Show Me the 00100100

choice number one. Oddly, you're wearing the same thing in both scenarios, is that you're wearing a rumpled Oxford shirt that's been rolled up to the elbows. You've gotta tie that's a little, you know, dislodged, and your hair's all must up, and you're standing in the middle of a crowd of people all with sheets of paper in their hands, and you're making these crazy hand signals to some guy who stands in front of a bunch of numbers that keep on changing, and you just keep screaming out buy

and sell in random increments. Yeah, you're imagining me as an extra from the movie Wall Street, which was about the place Wall Street which is about money. Yes, yeah, you've nailed it. Actually, that's my personality. I write about science, technology in the future, but my real passion is international finance. Yeah I can, especially the part that involves standing someplace and shouting. Oddly enough, you would be doing the same thing in the swamp, just to whatever creatures happened to

be stopping by. Okay, So yeah, this isn't um the easiest topic for everyone to find something interesting about international finance, stocks, commodities markets. But it turns out that there are crazy machines involved in stocks markets. Yes, and as you read about these crazy machines, you start to envision a future, a future that's similar to say, the Terminator film series. So we wanted to talk today about that future and

about some of these machines and what they do. So before we get into the future, let's talk about the past. You guys know what happened on May six? Can you tell us? John? I can. I was waiting for response, which your your stunned silence tells me that I need to explain. That was the day. It was when the superhero the Flash smashed into the side of a mountain. Oh, the Flash crash. Okay, well, you got the name right, but but the particulars are a little different from what

you are envisioning. It didn't involve Barry Allen smashing into the side of a cliff. No, the flash crash was really about a bad day on the for the TAL the dal Jones Industrial average. It dropped about nine percent, which was around a thousand points total that day, and recovered most of those losses within the span of less

than an hour. Actually, but six hundred points of that one thousand point loss came in the space of five minutes, and it required about twenty minutes for the market to recover. And there are a lot of different reasons why the stock market dropped that day. One of the reasons was that Greece was undergoing that terrible debt crisis that was gripping a lot of Europe. Actually, Greece was just sort of the hotbed for it. And then there are other

issues as well. But some of it has to do with these crazy machines all acting on what they what they perceived to be market trends in mass and it began to kind of create an escalating effect. I guess it's not so much the machines that are crazy, but it's what the machines do. These your computers executing what we call high frequency trading. Yeah, which you might wonder, well, what does that mean? High frequency trading is pretty much

what sounds like. You're trading lots and lots and lots of shares throughout a day, like your shares are are. You're buying and selling uh, sometimes multiple times within a second. Yeah, and actually usually not that many shares that you usually keep the number of shares to a couple of hundred, but you're but you're flipping them tens of thousands of times a day in order to rack up these fractions

of its scent each time. Right, because stock prices are actually averaged out quite a few decimal places these days, so you could be talking about point zero zero zero five cents in a trade. But if you have enough of those over the course of a day that that starts to amount to some real money. Okay, so high frequency trading was partially responsible for the flash crash, but surely that's the only time something like that has happened. Oh, actually there is another blip. In by blip, I mean

like a hundred thirty six billion dollar market value blip. Um. By the way, I accept blips. If anyone has a hundred thirty six billion dollar blip they'd like to give me, also blimps. I am a Dirgible fan, so just throwing that out there, I actually will accept cheap Dirgibles here. That's true. That's true. I like to think of them as reasonably priced. But go on, okay, So in this instant, the APIs the Associated Press's Twitter account was hacked and it sent out news that bombs had gone off in

the White House and that Obama had been injured. Um. Of course that never happened, but programs that were monitoring the news sold off holdings within seconds and the market plummeted. These are these are two two separate but connected uh trends here, right, We've got computers that are following algorithms that are looking for market trends, like if they see something happening in the market, they pounce on it, whether that's buying or selling whatever stocks they happen to have.

