←2020-02-22 2020-02-23 2020-02-24→ ↑2020 ↑all
00:19:39 -!- tromp has quit (Remote host closed the connection).
00:33:57 -!- b_jonas has quit (Ping timeout: 272 seconds).
00:48:34 -!- longname has quit (Read error: Connection reset by peer).
00:51:27 -!- longname has joined.
00:54:21 -!- tromp has joined.
00:58:25 -!- tromp has quit (Ping timeout: 240 seconds).
01:27:23 -!- tromp has joined.
01:32:13 -!- tromp has quit (Ping timeout: 272 seconds).
01:57:18 -!- longname has quit (Read error: Connection reset by peer).
02:00:34 -!- longname has joined.
02:29:29 -!- FreeFull has quit (Quit: Rebooting).
02:30:54 -!- FreeFull has joined.
03:15:17 -!- tromp has joined.
03:19:53 -!- tromp has quit (Ping timeout: 272 seconds).
03:28:40 -!- xkapastel has joined.
03:42:47 -!- tromp has joined.
03:44:12 -!- tromp_ has joined.
03:47:45 -!- tromp has quit (Ping timeout: 272 seconds).
03:48:47 -!- tromp_ has quit (Ping timeout: 265 seconds).
03:54:33 <tswett[m]> I feel a theorem coming on.
03:54:37 -!- LKoen has joined.
03:56:33 <tswett[m]> Suppose you have a stock market moving according to geometric Brownian motion. You can buy and sell stock, and borrow and repay money (with interest charged), with no transaction costs and no delays.
03:57:27 <tswett[m]> Theorem (?): You can make an arbitrarily large amount of money in an arbitrarily short amount of time with an arbitrarily large probability of success and an arbitrarily tight upper bound on the amount of money that may be lost.
04:00:15 <tswett[m]> Obviously this can't be done in the real world; it's a theorem about a flawed model, and the result isn't applicable to the actual stock market.
04:02:26 <tswett[m]> Here's the idea. Start by borrowing a dollar. Then buy one trillion dollars' worth of stock. Borrow and repay money constantly so that the amount of stock you own is always equal to one trillion times your net worth, disregarding the dollar you borrowed.
04:04:25 <tswett[m]> After a second, you'll have either an extremely large amount of stock or an extremely small amount of stock. If it's an extremely large amount, great job; sell it and retire. If it's an extremely small amount, borrow another 50 cents and repeat.
04:07:53 <tswett[m]> I might be falling prey to some kind of gambler's fallacy here.
04:20:23 -!- LKoen has quit (Quit: “It’s only logical. First you learn to talk, then you learn to think. Too bad it’s not the other way round.”).
04:34:33 <imode> currently, I am creatively and motivationally dead.
04:34:46 <imode> halted all my projects. spent most of the day in bed.
04:35:03 <imode> here's to those of you who are enjoying your time.
05:30:52 <ais523> tswett[m]: thinking about the Martingale gambling strategy (i.e. always bet enough that a win would cover all your losses so far; note that this doesn't work in the real world either), I realized that casinos effectively run Ponzi schemes
05:31:07 <ais523> if every better had infinite capital they would all make a profit in the long term, and so would the casino
05:32:55 <tswett[m]> Do you mean they're Ponzi schemes at the bettors' expense or the investors'?
05:38:45 <tswett[m]> Let's see, I'm trying to figure out if my theorem is actually true or not.
05:39:42 <tswett[m]> Suppose you try to do the same thing with, say, roulette. Start with a dollar and bet 99% of your money each time; stop when you have either a billion dollars or a billionth of a dollar.
05:40:17 <tswett[m]> Well, that's not a useful strategy because you have less than a one in a billion chance of getting to a billion.
05:40:46 <tswett[m]> Let's say it's a 50-50 coin flip game instead of roulette.
05:43:03 <tswett[m]> You have a 50% chance of losing 99% and a 50% chance of winning 99%. But there's a lot of compounding burden. If you look at logarithmic percentages, the loss is 461%L, whereas the gain is only 69%L.
05:44:47 <tswett[m]> You're going for a billionfold increase, which means you want to gain 2072%L before you've lost 2072%L.
05:45:39 <tswett[m]> But each play loses you about 200%L on average, and it's very unlikely that the variance will make up for that average loss.
05:46:10 <tswett[m]> Okay, suppose you try to reduce your compounding burden by making much smaller bets. Let's say you only bet 1% of your money at a time.
05:46:25 <tswett[m]> Of course, you still have less than a one in a billion chance of getting to a billion.
