is quite a lot. Using $0.15 per kWh that works out to a daily average of ~28 kW constant power. The average US house has a 200 Amp service at 220 V providing a max of 44 kW, although not really expected to provide this max continuously. In other words if you were to use a standard 110V plug on a normal 20 Amp circuit you would need 13 plugs each on a different circuit to power one receiver. This is not something that you can just plug into a hotel room outlet, and the power company would notice the massive power draw pretty much instantly, you would need a specialty power connection and probably a rented industrial space for these types of power draws.

Don't forget the backup power systems, it would be a great story if after winning the long odds bet a power loss shut down the signal before you could build the transmitters, or after attempting to build the transmitter the inventor realizes his theory was wrong and time travel is impossible, he just beat the impossible odds.


Your problem is, you're trying to use your fully-functioning Chaos Machine like a half-baked Time Machine.

If, going from the Case 3 you describe, the timeline arranges for chance events to occur that preserve its own integrity, you've constructed a machine capable of manipulating the very laws of probability! It will take some engineering; build several redundant copies of your device, get as many backup power sources as possible, and hook all the controls and readouts into a very reliable computer (or set of computers, a la the redundant systems used in high-reliability aerospace conditions.) Hook those computers up to the internet.

Now, decide something you would like to have happen. It can be anything that meets two criteria;

  • You can program your computer to deterministically detect that it occurred.
  • The probability of it occurring is greater than the probability of your whole setup breaking somehow.

For instance, say we decide to program the computer to check our bank account, and give a signal if it's somehow over a million dollars. Turn the receiver on; either a 1 or a 0 is received from the future. We've programmed the computer to take in the receiver bit and the bank account signal and transmit either A) the same bit that was received, if our bank account is now fantastically gigantic or B) the opposite bit, if we're still of modest means. Since B would break the universe, the universe will squirm and instead somehow make A happen.

(Or the universe will call your bluff and just break, maybe. This isn't a business for the weak-hearted.)

What condition to wish for with this technique is, of course, its own complicated question. Tying it to your bank account is just asking for an accounting error that would get swiftly charged-back. Having it scan for the name of your rival in the obituaries would just lead to the belated realization that, really, you loved him all along. Who knows what would happen if you tried to use it to get a snow day from work. But there are endless possibilities!

BONUS: Wait, forget what I said about all the reliability engineering. There's an easier way; just set up two of your machines, each in different cities. Alongside the program to achieve your dreams, program each one to transmit a paradoxical bit if the other one breaks. Now they're both effectively invincible! (And while you're at it, perhaps invest in an internet-enabled heart monitor...)

  • $\begingroup$ The thing is, the universe doesn't try to work around the paradox. It's just fully deterministic. It's not that the lightning happened because of me and my device, it's just that false prediction means something will happen in the future that makes the machine to fail; scenario 3 is not impossible - but very unlikely, about as unlikely as a random lightning hitting there and then. 99.999% of the time I'd be getting scenario 1. $\endgroup$
    – SF.
    Mar 31, 2017 at 17:23

Not such a breeze though - as soon as you are winning too regularly you will be noticed, and that avenue of revenue will be cut off. General discovery of the fact you have a time machine will also put an end to convictions in non-determinism which means an end to gambling, the insurance industry, and civilisation as we know it.

Of course as soon as you have developed the time machine, you have sealed off non-determinism pretty much completely - (because nothing can change until the information you are sending is guaranteed to happen) unless your information is unreliable. If it is unreliable, then lottery tickets won't help anyway.

So, it's not just about a business monopoly. Basically, there's a trumping issue. Whoever makes a specific prediction first now has control of the entire universe up until that event. The further away that prediction the longer that control is sustained.

Ie, you have just become a god. Even with your one bit of information.

The easiest, and by far and away the most reliable, means of making a shedload of money is to find a very rich business partner. Then you can leverage the cost of manufacturing and running a warehouse of devices.

Then you would just want to receive a plaintext ticker-tape - why? Because you have NO idea whatsoever what will be of value to you now - only your future self will be able to tell you. If your future self tells you to use a compression method or encryption method, then you can follow instructions.

Now, chain that. So your future self is now receiving messages from a future self, and passing them down to you. As long as there are successors (or an end to mortality) you have established (and locked the universe into) a (chained) link to a very deep future.