And then the other trend is machines that are taking in information from the real world and acting upon it. So both of these things are happening in increasingly large numbers and frequencies in the United States in particular, in

other countries as well. We're not the only ones who do it, but investors of this kind sometimes call themselves quants, alluding to some kind of metaphorical quantum unit of money, which of course does not exist, but it's kind of cute um and and and the strategy here is too, is to yeah, flip all of these shares back and forth based on minute changes in the stock market thousands of times a day, and then remove all your funds from the market by the close of business. Right, so yeah,

you you get in and out every single day. It's kind of it's kind of mind boggling, and it's and it's operating at a speed that human people literally cannot compete at. Right. Yeah. No, that by the time you would see a offering on a stock, it would already be too late in this world, you know, too late a thousand times over. We're talking about machines that can that can make a decision in a fraction fraction of

a second, millionth of a second. And it might be responding to like we talked about something in the news, or some kind of external predictor or me responding to the just the market prices themselves, right right? Yeah, Now, according to one story I read in two thousand twelve, high speed computer algorithms, or al goes as some call them, I don't. I can live for about of all trading in US stocks, which sounds like a lot until you

hear other numbers. I've heard that estimated at as high as eight percent, and as much as for UK stocks for other European stock markets. Right, and so here we're talking about not just uh high frequency here, it's obviously all about speed as well. So it's not just how many times are your trading stocks throughout the day, but how quickly can you pounce upon a trend opportunity in

order to make that fraction of assent. So let's before we get into actually how that happens and what and what lengths people and companies have gone to in order to get that that advantage of speed. Maybe it's time to take a little step back, right, Let's let's turn the clock back and take a look at how stock markets even came into being and how they work in general, so that we can understand what's going on at these

incredible speeds. So if you were to go back to uh say, the Middle Ages, and um, you know, you know, dust yourself off a little bit and look around. The earliest examples of what would eventually become stock exchanges really grew out of the the money lending and later the banking movements that we're just getting started. Really, money lending had been around for a while, but banking was pretty darn new, um, and so at that point one of

the things that started to become commonplace was trading of debts. So, Lauren, let's say that you had borrowed a small sum of money and you fully intend to pay it back, and you are probably going to be able to, and I happen to hold the the debt the promise you're going to be paying me. Joe, let's say that you have borrowed a ludicrous sum of money and you have very little intent to pay it back, but no one knows that. Well,

I needed to invest in my alchemy lab. Okay, so you have gone into I turn all this lead into gold. You'll get your money back and then some So Joe has invested heavily in Al's chemicals, and uh so, Joe's money is potentially never going to be seen again. Now I have a chance, a pretty good chance of getting the loan paid off by Lauren. But someone else holding Joe's loan may say, hey, I'll trade you this this loan for that loan. Whereas you know, yours has almost

a guarantee of being paid off. Mine doesn't have a very strong guarantee, but the payoff would be much larger. And so you would have people start to trade those kinds of those debts, and this has got to be a pretty big thing. However, there was no there was

no formal place to meet to do this. You wonder, you wonder if the higher risk, bigger debts are more attractive to people who are better at extracting payment, yeah right, or less good at math yeah well, uh yeah, Like the Borgias were probably pretty darn good at it at the by the at their height. But then you had, around the the early fift hundreds the Belgian Stock Exchange, which traded in bonds and promissory notes. No promisory notes, that's essentially saying I promised to pay you X amount

of money. Yeah, it's it's a legal document, but it is essentially an iou uh. And then in shipping you started to see the people who owned ships would sell shares in a voyage, the idea being that by raising this money they could outfit the voyage and then if the voyage were successful, they would pay out profits from that voyage to their backers. Just very similar to the kind of stuff we would see and say a movie like The Producers, where you get this idea of selling

shares in a in a show. In that case, they were trying to uh sell shares and show that was guaranteed to fail, so they wouldn't have to pay anyone back. But this was this is the basis of what would eventually be come kind of again a stock market. And in fact, in London you get your first stock market in seventeen seventy three and nineteen years later that's when the New York Stock Exchange was formed. And so that

takes us to how these these systems work. So you've got buyers and sellers right, and you have a market there. Often the market is actually the entity that's holding stocks and is selling to you. Uh, you have a price which is guided by the market itself and basically guided by the amount of money someone was last willing to buy or sell a stock at. So if you want to buy stock. Joe. Let's say that you're raid by into Al's chemicals today, you know, because that that business

really turned around after the Middle Ages. I want to invest in lead. Okay, So Joe wants to invest in lead. He he thinks lead is going places. So Joe, you you say that you're willing to pay how much are you willing to pay for a share of lead? A dollar? So Joe willing to pay one dollar for a share of lead? And how many shares are you wanting to purchase in total? So Joe is willing to pay one dollar per share of lead for up to one hundred