05:48:06 <tswett[m]> Now a win gives you a 0.995%L increase and a loss gives you a 1.005%L decrease. You're still losing about 0.01%L on average, and now the variance is much, much less.
05:49:11 <tswett[m]> So what happens if you take the limit as the bet sizes approach zero? Well, the expected value of the game is 0 no matter what your bet size is, so even with an infinitesimal bet size, the expected value remains 0.
05:53:25 <tswett[m]> So now if you apply that to investing in geometric Brownian motion with leverage... I don't know what that limit looks like.
06:07:12 <int-e> imode: Welcome to IRC, the place for people who have trouble moving on in their life.
06:10:39 <imode> int-e: you know, I've been on IRC for 10 years. and you're right. and I hate that.
06:11:47 <int-e> Well, about 15 years here. I kind of know what I'm talking about.
06:12:03 <imode> what happens when we die.
06:12:37 <int-e> Earth keeps moving. We actually have a good chance of going out with a (nuclear) bang.
06:12:54 <int-e> `? marvin
06:12:57 <HackEso> marvin? ¯\(°​_o)/¯
06:16:23 <imode> I just realized my first few messages almost made a cute rhyme.
06:17:00 <int-e> a rap song in the making, perhaps.
06:17:34 <imode> bars.
06:17:46 <ais523> hmm, this place got so depressing while I was gone
06:18:02 <ais523> imode: for what it's worth, I've been there too, and I shook myself out of it eventually
06:18:14 <ais523> now I'm there again now, but still managing to at least get my day job done, and hopefully I'll recover soon
06:18:25 <imode> the oscillation hurts.
06:19:14 <int-e> . o O ( Look on the bright side: Without depression, we would have destroyed Earth centuries ago. )
06:19:22 * int-e should probably stop talking.
06:26:34 <int-e> ais523: I feel that one important difference between Casinos and Ponzi schemes is that Casinos don't punish late "investors"; everybody gets the same (somewhat lousy) chances of making a profit, except for the people running the scheme, whose chances are much better (that's where the analogy works again).
06:27:34 <ais523> nor do Ponzi schemes, intentionally
06:27:36 <int-e> (But don't ask me why I felt I had to capitalize casinos.)
06:27:55 <ais523> it's just that when they inevitably run out of money or get found out, late investors are the people who end up screwed
06:28:17 <int-e> Well, Ponzi schemes inevitably run out of investors. And often you get an exit scam on top.
06:29:04 <ais523> I guess the difference is that a casino's advertised average rate of return is negative
06:29:11 <ais523> so they accumulate money over time rather than losing it
06:29:26 <int-e> (Which is where the people running the scheme run off with all the money, and later turn up shot dead or swimming with the fish in a remote place.)
06:29:50 * int-e may have skipped a step or two there.
06:30:32 <int-e> I suppose you can put it that way, yeah.
06:31:35 <int-e> Also, casinos are far more romantic.
06:31:48 <ais523> in the UK, gambling machines have to advertise their rate of return on the outside of their case
06:31:53 <ais523> (it's normally around 90%)
06:32:11 <ais523> I wonder if that means that they can't be games of skill to any extent, as that would influence the rate of return?
06:32:29 <ais523> at least, from what I've seen of other people playing them, the choices it offers seem to be rigged to provide the same return no matter which choice you make
06:32:56 <int-e> Does it have to be exact or a lower bound?
06:33:29 <ais523> I don't know
06:33:49 <int-e> And yet it makes all the difference for this question. :-P
06:34:07 <int-e> (That is, whether you are allowed to reward skill.)
06:40:53 <int-e> I guess the people most interested in a stable payout rate are those running the machines.
06:41:26 * int-e shrugs.
06:42:10 <zzo38> I think that it is designed so that it can't pay out more than a certain percentage of all money put in, so that will ensure the rate of return is what it says it is, I think. I think that is what I read somewhere, at least.
06:46:02 <zzo38> I should think, they should fix it so that it is permitted to reward skill if the machine designer wishes to do so, although it is unclear how the rate of return should be advertised in that case. Do you know?
06:52:48 <ais523> I don't know
07:30:40 -!- Lord_of_Life_ has joined.
07:31:38 -!- Lord_of_Life has quit (Ping timeout: 268 seconds).
