# Knight Capital loses $440 million in 45 minutes due to computerized trading **URL:** \ **Category:** Great Debates\ **Created:** [August 2, 2012, 7:39pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213 "2012-08-02T19:39:10Z")\ **Posts on this page:** 20\ **Page:** 3
**Author:** ![ralph124c](https://avatars.discourse-cdn.com/v4/letter/r/8797f3/32.png) [@ralph124c](https://boards.straightdope.com/u/ralph124c)\ **Post date:** [August 8, 2012, 6:57pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/41 "2012-08-08T18:57:13Z")
I wonder if some of Knight’s execs secretly bet against their own firm? It has happened before. ---
**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\ **Post date:** [August 8, 2012, 7:22pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/42 "2012-08-08T19:22:57Z")
[Here](http://www.theregister.co.uk/2012/08/08/knight_capital_analysis/) is an update from the Reg, with a plausible explanation of what went wrong with some links. It seems that they had a new trading program, and had a testing program to exercise it. The testing program had accidentally included in a package that hooked up to the actual exchange, and then merrily started to make trades - trades it thought were fictitious, but which were real. The link points out that Knights is a market maker which does trades for others, and not a company which is doing risky investing. Just a cock-up, which got big because the trading is so fast. ---
**Author:** ![BigT](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/bigt/32/12044_2.png) [@BigT](https://boards.straightdope.com/u/BigT)\ **Post date:** [August 9, 2012, 1:51am UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/43 "2012-08-09T01:51:12Z")
It’s still a cockup that shouldn’t be able to happen. And I really, really don’t get Sam Stone’s attitude. Large falls in the stock market, because you don’t feel an immediate effect, doesn’t hurt anyone? You can’t see any reason why the thing that caused the Depression should have safeguards to prevent a dumb computer from destroying it all? You may choose to continue to take it for granted that these people know what they are doing, but they very clearly proved they don’t. This sort of thing shouldn’t have been possible–I’m barely a programmer and I could put in safeguards that would prevent this. ---
**Author:** ![Kobal2](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/kobal2/32/20_2.png) [@Kobal2](https://boards.straightdope.com/u/Kobal2)\ **Post date:** [August 9, 2012, 2:43am UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/44 "2012-08-09T02:43:45Z")
> [@BigT](#): > > This sort of thing shouldn’t have been possible–I’m barely a programmer and I could put in safeguards that would prevent this. No you couldn’t. You could design 5 successive “If you do this it will fuck things up something ridiculous. Do you REALLY want to do this ?” pop-up safeguards with flashing lights and gongs, order a mandatory, week-long “DON’T DO THIS. DOING THIS WILL CAUSE FUCK UPS AND SHIT.” training seminar for all end-users and require the process to be authorized by two separate people on opposite sides of the building ; and there’s still going to be one genius to do it anyway and be surprised by any resulting fuck-up. Probably more than one, akshully. Fool-proof is not realistic - they’re always releasing better fools, ya can’t keep up. ---
**Author:** ![septimus](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/septimus/32/410_2.png) [@septimus](https://boards.straightdope.com/u/septimus)\ **Post date:** [August 9, 2012, 3:00am UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/45 "2012-08-09T03:00:50Z")
The hyper-losses of Knight Capital are another example of the risks imposed by hyper-activity in the financial sector. I didn’t think these dots were so hard to connect. Should government institute laws to force people to fasten seat belts? Opinions differ; some might accept such laws because unbelted drivers increase hazard for others; some not. My approach is the _libertarian_ (small-l) approach where costs are adjusted through taxation. Taxes on cigarettes are an obvious example; they reduce unproductive behavior, raise revenue, and help compensate society financially for the costs imposed on it by smokers. Those who think the “hyper-efficient” high-volume financial markets we see today provide a strong benefit to society may leave the room now! My remarks are directed at rationalists who grasp that today’s hyper-markets serve the general public poorly, serve average investors poorly, and benefit no one but the big financial traders themselves. And please do note that the high-stakes game played by the too-big-to-fail industry are ultimately variations on “Heads we win; Tails the taxpayer and general public loses.” I do not propose that government outlaw bad software. I do not propose that government outlaw high-speed trading. _A simple small transaction tax would reduce hyper-activity_ and thus its associated risks, and would raise significant revenue. And **please** don’t trot out, for the 3rd or 4th time, a tired pdf showing poor results from a 0.5% tax when current proposals are for a 0.03% tax. ---