In the end though, your life is over - you will become a 'wirehead' - fixated by the very next message that comes from your devices - you have lost any freedom and will to an endless chain of future selves who will determine your actions.

What sort of life is that?!


So, one machine alone isn't enough for anything. However, let's say you batch your machines into groups of six. Now you can run a parity bit check on the machines as soon as you turn them on. If the parity bit fails, you can re-use those machines, so you have semi-reliable data transmission.

The five machines (labelled) ABCDE give you a plaintext alphanumeric (eg 0-space, 1-26=alpha, 27-31 digits 0-5 in base 5 (or something like that), whereas the sixth machine F gives you the parity bit.

A B C D E F 
0 1 0 1 1 x

Or with seven machines you could use a [7,4] Hamming code.

Once you have enough money to build a few groups of these, you can use a checksum such as Adler-32.

This is when it gets weird though. Because the machine gets the result on start-up, we can engineer a chronological paradox to our favour.

As soon as we get a result (on startup) we can test the checksum. If the message fails the checksum, we just reset the machines.

If universe is not many-world universes, then once the message passes the checksum, we can also reset the machines, because we have now got the message that was sent to us. This sounds completely impossible based upon your initial conditions, and it DOES depends upon which philosophical universe you want to choose: We now know that the message is reliably from the future (discarding the tiny percentage of clashes which would reveal themselves in the actual content of the plaintext).

So if we turn the machines off, are we changing the universe to another (non-determined) one? Many-worlds would allow for that.. ie, we got a valid message, but by turning the machines off, the message is valid for another universe, and not this one...

Otherwise, we can just save on the electricity bill and re-use the machines having got a valid message from this universe.


(0) Take out a mortgage on your house $100k. Build the transmitter. You now have a 1-bit transmitter/receiver.

(1a) You start with betting on eg. Betfair on binary outcoming sports events, making the bet (turning the machine on) just before the event. Eg. Oxford will win the boat-race. You bet as often and as frequently as you can on binary outcomes. You could also use binary trades on the stock market

(1b) You write software to find, and place bets on, such events as above - which can (a) find a short-term binary bet (b) turn the machine on and read the result (c) make the bet (d) turn the machine off.

(2) You plough your winnings into more winnings until you reach the point where you can start investing in more machines. You continue to use eg. Betfair for an income stream.

(3) You now generate groups of 7 machines for [7,4] Hamming code text messages. You use those results to make longer term financial decisions and advice.

(4) You now generate clusters of machines that accept strong checksums.

Why you cannot use the receiver without the transmitter

Until the transmitter is available, the receiver is unreliable.

I have a receiver.

My monthly disposable income is $1,250, available in advance.

The monthly cost of running the receiver is $3,000, payable in lieu.

Let's say that I am committed to running the receiver for the next month. I have $1,250 in holdings (from my previous month's disposable income). I believe that it is likely that I will construct the transmitter this coming month.

I need to find a bet that I can cash in on as soon as possible. It must be a binary bet (eg, Oxford will win the boatrace) because I only have one bit available.

I need 100k dollars to build the transmitter. That's 8,000 percent my disposable income. I will also need to pay the $3,000 power bill, but let's assume that I don't finish the transmitter on the last day of the month, so I can make a second bet to pay that off.

So I need to find an 80:1 bet on a binary outcome event

Such bets are incredibly rare. Binary stock market returns never return 100% - normally the best one can find are around 80%. I need one hundred times those odds in my favour.

Also, if I lose, then I am going to be $3,000 in debt at the end of the month, and I know that I will have failed to develop the transmitter this month.

If I save my money, then next month I will have $2500 to bet, and only need to find a 40:1 bet to break even.

But even with those diminishing odds, I may as well save up (or use a normal investment path) until I can pay for the receiver, since if I lose, I lose the lot.

I only have to wait 6 to 7 years; that gives me an excellent chance of developing and researching everything I need to ensure that I can build the transmitter, with none of the drawbacks that the overpowering electrical bill brings.