07:31:59 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
07:33:49 <int-e> Rewarding skill is allowed... apparently incluindg ripping off unskilled players: "THE RETURN TO PLAYER BASED ON BEST STRATEGY IS (VALUE) %" http://www.gamblingcommission.gov.uk/for-gambling-businesses/Compliance/Sector-specific-compliance/Arcades-and-machines/Gaming-machine-technical-standards.aspx (for example Section 8.3 in the B2 version.)
07:35:06 <int-e> Obviously this is for the UK (and not straightforward to find... the primary law is the Gambling Act 2005 which is easy enough to find, but the details are filled in by regulations such as these.
07:43:22 <zzo38> Yes, if it specifically says "THE RETURN TO PLAYER BASED ON BEST STRATEGY IS" then it is clear what it means.
07:46:59 -!- ais523 has quit (Quit: quit).
07:53:36 -!- sixyears has joined.
07:57:22 <sixyears> good eve
08:04:27 -!- tromp has joined.
08:06:23 <esowiki> [[Clue (oklopol)]] https://esolangs.org/w/index.php?diff=70026&oldid=69965 * IFcoltransG * (-24) /* Syntax */ Clarified, maybe guessing at meaning
08:42:03 -!- imode has quit (Ping timeout: 260 seconds).
09:08:18 -!- sixyears has quit (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )).
09:23:33 <int-e> . o O ( such silliness: M -= M.Identity().diagonal().asDiagonal(); )
09:46:59 -!- arseniiv has joined.
09:56:51 -!- LKoen has joined.
10:14:06 -!- LKoen has quit (Remote host closed the connection).
10:16:05 -!- LKoen has joined.
10:24:34 -!- LKoen has quit (Remote host closed the connection).
10:35:27 -!- LKoen has joined.
11:01:15 -!- LKoen has quit (Remote host closed the connection).
11:05:35 <esowiki> [[Mindwhipper]] https://esolangs.org/w/index.php?diff=70027&oldid=70024 * Asasnat * (-167)
11:12:01 <esowiki> [[Mindwhipper]] https://esolangs.org/w/index.php?diff=70028&oldid=70027 * Asasnat * (+18)
12:28:12 -!- xkapastel has quit (Quit: Connection closed for inactivity).
12:51:20 -!- LKoen has joined.
13:16:39 <esowiki> [[Nybblang]] M https://esolangs.org/w/index.php?diff=70029&oldid=70022 * PythonshellDebugwindow * (+12) /* Program import */
13:41:41 -!- LKoen has quit (Remote host closed the connection).
13:43:21 -!- LKoen has joined.
14:33:00 <olsner> int-e: incidentally I spent most of this week working with (against?) Eigen
14:34:38 <int-e> olsner: I was just playing around. However, I made the mistake of looking at the machine code resulting from the naive M -= M.Identity()... and it had two nested loops :-/
14:34:51 <olsner> to modify a diagonal you can also do something like M.diagonal() -= 1; (but probably with a random extra .array() somewhere, and maybe it randomly doesn't work with scalars at that place)
14:35:26 <int-e> Ah, that looks more proper actually.
14:37:14 <int-e> Yeah, M.diagonal().array() -= 1 works.
14:37:54 <int-e> Thanks :)
14:42:25 <int-e> I find the documentation of Eigen hard to digest as well, btw. The problem being unfolding the abstraction to find out what methods actually exist.
14:43:27 <int-e> For example I almost missed Matrix::row(int) and Matrix::col(int).
14:45:37 <esowiki> [[CopyPasta Language]] M https://esolangs.org/w/index.php?diff=70030&oldid=69586 * Rerednaw * (+543) Examples and addition of the category Turing complete
14:45:40 <int-e> (Because it's defined in DenseBase)
14:45:52 <olsner> I tend to browse around randomly to find methods accidentally mentioned in examples :)
14:46:13 <esowiki> [[CopyPasta Language]] https://esolangs.org/w/index.php?diff=70031&oldid=70030 * Rerednaw * (-1) /* Mini Fibonacci sequience */
14:46:40 <int-e> Im also mildly curious how (in)sane auto b = decltype(M.col(0))::Constant(-1); is.
14:46:47 <esowiki> [[CopyPasta Language]] M https://esolangs.org/w/index.php?diff=70032&oldid=70031 * Rerednaw * (-11) /* Mini Fibonacci sequence */
14:46:56 <olsner> maybe the method to use is to just try stuff, like "I wish I could write X.row(foo) = bar" and then you find out both that row() exists and that it can be used as an lvalue
14:47:00 <int-e> (To get a compatible column vector with all entries equal to -1)
14:47:00 <esowiki> [[CopyPasta Language]] M https://esolangs.org/w/index.php?diff=70033&oldid=70032 * Rerednaw * (+1) /* Mini Fibonacci sequence */
14:49:08 <int-e> olsner: Unfortunately I don't have that kind of faith in library designers.