shares of lead. Now let's say that you have a that that would, by the way, be called a bid. That that's your bid for a share for a hundred shares. Now the seller might say, you know, I'm willing to sell uh fifty shares of lead, but my asking price, my ask is a hundred a hundred dollars. So there's a big gap there. Right, you're asking you we're willing for a dollar, they're asking for a hundred. This would

be a bad what they call spread. But let's say as much you were you were being very realistic, or maybe maybe the price is much lower. We'll say that asking for a dollar oh five per share, Then you have a spread of five cents per share. It's not that big a deal. If it's the market that's selling this. The market ends up in kind of pocketing that five cent difference every time these trades are happening, and that's how the market makes money. And again, the stock price

is reflected by that last transaction. So if your transaction was the last one, and you actually bought say a hundred hundred years of lead for a dollar or five apiece, dollar or five would be what the stock price would be.

So that's the basis of this. So if you were able to act very very quickly, and and you saw that a stock price had changed to a specific amount, and you were able to leap on it right at that moment before anything any other changes in the market happened, before anyone else made a purchase, and then sold it for a different amount, which would change the stock price. You could stand to benefit from that, either by selling something slightly higher than what it was or buying it

when it was slightly lower. You know, there's a lot of different maneuvers you could do within the stock market, but you know, you're limited by how fast you can do them, which is where things like computers come in. Oh well, I mean because in in ye olden days, and by ye olden days, I mean for a great part of the twentieth century. In fact, this involved a lot of literal running around. I mean high high technology at the time was hiring the fastest and loudest kid

that you could to go work on the exchange floor. Um. And or sometimes pigeons. Legends say that pigeons were used to trade in real time on the outcome at Waterloo. The Rothchild family supposedly used pigeons to uh, to learn the outcome of the Battle of Waterloo before anyone else that they could then act upon the market accordingly. So I guess they had a strategy like, okay, if the French win, go with this. So it was kind of

an interesting Uh. That's that's an interesting legend. Whether or not it's true, you know, it could be apocryphal, we don't know sure, but um, but all of this did definitely change when computers started to become a thing. Um. In October, a computer system replaced that original trading floor of the of the New York Stock Exchange, and uh it tripled the number of exchanges that could be made

within months or that were literally made within months. Yeah, so you suddenly see a lot more volume as a result, you see uh more you know, a higher frequency but not high frequency trading. It that would come later, but it's certainly that that's set the stage and now we've gotten to a point where we really want to eliminate

latency as much as possible. Latency is that period between recognizing a potential trend and acting upon it, and latency depends on a couple of different things in this world. It depends on how good your computer algorithm is a recognizing trends and then and then executing a command. And it depends upon the actual communication connection, physical hardware that you're using, right, even the infrastructure, right the cables that

connect you to the New York Stock Exchange. In fact, at this point, you have companies that are willing to pay top dollar to have to have machines all stored in Essentially, they're all putting them in the same building that's near the New York Stock Exchange in order to have as as close to connect to the stock exchange is possible, and to keep things fair, the engineers in that building equalize the cable that runs to each of these servers, like like adding a few feet of cable

when necessary, so that these messages that are moving essentially y'all at the speed of light are going to equal out. Yeah, that's kind of incredible that a few feet can make a difference. So the actual physical speed of data transfer depends upon the medium you're using to transmit that information, right, because the speed of light is a constant, and it's a hypothetical ideal that we're never really going to reach. Right,

that's the speed of light and a vacuum. And so what we're dealing with is like various extremely fast transmissions, say like the stream of a microwave transmission through the air, a fiber optic cable across the street, you know. Right, As it turns out, light does travel at different speeds through different media, right, I mean, it all depends upon the medium or electromagnetic waves, because we're gonna be talking

not into really about light. That's true. That's true. So if you are talking about optic cable, so fiber optic cables you can transmit lots of information through fiber optic cables by doing lots of tricky things like having multiple cores and even a different geometric layout of how the light itself travels through that cable. So you may have heard about the fastest supercomputers in the world using fiber optic to move massive amounts of data in very little time.

