Sending Info through time

I just read a book in which the hero dies and wakes up as himself as a young man and this cycle repeats several times. At first he is able to get rich and attempts to change history based on his general knowledge and in subsequent lives memorizes much info to take with him after his next death. I have been trying to think of a way that large amounts of information could be keyed to documents which would be available in an earlier time rather than trying to memorize it. For example all the winners of the first race at a race track for a year might be listed as a 20 digit number which (perhaps?) I could find in a book of logarithms by simply remembering the page and line numbers. Or maybe not. Any ideas?

) I just read a book in which the hero dies and wakes up as himself as a ) young man and this cycle repeats several times. At first he is able ) to get rich and attempts to change history based on his general ) knowledge and in subsequent lives memorizes much info to take with him ) after his next death. I have been trying to think of a way that large ) amounts of information could be keyed to documents which would be ) available in an earlier time rather than trying to memorize it. For ) example all the winners of the first race at a race track for a year ) might be listed as a 20 digit number which (perhaps?) I could find in ) a book of logarithms by simply remembering the page and line numbers. ) Or maybe not. Any ideas?

Woudn't it be much easier to search for the 20 digit number somewhere in the decimals of PI, and then just remember that it's '20 digits, starting from the N-th digit of PI' ?

Or maybe you could find the prime factors of your 20-digit number, and then just remember which primes they are ? (You only need to remember that it's the N-th prime, so N is a lot smaller than the prime itself)

(Gotta tell this to the boys over in comp.ression. Could use a good laugh)

sammacel did tell us: : I just read a book in which the hero dies and wakes up as himself as a : young man and this cycle repeats several times.

Is this _Replay_? Cool book. But it shouldn't be that hard to actually remember enough winning horses, stocks to buy and dates, etc. outright to make a significant difference, should it?

You can't map *all* 20 digit numbers to a page number that is shorter that 20 digits. That's everything.

If you want to compress the winning horses, you could include their odds in your calculation. (see: arithmetic compression)

very likely N would be a number with more than 20 places then. And it is not known, if every natural number can be found in PI.

Encoding the information as a prime first makes the encoded string longer, since not every number is a prime. This makes N also very likely to be larger than the encoded information. Prime factors don't help either, because the sum of their places is always >= the number of places of the product. (11x13 = 143 -> you remember 4 places to get 3)

The number of 'bits' required to store the location/type of book, the page number, and the line numbers would be at least as large as those require to store an arbitrary 20 digit number. The man might find it easier to remember the number using a book, since it is less abstract than a pure number (although learning to use a abacus helps numbers become less abstract). There are memory techniques that are based on associating numbers with objects and then remembering the objects by creating a memorable story. These techniques (e.g., 'link and peg' are described in such books as _The Memory Book_ by Harry Lorayne and Jerry Lucas (I'd describe more about the techniques, except I can't remember where I put my copy of the book ).

Betting on sporting events isn't likely to work. Just making a bet that didn't occur 'previously' could change the outcome of a horse race (a large bet will certainly change the pay-offs). Over several weeks chaos will have a strong effect. Just making or not making a bet on one racing day could change the weather on future racing days. A change in the weather could certainly change the outcome of a race. Events based entirely on chance, like a lottery, would change even more. The man might do better if he patented and marketed key inventions, but it is possible that an invention may not be as popular as it was 'before' (e.g., a Rubix cube). Investing in natural resources, like oil, would be one of the safest bets. The man could memorized the locations of significant oil or gold deposits (actually visiting the sites would make them easier to remember). Chaos wouldn't affect the locations of such deposits.

Carl G.

No. The odds are that N will be also a large number, probably 20 digits too. You might as well remember the original number.

This may save you a few digits, but it won't make much of a difference. Try it.

The question is: how much information do you really need? Why pick horses in races when you can remember a winning lottery number? Why not invest in a trend before everyone knows about it (i.e. buy internet stocks starting in 1994, sell in early 2000). In this case, compression would mean remembering a small number of extremely significant events. The key is to pick the right set of events. E.g., pick the one stock that will perform best over your target period.

If you had enough computing power, you could try to write a string of information that an existing algorithm would encode into another widely known or easy to remember string (a famous book, a mathematical formula that would recreate your string, etc). You could tweak your string without altering its meaning in order to maximize the possibilities. Still, even with an insane amount of computing power, intuition tells me that you would have to get very lucky. The order of complexity of this problem could be unmanageable.

Although it might make for an interesting SciFi story where the hero plans and does everything right, yet some little mishap changes the value of the deposits ... maybe a nuclear accident irridates the minerals?

<snip>

The classical example story is that of the man who went back in time to 'hunt' dinosaurs and stepped off the designated path, killing a butterfly. Returning to the 'present', he finds the world a very somewhat different place.

Which brings up a question- wouldn't the change (if there indeed is any- see below) be rather subtle? Wouldn't it take a long time to become apparent? Like years or decades? I don't see how my winning the Pick-Four lottery at Noon can cause me to get/not get into a car accident at 12:10.

And then again, maybe the 'real' timeline already _included_ the 'changing' event. Time travellers don't _change_ the past, they _fufill_ it.

Damn. Where's my copy of 'All you zombies...'??

Fred Klein

Yes that was the book. I realize you could easily memorize enough to make tons of money but the idea of encoding information somehow instead of carrying it with you is interesting. Perhaps it is impossible but maybe not. I still don't know.

You can really remove all the context from this question and reduce it one about data compression/encoding which is a rather exhaustively studied field. But the bottom line is that if you know nothing a priori about the distribution of data that you wish to compress then you can't, on average (a term which is vague as I've written it but can be defined precisely if necessary), achieve any better result with a compression scheme than by just leaving the data unencoded. In fact you do worse because not only do the strings not get any shorter but you have to remember your encoding scheme as well. Look up topics like information theory and Shannon entropy and you'll find lots of interesting stuff pertinent to your question.

Just do the groundhog day thing and kill yourself multiple times, that's be cool. I think going through the stages repetitively, you eventually be able to remember it easily, as you'd have seen the same race many many times, and the winner would no doubt be in the back of your mind someplace.