**Author:** ![Latro](https://avatars.discourse-cdn.com/v4/letter/l/a87d85/32.png) [@Latro](https://boards.straightdope.com/u/Latro)\ **Post date:** [August 9, 2012, 8:52am UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/46 "2012-08-09T08:52:17Z")
> [@Voyager](#): > > In the Times today I read that there was a new capability or type of trading launched, and Knights wanted to be first, ahead of more cautious competitors. If that is correct, and they were shooting for a deadline, then I wouldn’t be surprised if they skimped on internal testing to meet it. The problem described in the Register should have been easily caught, but testing takes time and a good test environment costs money. > CEOs get a fortune because they are supposed to be so much better than lackeys like us. But when they screw up they get another fortune. Ancient kings and emperors who took the throne got the palace and the gold and the serving wenches - but when they screwed up their heads went on a pike. When they lost the big battle they didn’t get to depart with half the gold. If we stuck the heads of failing CEOs on a financial pike maybe they’d try a little harder. > [@Enkel](#): > > The funniest part is that somewhere, there is a programmer going “Oh, crap!” Because it is the programmer who will get the blame for this. Not the CEO who went against the warnings of the programmer and wanted it live ASAP, NOW!!, YESTERDAY!! “Why isn’t this finished yet?!” “I’m starting to wonder about your programming skills and if you are the right man for this job!” ,“We have reached the deadline, just put it live already.” “No, don’t give me all that technological crap, I don’t have to know about that, that’s your job, not mine.” No, the programmer will be blamed for not doing his job properly. “What, you warned us?” “Can’t remember, you obviously weren’t clear enough.” “You should have been more vocal in your objection.” “I’m starting to doubt whether you are the right man for this position.” ---
**Author:** ![ralph124c](https://avatars.discourse-cdn.com/v4/letter/r/8797f3/32.png) [@ralph124c](https://boards.straightdope.com/u/ralph124c)\ **Post date:** [August 9, 2012, 11:44am UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/47 "2012-08-09T11:44:42Z")
Why are financial markets designed in this way (such that “program trading” can take advantage of miniscule price differences in a stock)? It seems to me that the stock market should be designed to promote liquidity, not encourage hyper trading schemes. In any event, the big funds that hold shares for long periods of time (like Buffet’s Berkshire Hathaway, Fidelity Magellan, etc.) are affected by the activity-why don’t they put pressure on the trarders to stop this nonsense? ---
**Author:** ![TriPolar](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/tripolar/32/3008_2.png) [@TriPolar](https://boards.straightdope.com/u/TriPolar)\ **Post date:** [August 9, 2012, 12:56pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/48 "2012-08-09T12:56:07Z")
> [@ralph124c](#): > > Why are financial markets designed in this way (such that “program trading” can take advantage of miniscule price differences in a stock)? It seems to me that the stock market should be designed to promote liquidity, not encourage hyper trading schemes. In any event, the big funds that hold shares for long periods of time (like Buffet’s Berkshire Hathaway, Fidelity Magellan, etc.) are affected by the activity-why don’t they put pressure on the trarders to stop this nonsense? Because those trarders own the country. ---
**Author:** ![Trom](https://avatars.discourse-cdn.com/v4/letter/t/958977/32.png) [@Trom](https://boards.straightdope.com/u/Trom)\ **Post date:** [August 9, 2012, 2:06pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/49 "2012-08-09T14:06:25Z")
> [@ralph124c](#): > > Why are financial markets designed in this way (such that “program trading” can take advantage of miniscule price differences in a stock)? It seems to me that the stock market should be designed to promote liquidity, not encourage hyper trading schemes. In any event, the big funds that hold shares for long periods of time (like Buffet’s Berkshire Hathaway, Fidelity Magellan, etc.) are affected by the activity-why don’t they put pressure on the trarders to stop this nonsense? Because on a whole, HFTs are beneficial. From a Q&A with gigantic, buy-and-hold Vanguard’s CIO: > [@](#): > > Q: What is your take on high-frequency trading [HFT]? Is it helpful or harmful to individual investors? > > Sauter: The term high-frequency trading is applied to many different types of participants and strategies in the market. Market makers are high-frequency traders. We need market makers to provide liquidity and for price discovery. High-frequency trading is also used for arbitrage opportunities. We want the derivative markets to be in line with the cash markets. We want futures contracts to be in line with the underlying security. It’s the same with ETFs. Arbitrageurs use high-frequency trading to keep markets in synch with each other. Maybe at the other end of the spectrum there are people abusing the system. The SEC is trying to identify people doing that. It’s a very small percentage of high-frequency trading. Investors have benefitted from changes in the market including HFT. Transaction costs have declined precipitously the past 10 to 15 years. **Net-net, we’re better off with HFT than without. Individual investors have benefitted more than institutional money managers.** Spreads have narrowed dramatically. The bid and offer are thinner than they used to be; it used to be deeper. But most traders are not trading in enough size to worry about market impact. We benefit from tighter spreads but we still have to be careful. my bolding ---