But this is this is that difference between speed and throughput that we've talked about before in this podcast. It's not that the data is moving faster, it's it's able to move more data at the same time, so it's it's not necessarily traveling at the speed of light. So speed of light through optical cable is around two hundred thousand kilometers per second. But you know if that sounds pretty fast two kilometers per second. I mean, I'm not

saying that's slow. But if you were to say, set up a microwave tower and a receiver and send a message that way, then travels at around three hundred thousand kilometers per second, so markedly faster. Right, You're going much faster than you would be if you were going through

optic cable. So there are companies that are setting up these microwave relay towers to transmit information between say the New York Stock Exchange and a trading houses in Chicago that deal with futures, right, because the activity in one can affect the other. And if you know about it ahead of time, ahead of your competition, you stand to make a lot of money. Yeah, I mean even by

by fractions of a second faster than your competition. And there are problems with microwaves that you don't run into with optical cable, Like it's a direct line of sight kind of issue, right, and if there's a storm, it can interfere with transmission. Right. So I mean the same reasons that we don't use satellites for this kind of thing because latency. I mean, you can move a lot

more data. But yeah, but for that information to travel all the way up into space and then relay with the satellite and then be be back down to some receiving station on Earth, that that information has to travel a huge distance obviously all the way up into lower th orbit and then back down and then or not even lower thorbit, that could be much higher than lower orbit.

Uh so then you have a real latency problem. Sure, and people are working with microwaves on that entire line of sight issue, right, No, sure, A couple of different companies have been working on building microwave towers, like I said, between New York and Chicago. They're also looking at the possibility of using microwaves to transmit information across the Atlantic, which would be a lot trickier because you know, building

towers out in the middle of the ocean not easy. However, there's at least one guy who suggests we use drones. Microwave drones, Yeah, drones that can actually the stock so you think, you know, you'd have to have a drone that would be very energy efficient to be able to maintain flight over the Atlantic Ocean without needing to recharge every few minutes. It would have to have some kind of crazy power source, perhaps one of the ones we've

talked about in one of our previous episodes. But if you were able to do it, then in theory, you could have a drone microwave relay station to transmit information across the Atlantic Ocean, which would travel faster than it

would with a fiber optic cable. So, assuming that there weren't these atmospheric problems in the way obviously that we mentioned earlier, right, sure, And I mean, as as crazy as this all sounds, it might in fact be more reasonable than the three hundred million dollar cable laying project that's underway to shave uh sixty five milliseconds to just under sixty from London too. So here's the thing. If

there's money to be made, someone's doing it. That's really the message we're getting at here is that you know, you're you're hearing like these incredible like you're just taking maybe a little bit more than five milliseconds off of a transmission line that seems so tiny to us, but the fact is that can trans translate into me, got dollars? Yeah, yeah, for for these for these crazy high frequency trader people. Okay, So I have a question. So obviously lots of people

are going to be trying to get there. They're buying signal out there faster to get out ahead of the competition. But I can foresee that one way to get out ahead of the competition is not to get there faster, but to make smarter decisions, or to trick your opponent into making dumber decisions. Right, So I have to guess if people are spending millions of dollars shaving half of a second off of a transmission time. They're probably also spending money on software that's designed to do this, am

I right? Yeah, yeah, So you've got software, like we said, the algorithms that are designed to specifically look for trends in the market. That's one type, right, They're looking for any kind of indication that a certain type of market condition is about to happen and so therefore act in advance of that, right, which I mean you've got these

algorithms watching Twitter blogs and stuff like that. Right, then you've got those exactly the ones that are looking at the news itself and not just market conditions, but also

external conditions that could then affect the market. Remember that the market is going to lag behind any news, but it's not gonna lag a lot as news, especially if machines making these decisions, because as news propagation gets faster, and if they are if there are ways of getting that news into a machine readable format faster and faster, whether or not that means that you've translated the news into machine readable format or you've created a program that

has really good natural language recognition software, then you're going to have machines reacting to real world news essentially immediately after it happens, or at least immediately after it's been reported. So then you have this market being even more reactive to real world conditions, sometimes perhaps overreactive, like in the

example of the false story about the bomb um. Further marking the software waters here, we've got some some algorithms that are undertaking something like like denial of service attacks. They're basically clogged clogging exchanges with quotes to to test the market or confuse other software, or to just plane slow down the entire system. Yeah, this is kind of like when you hear it is very much like a denial of service. I'm glad you put it that way,