  • $\begingroup$ First, I don't need to create the data capacity to wire myself into unchangeable "optimal" future, because yeah, what sort of life is that? I just guide my past self through smart advice. Next, energy intake is huge. And then there is STILL the bootstrapping issue. Never mind getting to the point when it's me sending the messages to my past self, not whoever assassinated me (or whole thing being a big fluke up to that point, the bits randomly arranging themself to whatever led up to my failure. $\endgroup$
    – SF.
    Mar 31, 2017 at 12:47
  • $\begingroup$ Yes, SF. Great point re. how to authenticate the messages. As for energy expense, etc., a rich 'investment angel' would resolve that. I don't really see what the difference is between being wired to an 'optimal' future, and being guided by 'smart advice' - both of them you are chaining yourself to decisions made beyond your ability to know them (because they are in the future). Plaintext helps to alleviate problems of data-reliability, whereas eg. lottery tickets would not. $\endgroup$
    – Konchog
    Mar 31, 2017 at 12:49
  • $\begingroup$ I'm afraid since non-determinism is dead, I don't really care - nor want to authenticate the messages (maybe save for my peace of mind.) If the assassin sent me the message, I can either get killed or break the universe. I must either be smart now, or in the future, so that I'm not assassinated. $\endgroup$
    – SF.
    Mar 31, 2017 at 12:52
  • $\begingroup$ Being smart, understanding the consequences, would you really want to have a time machine at all? Isn't it just one of those fantasies which appear really good in the imagination, but would/could have such dire consequences in the real world that an intelligent person would choose not to have it even if they could? It may not be the story you want to write, but I think that could be really good fiction. The person who chose not to be a god. $\endgroup$
    – Konchog
    Mar 31, 2017 at 12:57
  • $\begingroup$ Where does he get $50,000 to build five more receivers? $\endgroup$
    – Mike Scott
    Mar 31, 2017 at 13:37

Let's speed up this process

The receiver should probably be used to guide the creation of the transmitter. You have a tri-state here, so I'm going to suggest:

Forget about money for now. That will come later.

Instead, get to poring through your research books to find the approach that works. Don't even bother reading papers. Go to a list of physics articles, and devise a scheme to map time-of-day to paper. If this isn't "normal" physics, then you should instead make an exhaustive list of the ways you are thinking of making the transmitter. We are going to use the time of data transmission as a hack to get more information, along with the bit value itself. It sounds like you are going to be able to have very high temporal precision with your data-sending, so let's give ourselves a 1-minute window.

Now you have a mapping from time to "ways-to-proceed-to-make-the-transmitter". Your future-self already knows how to make the transmitter, and will have the mapping saved from his past (your present), so all you need is something to monitor the receiver and note the exact time (to the minute, or whatever temporal precision is appropriate) that the device receives the data transmission. Look this up in your map, think about it for a while, have the "Eureka" moment, then go off and build a transmitter (and another receiver). Now you can go to the casino. (Don't forget some day down to line to send a transmission back to the past to allow this all to happen!)

But what if the specific breakthrough needed to make the transmitter has not yet been invented, by you or another? Well then we need to refine our process. We can send a 0 or a 1. Let's make the plan:

Send a 0 if the breakthrough is from the map you created. Otherwise, send a 1 if the breakthrough is in a publication (book, article, news story, whatever) from that day - i.e., if you receive a 1, it's time to go to every physics journal, every news outlet, every book publishers, and find what was published today. Then again, have a think for a while, have your "Eureka" moment, build a transmitter (and another receiver), now go to the casino and make money.

You can, of course, speed up the process by adding more receivers - adding one more receiver increases the number of items you can map to each minute from 1 to 4 (binary). Having three receivers, your map increases to 16, four receivers gives 32 items, etc. Depending on how many items you have to sift through and how fine-grained you need your guidance to be (and how patient you are), you can spend the money accordingly.

  • 1
    $\begingroup$ There is no "temporal precision" bits are set when the machine is turned on. $\endgroup$
    – sphennings
    Mar 31, 2017 at 12:36
  • $\begingroup$ Doesn't seem so: it needs to be continuously powered at least until the value of the bit of information is set (in the future) $\endgroup$
    – Hugh Nolan
    Mar 31, 2017 at 12:38
  • $\begingroup$ It can provide exactly one bit of information. $\endgroup$
    – sphennings
    Mar 31, 2017 at 12:39
  • $\begingroup$ @Hugh: spehennings is right: if I power it off, the bit read from it will be random. I need to keep it powered on until I enter the bit I want to receive in the past. $\endgroup$
    – SF.
    Mar 31, 2017 at 12:39
  • $\begingroup$ ...in other words, if I want another bit, I need to build another unit. At least until I get the transmitter - after entering the bit, I can reset it and read another, from more distant future. $\endgroup$
    – SF.
    Mar 31, 2017 at 12:41

First send blueprint/manual of receiver upgrade.