**Author:** ![JKilez](https://avatars.discourse-cdn.com/v4/letter/j/e5b9ba/32.png) [@JKilez](https://boards.straightdope.com/u/JKilez)\ **Post date:** [August 9, 2012, 2:38pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/50 "2012-08-09T14:38:06Z")
> [@septimus](#): > > Retail investors wouldn’t have overpaid by 29% if they used limit orders, but a previous thread covered reasons not to limit; and there are others. What’s wrong with limit orders? ---
**Author:** ![Really\_Not\_All\_That\_Bright](https://avatars.discourse-cdn.com/v4/letter/r/e8c25b/32.png) [@Really\_Not\_All\_That\_Bright](https://boards.straightdope.com/u/Really_Not_All_That_Bright)\ **Post date:** [August 9, 2012, 2:58pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/51 "2012-08-09T14:58:56Z")
How _could_ individual investors possibly have benefited more from high-frequency trading? Individual investors can’t do it. ---
**Author:** ![septimus](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/septimus/32/410_2.png) [@septimus](https://boards.straightdope.com/u/septimus)\ **Post date:** [August 9, 2012, 3:36pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/52 "2012-08-09T15:36:14Z")
> [@Trom](#): > > Because on a whole, HFTs are beneficial. > > From a Q&A with gigantic, buy-and-hold Vanguard’s CIO: The advantage of narrowed spreads is inversely related to one’s hold time. Very important for millisecond traders; good for day traders; almost irrelevant for buy-and-hold investors like Joe Average and Vanguard. If Vanguard’s CIO doesn’t grasp that, don’t blame me. > [@JKilez](#): > > What’s wrong with limit orders? I didn’t mean to issue a blanket condemnation; I just wanted to forestall any blame-the-victim comment that retail investors who ended up losing big along with Knight Capital “should” have avoided trouble by using limit orders. In a recent thread, a Doper proposed issuing buy orders for, say, 0.5% under the opening price. It was pointed out that he would win a few pennies most of the time, but lose big when the stock he wanted to buy took off and he never got in. Another potential disadvantage of limit orders is partially completed trades, or multiple commissions. (You can avoid this with an “All or None” restriction, but then run a risk that the trade will not complete at all.) I can attest from personal experience that even trades that seem quite small may complete only partially. ---
**Author:** ![Trom](https://avatars.discourse-cdn.com/v4/letter/t/958977/32.png) [@Trom](https://boards.straightdope.com/u/Trom)\ **Post date:** [August 9, 2012, 4:21pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/53 "2012-08-09T16:21:34Z")
> [@septimus](#): > > The advantage of narrowed spreads is inversely related to one’s hold time. Very important for millisecond traders; good for day traders; almost irrelevant for buy-and-hold investors like Joe Average and Vanguard. If Vanguard’s CIO doesn’t grasp that, don’t blame me. Huh? Cite? That’s a bizarre statment. Why would anyone ever want to get hosed by paying a large spread? How could a market that is $10.00 bid x $11.00 asked possibly be better for a retail investor than one that is $10.49 bid x $10.51 asked? ---
**Author:** ![septimus](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/septimus/32/410_2.png) [@septimus](https://boards.straightdope.com/u/septimus)\ **Post date:** [August 9, 2012, 4:51pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/54 "2012-08-09T16:51:04Z")
> [@Trom](#): > > Huh? Cite? That’s a bizarre statment. Why would anyone ever want to get hosed by paying a large spread? How could a market that is $10.00 bid x $11.00 asked possibly be better for a retail investor than one that is $10.49 bid x $10.51 asked? Saw a lot of those 10% spreads in typical retail issues before the HFT, did you? ETA. And BTW how did “almost irrelevant” become “better”? I doubt you want intelligent debate but if you do, calm down with the misstatements and hyperbole, please. ---
**Author:** ![Trom](https://avatars.discourse-cdn.com/v4/letter/t/958977/32.png) [@Trom](https://boards.straightdope.com/u/Trom)\ **Post date:** [August 9, 2012, 5:12pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/55 "2012-08-09T17:12:10Z")
> [@Really\_Not\_All\_That\_Bright](#): > > How _could_ individual investors possibly have benefited more from high-frequency trading? Individual investors can’t do it. Tighter spreads and liquidity. Buying and selling stock used to be a lot more expensive for individuals. The only counter parties to trade with were specialists and big banks. Now HFTs compete with them and provide better prices. “Wall Street” generally hates the HFTs. ---