because it's like pinging a server for a response. That's you know, a basic command, right you paying a server to see how long it takes for your message to go out and for a response to come back. And if you send lots of pings to a server from lots of different sources at once, you can overwhelm the server and that's a denial. That's a kind of denial

of service. Attach quotes are kind of similar. It's sort of like, uh, almost like a false bid for a stock and if you're able to send a lot of them out, it may end up clogging up the whole system or just fooling other algorithms into thinking that a particular stock is something it should move on. And then as those algorithms move on your decoy, you can then you know, twirl your mustache and then go after the

actual target you were interested in all along. Right, I should mention that some of these, some of these pings could be completely legit trying trying to test and see what price of stock is at it an given point, and you do have to do that pretty frequently and pretty quickly in order to engage in this high frequency trading stuff. And furthermore, unlike a denial of service attack, this is in no way illegal. It's just considered a little bit rude. Yeah, maybe some people don't even look

at it as rude. Now, now, granted, this is one of those conditions that could change over time. Right if if enough, if enough people or let's just say computers abuse this system and it becomes an actual problem as opposed to something that is, you know, generally looked upon as dirty pool, but it's all allowed, then you might you might actually see companies get a little bit of

a wrist slap for this kind of behavior. Um, so it may turn out that perhaps five years down the road, algorithms cannot have this sort of behavior built into them. You know. I I don't want to be unfair to the stock trading community, but I kind of have to see that within this world, there's not a lot of like middle ground between not allowed and we're going to do it. Yeah, there's not like this huge gray area where it's like, well, we could do that, but that

wouldn't be very nice. I mean, it seems kind of like, if it's not illegal, you should do it till it's illegal. It's essentially the message. Well again, it's all about if it's if it's profitable, then it makes sense, right, I mean literally makes sense in the case of fractions of the center. Okay, so now let's talk about how thing this is actually affecting the world today. We've already talked about the flash crash and the other the the blip,

the hundred thirty six billion dollar blip. But what is the future going to be like? In this world of high frequency trading, computer algorithms are reacting to news and other changing conditions. What's that future, Like some industry insiders are really gloomy about it. Actually they're they're worried about about jobs. And I mean there's definitely been a trend towards these algorithms replacing some of the people who originally worked in in fields trying to figure out what stocks

in particular markets were doing. Um. So, so they're worried about that. They're worrying, worrying about the dependence on technology changing the market into something unidentifiable and coldly automated, kind of the sort of buzz that you hear about any segment of the tech industry. Um. Not that I'm saying that they're you know, wrong to to have this concern. And then they're also concerned about stuff like glitchens and

hacks undermining market stability and are already really delicate economy. Yeah, that that delicate economy that was brought about by a lot of people just saying maybe maybe their messages, Look, we don't need help messing up our economy. We've been doing it for years done and so yeah, and like we said, it's it's not even a fair playing ground now if you are a person trying to play in this world. I mean you. You just can't move at

the speed of these algorithms. I mean, you literally cannot read something in the space of time that it would take a machine to read it and make a decision and enact that decision. Right. So, you know, it's one of those things where if you look at a stock price, you can pretty much guarantee that whatever that stock is trading at is no longer the price that you're looking at,

at least on Uh. You know, it might be a fraction of ascent different, but it will be different, and depending upon the type of trading you're engaged in, that could be a big deal. Uh. Now, we might see

if these systems do cause big issues, some regulations step in. Uh, it will probably be I mean, it's going to be a reactive kind of thing, I think, and I doubt we're going to see any proactive regulation, but we'll probably In fact, so far, the regulations seem to be opening up the market to more of this kind of trading, right, So it'll only be I think in the wake of something uh spectacular or catastrophic, however you wanna describe it. I think that's going to be the required event before

we see any regulations. Step in. But I imagine that that would happen eventually. Should this prove to be more harm than good, I mean, because of ultimately this is about making money, and if it means that people are not able to make money because they're all these bullips that create a false market where nothing the value of everything is not accurately reflected in the prices, which, by the way, you could already argue for a lot of companies.