14:50:53 <int-e> olsner: But I was actually clued in on this by an example... which pointed me to the block interface, of which row() and col() are members.
15:24:42 -!- ArthurStrong has joined.
15:28:04 <ArthurStrong> Hi all. What is the name of problem of finding shortest path in a graph that has no Hamiltonian path? Like TSP, but edges can appear >1 in resulting path
15:49:21 -!- Phantom_Hoover has joined.
17:28:56 -!- kspalaiologos has joined.
17:35:59 -!- tromp has quit (Remote host closed the connection).
17:49:09 -!- tromp has joined.
17:49:40 -!- kspalaiologos has quit (Quit: Leaving).
18:09:22 -!- imode has joined.
18:29:25 -!- tromp has quit (Remote host closed the connection).
18:31:25 -!- fizzie has quit (Quit: Coyote finally caught me).
18:31:26 -!- zemhill_________ has quit (Remote host closed the connection).
18:33:22 -!- fizzie has joined.
18:36:18 <fizzie> ...restarted the wrong system. My own fault for not putting something more obvious in the shell prompt, I guess.
18:39:00 <int-e> Yay, reboot.
18:39:40 <int-e> It's way more fun with halt/shutdown -h/poweroff
18:40:57 <fizzie> I have a sinking feeling getting zemhill back up and running is going to be a headache. Fortunately, it's not really seen any use for some months.
18:58:23 -!- tromp has joined.
19:29:27 -!- Lord_of_Life_ has joined.
19:32:40 -!- Lord_of_Life has quit (Ping timeout: 255 seconds).
19:32:43 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
20:29:14 -!- b_jonas has joined.
20:40:13 -!- LKoen has quit (Quit: “It’s only logical. First you learn to talk, then you learn to think. Too bad it’s not the other way round.”).
20:54:25 <b_jonas> olsner: you can come to #eigen and ask questions, it's a friendly channel, or at least used to be a few years ago when I used eigen a lot
20:56:08 <b_jonas> fizzie: I actually have a shellscript to avoid this: (sudo kingreboot) reboots the machine only if its hostname is king
20:58:49 <b_jonas> also, it appears that twitter has changed their website again, so it now doesn't even load anything without javascript. it used to load a few pages long part of a thread, just hid it under an overlay that I could remove with user css
21:03:01 <zzo38> b_jonas: The mobile version still works
21:03:19 <int-e> b_jonas: Yes, they have some recent changes. So now I'm using the mobile version, which works reasonably well without JS, but only lets you in with cookies. Can't win :/
21:05:04 * kmc eats all the cookies
21:05:14 <int-e> @botsnack
21:05:14 <lambdabot> :)
21:06:10 <int-e> kmc: Don't worry about surveillance on the Internet. Here, have a cookie.
21:06:29 <zzo38> Mobile version probably isn't a problem if you only want to read and not write, though.
21:06:58 <zzo38> (Mobile version also works even if you do not have an account.)
21:07:30 <int-e> (I don't have an account.)
21:14:41 <b_jonas> I'll try the mobile version, thanks
22:53:35 <esowiki> [[Garbage]] M https://esolangs.org/w/index.php?diff=70034&oldid=68229 * IFcoltransG * (-6) /* Syntax */ Wording
22:56:22 -!- zzo38 has quit (Disconnected by services).
22:56:26 -!- zzo38 has joined.
23:09:06 -!- Phantom_Hoover has quit (Ping timeout: 258 seconds).
23:14:45 -!- Sgeo has quit (Read error: Connection reset by peer).
23:15:11 -!- Sgeo has joined.
23:29:48 <esowiki> [[Semper dissolubilis]] N https://esolangs.org/w/index.php?oldid=70035 * Hakerh400 * (+10387) +[[Semper dissolubilis]]
23:29:51 -!- xkapastel has joined.
23:29:59 <esowiki> [[User:Hakerh400]] https://esolangs.org/w/index.php?diff=70036&oldid=69979 * Hakerh400 * (+26) +[[Semper dissolubilis]]
23:30:26 <esowiki> [[Language list]] https://esolangs.org/w/index.php?diff=70037&oldid=69978 * Hakerh400 * (+26) +[[Semper dissolubilis]]
←2020-02-22 2020-02-23 2020-02-24→ ↑2020 ↑all