**Author:** ![Trom](https://avatars.discourse-cdn.com/v4/letter/t/958977/32.png) [@Trom](https://boards.straightdope.com/u/Trom)\ **Post date:** [August 9, 2012, 5:20pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/56 "2012-08-09T17:20:33Z")
> [@septimus](#): > > Saw a lot of those 10% spreads in typical retail issues before the HFT, did you? > > ETA. And BTW how did “almost irrelevant” become “better”? I doubt you want intelligent debate but if you do, calm down with the misstatements and hyperbole, please. That was clearly a hypothetical example. But, yes, I was say there were more 10% spreads before HFTs. I contend that spreads were significantly wider before HFT market making existed. Do you disagree? Smaller spreads are better for retail investors than larger spreads. Do you disagree? You said: [QUOTE=septimus] The advantage of narrowed spreads is inversely related to one’s hold time. [/QUOTE] Back that up with a cite, please. You hand wave away a cite from an industry leading professional that directly contradicts you and _I’m_ the one that’s not interested in honest debate? :dubious: ---
**Author:** ![justrob](https://avatars.discourse-cdn.com/v4/letter/j/a9adbd/32.png) [@justrob](https://boards.straightdope.com/u/justrob)\ **Post date:** [August 9, 2012, 5:36pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/57 "2012-08-09T17:36:47Z")
> [@septimus](#): > > It would be fun to construct such a hypothetical and expose a “double standard” but I don’t know how you can. Accidents are bad – it’s almost like a law of thermodynamics. In the event, normal traders were happy to take easy profits. In your hypothetical, they’d have to be happy to take easy losses. > > There is one way that might happen: If the other computer algorithms had an hitherto-undiscovered flaw that caused them to happily lose in some scenario and Knight’s new algorithm, accidentally or not, exploited that. Well obviously an investment company wouldn’t by a stock for more than it is trading for then turn around and immediately sell it for less than they paid but that just happened to the tune of $440M. In the ridiculous hypothetical that Knight Capital had made a $440M profit ( **which could never happen in the real world so just go along with it for the hypothetical** ) would they be getting the same re-do? I am clearly trying to expose a “double standard” which you could never show in the real world. ---
**Author:** ![Simple\_Linctus](https://avatars.discourse-cdn.com/v4/letter/s/9f8e36/32.png) [@Simple\_Linctus](https://boards.straightdope.com/u/Simple_Linctus)\ **Post date:** [August 9, 2012, 5:37pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/58 "2012-08-09T17:37:34Z")
I understand why tighter spreads and added liquidity is hypothetically a good thing for other investors. \*\*However - \*\* the HFTs are making money. Where is this money coming from? It can only be other traders. Consequently I have to conclude there`s a hidden cost to investors somewhere. Anyone wanna explain it? ---
**Author:** ![Trom](https://avatars.discourse-cdn.com/v4/letter/t/958977/32.png) [@Trom](https://boards.straightdope.com/u/Trom)\ **Post date:** [August 9, 2012, 6:08pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/59 "2012-08-09T18:08:43Z")
> [@justrob](#): > > Well obviously an investment company wouldn’t by a stock for more than it is trading for then turn around and immediately sell it for less than they paid but that just happened to the tune of $440M. > > In the ridiculous hypothetical that Knight Capital had made a $440M profit ( **which could never happen in the real world so just go along with it for the hypothetical** ) would they be getting the same re-do? > > I am clearly trying to expose a “double standard” which you could never show in the real world. In what way is Knight getting a redo? They pretty much failed as a company and were taken over by competitors at fire sale prices. Exactly the way it should work. ---
**Author:** ![Trom](https://avatars.discourse-cdn.com/v4/letter/t/958977/32.png) [@Trom](https://boards.straightdope.com/u/Trom)\ **Post date:** [August 9, 2012, 6:10pm UTC](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213/60 "2012-08-09T18:10:39Z")
> [@Confused dart cum](#): > > I understand why tighter spreads and added liquidity is hypothetically a good thing for other investors. > > \*\*However - \*\* the HFTs are making money. Where is this money coming from? It can only be other traders. > > Consequently I have to conclude there`s a hidden cost to investors somewhere. Anyone wanna explain it? Here’s Getco’s, one of the largest electronic market makers, explanation. [http://www.getcollc.com/index.php/getco/tertiary/market\_making\_explained/](http://www.getcollc.com/index.php/getco/tertiary/market_making_explained/) [Previous page](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213.md?page=2) [Next page](https://boards.straightdope.com/t/knight-capital-loses-440-million-in-45-minutes-due-to-computerized-trading/630213.md?page=4)