But if if this made that problem bad enough, I would imagine that would cause some regulations to arise as a result. Also, you know, there is a difference between high frequency trading and the kind of investment that your average investor is doing. Right, It's not right. This isn't care right. Well, you know, this is like we said, you know, like you were saying, Lauren, It's all about getting in and getting out in the same day and making lots of small trades throughout the day, you know,

thousands of times a set. And sometimes there are hobbyists are professionals who are doing this on on purpose a lot really hard. This isn't This isn't about investing in a company or even people who want to invest in a company they believe in as a way of supporting that company and then earning rewards off of that support in the long run, assuming that company is actually successful. You know, that's a very different approach. That's a very

different strategy. Although I did want to put in that if all this high speed stuff sounds weird and like counterproductive to that ostensible goal of of an investor and a company, you know, mutually benefiting each other with that money exchange. Traditional stock theory does state that all trades are of benefit to the overall economy because they keep the market flexible and more welcoming to investors. Yeah, it's all about liquidity, right exactly. So yeah, this is a

so it's you know, is it scary? Sure, because you're talking about a lot of different automated processes that are, you know, we hope working to their best advantage, but could conceivably if they all all react in a similar way, either drive the market up and and it would end up being a false market that way, or make it

crashed down. Hopefully, should that ever happen, the market would be able to correct itself relatively quickly the way it did back in two But you know, it's certainly one of those things where you know, you feel like the control is slipping further and further away from human hands. And that's always something that gives us a little bit

of a pause. Oh sure, and it's not even with high frequency trading, like like I was saying a second ago, Um, you know, algorithms are starting to replace lots of traditional stock analysts, so so changes are definitely happening across the board. Yeah, yeah, it's just that you might not for those other methods. You might not have something where it's trading as frequently.

It maybe all right, right, you're not going to cause these kind of data sensitive crashes, right, Yeah, you might just have an algorithm that's looking at the market in general and saying which of these prospects seemed the most promising for X payoff? Because you know, again, like if you've ever in uh been part of a four oh one K, you've you've probably seen multiple choices that you

can put your money into. Some of them are higher risk but higher payoff at the risk pays off, some of them are lower risk but lower payoff, same sort of thing. These algorithms are just trying to identify those

those trends. Uh, mathematically, you know, Uh. You know, I remember reading from one analyst who said that according to what they could tell that you could be just as effective in the stock market if you threw a dart at a list of stocks as you could if you were a strategist really studying the market, which is a little cynical. I'm sure all of us have felt that way, if we've looked at our four oh one case at some point, especially over the past three years. Yeah, let's

not talk about it. I've Yeah, I like to think I've funded several kickstarters magically through the money lost in my four oh one case. So that's that's the way I like to think of it. Y'all obviously have not been funding Alchemy like I have. Yeah, it's it's do I tells you. Returns on Alchemy chemical is just really but you know, I don't know. I think that overall, having more data and more software devoted to crunching that

data in any system is a good thing. I mean, I cannot I cannot look at that and say, no, knowing stuff is bad. Um And and you know the fact that we've got these um online investing tools is also a pretty fabulous thing. You know, It's it's letting people that might have never gotten into this kind of gig experiments. Yeah, yeah, it's definitely. It's definitely opened up

those the opportunities quite a bit. I mean, you know, it doesn't take you don't have to think back that far to imagine a world where the stock trading the only thing anyone knew about it was from films like Wall Street because it was something that only, yeah, a very elite few were engaged in, and most of them were big companies or very wealthy individuals who were engaged in, and the rest of us were, you know, feeling like it was a world that that was mythical, like the

land of dragons and unicorns. But these days we can all have access to it. Well, that kind of wraps up our discussion about the stock market, the computer algorithms, the robots that are going to own everything that we have built over the last a few centuries. Never hook sky net up to it. Okay, well I'll take that under consideration. I don't know Skynet's really convincing. I was talking to it the other day and it had some great ideas. So guys, here's another great idea. Go visit

f W thinking dot com. That's a website where we've got all of the blog post, podcast videos, other content that is really cool. It's all about the future and science and technology and you should definitely check that off you haven't been there, and also remember you can you can join in on the conversation. We are on various social media platforms including Facebook, Twitter, and Google Plus. You can find us with the handle fw thinking and we

will talk to you again really soon. For more on this topic and the future of technology, visit forward thinking dot com, brought to you by Toyota Let's Go Places