Like qr code with strong/maximum error correction

Otherwise lottery numbers.

EDIT: Ok if u dont want lottery - you create list (or there is a some list online) with random companies (and numbers 1-15) and you will give everything you have left into one that will be very rich and famous in near future = you will make money.
Or you make correct bet who will win world championship of something. Again - teams in aplhabetical order (1-15).

  • $\begingroup$ Ah, of course. As soon as I reach the data capacity to obtain them! $10k per bit. $\endgroup$
    – SF.
    Mar 31, 2017 at 12:17
  • $\begingroup$ Error correction works by using extra bits to validate the integrity of your data. Extra bits = Extra money. $\endgroup$
    – sphennings
    Mar 31, 2017 at 12:19
  • $\begingroup$ Yea, but if "the value I had read today is just a fluke of quantum uncertainty" - depends if you know about it, or not. Otherwise o/c you don't need error correction. $\endgroup$
    – Jan Ivan
    Mar 31, 2017 at 12:21
  • $\begingroup$ Powerball: 6 out of 95 numbers total, 7 bits per number, probably can be squeezed down to some 5 bits per number through some smart compression, say, 30 bits - some $300,000... nope. Need to start smaller. $\endgroup$
    – SF.
    Mar 31, 2017 at 12:28

I think you have some problems:

It's 1 bit per device, per reception-transmission period. It's only 1 bit per minute if I transmit a minute after receiving. Considering I don't have the transmitter, it's more like 1 bit per year or more.

If this works like I think it does, you cannot send any meaningful data.

What I think will work: your receiver can receive one of 3 things: nothing, 0 and 1. You send this binary to a specific time (and maybe place if you want to parallel the process) in the past. Your future self will know how your past self will decode the incoming data. Your future self will also know when you have your device on. Just save the incoming data in binary, and you are golden. If the rate is very slow you send only numbers for lotteries, or other key:value data usable for gambling.

If you even want to eliminate the key data being send, keep notes in a book on what data you will receive at what time. If your sender time machine works you will have the data streaming in the moment you turn your receiving machine on.

You don't know at this time when stuff is being send, don't matter, as long as you can trust it... (I hope you neighbour ain't playing with you)

Your "receive" book will look like this:

  • 13th march 2017 - 19:00 - Who will win the Dutch general Election and how many seats? Election results.
  • 14th march 2017 - 20:00 - Will Trump repeal Obamacare on the 24th? Defeat
  • $\begingroup$ Not that easy; check the description again. The receiver can receive only 2 things, 0 or 1, and not at a specific point of time but upon being switched on (and must remain switched on until transmission, and can't receive anything else in the meantime.) $\endgroup$
    – SF.
    Mar 31, 2017 at 15:26
  • $\begingroup$ 1) What do you receive if the transmitter does not send? 2) If the data received is not bound by time, but by what sender does, at what ever time it sends, no meaningful data can be had.*always possible that I just don't understand you. $\endgroup$ Mar 31, 2017 at 15:33
  • $\begingroup$ 1) Randomness, 50% for 0, 50% for 1. 2) read the "scenario 1" in my edit (which, assuming my success, is the most likely one), now replace coin toss with roulette roll with high bet on red or black. $\endgroup$
    – SF.
    Mar 31, 2017 at 17:30

Latitude, longitude and a date.

This way your future self can decide what is the optimal place to send you based on future information.

This is going to be somewhat pricy to set up. GPS coordinates require two 64 bit values to be set up properly. You could use a 32 bit value to represent Unix time but that will overflow in the year 2038. It's best to use 64 bits for time just to be safe.

You could shave off bits by truncating values or using nonstandard formatting. Just be sure that you clearly document these decisions. You only have one shot at this.

  • $\begingroup$ Wait, what is supposed to happen at that location? $\endgroup$
    – SF.
    Mar 31, 2017 at 12:31
  • 2
    $\begingroup$ Good question. Your future self has decided that it is the best location to send you to. Show up prepared to figure out why. $\endgroup$
    – sphennings
    Mar 31, 2017 at 12:33

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