Hacker News Reader

Intentions are hard to prove and are very subjective. Big financial players who believe they are being exploited by smaller players who game their algos can lobby the regulatory bodies to punish those smaller players. This contributes to an ever increasing concentration of power in the financial sector to the large, too big to fail banks.

It's the reason that, even after tens of billions of dollars in fines, there have been no criminal cases brought against those in the big banks, while this guy faces a maximum sentence of 380 years if found guilty on all charges [0]

Spoofing is best handled by the particular exchanges but should be weighed against the harm it would do to honest market makers who have to place and cancel orders regularly

> In looking at Mr. Holder’s legacy in white-collar crime cases, the pursuit of corporations, especially banks, stands out. The highest settlements for violations have come in the last year. These include the $13 billion from JPMorgan Chase and the $16.65 billion from Bank of America for their roles in issuing mortgage-backed securities tied to faulty subprime loans. BNP Paribas also paid a $9 billion fine for violating economic sanctions laws. BNP was also required to enter a guilty plea to a criminal charge, a rarity in the world of corporate criminal prosecutions. But prosecutors did not file charges against any individuals for the wrongdoing that led to the large settlements. [1]

[0] http://www.telegraph.co.uk/finance/financial-crime/11554471/... [1] http://dealbook.nytimes.com/2014/09/29/eric-holders-mixed-le...

People use the term spoofing in an overloaded way:

1) As this article describes, spoofing is the use of orders to change the perception of the supply and demand of an instrument. It is often also called layering (a term that is also overloaded). It does not involve mass amounts of orders or cancels. In fact, good spoofers go to great lengths to make their order flow look identical to someone who has regular trading intentions.

2) Spamming lots of orders into an exchange for some nefarious purpose (I've heard this described as a DDoS on the exchange). Which is both widely over reported (to the point that I believe it is largely urban myth) and trivial to detect and prevent on the exchange side (the exchanges know exactly who is sending the orders, the specific order gateways have rate limits, etc).

Finally, exchanges do penalize market participants that cancel too many orders. This is typically done via a fill ratio (the % of orders to those that trade) and enforcement can be either in the form of price levels, fees or potentially banning from the exchange.

To make it more complicated, there are people who run bargain hunter strategies - people who deliberately place orders way outside the NBBO in the hope of getting fills at really good prices. E.g., I'll buy TWC at $140 but not it's current $148.

This is functionally identical to spoofing except that these people are explicitly hoping to get a fill. (Hence it's legal.)

Question: Matt Levine keeps pointing out that this is some rando in a track suit working from his basement in London, and MF Global is providing him with execution services for ostensibly zillions of dollars worth of orders. There's no way he could have taken fills, right? The trading intent was fiction, but so was the trading capability, right?

This is where the "intent" of spoofing becomes very very murky. There is a trading strategy called laddering (or sometimes confusingly layering) where a market maker will put orders at many levels of the order book at once (and sometimes at extreme depths, like 100s of levels). If they were to get swept in one direction there is absolutely no way they could trade all of those levels (their risk controls would kick in long before that). But, if the fills come in the correct order (ie 1 buy/1 sell) then the market makers would happily trade all of them.

Its even worse in pro-rata matched markets because participants will often put orders in at much higher quantities than they actually want to trade, so that they can get any of the fills. That is, they'd love to get filled for 100 contracts, but are putting in orders for 10000 contracts in order to get any of the matching. Again, if that filled in the proper order they'd happily trade it, but if not they don't want to.

So in both cases, spoofing and not spoofing, the actual ability to honor the orders is in question. So that in and of itself is not evidence of spoofing.

The difference (I believe) is that the market makers in the laddering cases would LOVE to trade every single one of those orders if it came in the right order and are not intending to change the markets behavior with the placing of those orders.

The intention of a trader on these kinds of markets is to make money. End. Telling him that you'll arrest him, or that he should feel bad, if he doesn't have the intentions and morality you want him to have - does absolutely nothing to change his intentions.

Markets work, despite this incredibly misaligned incentive, because markets can eventually reward people who are eventually more correctly predicting the future value of the assets.

FYI - your usage of the phrase "true price" makes me strongly confident that you totally misunderstand markets.

There's an argument for spoofing from earlier this year [1] which I would tend to agree with if the proportion of spoofers and HFT front-runners actually balanced out. Not sure how it really is in reality though.

[1] http://www.bloombergview.com/articles/2015-01-23/high-freque...

From what I can tell, the argument for spoofing devolves pretty rapidly into an argument for private order books (that is prices are not published at all).

This is available currently in the form of dark pools. They don't tend to work out as liquidity is much lower on them and bid/ask spreads are wider.

That may not be true if ALL venues were required to be private, but that would be a pretty dramatic change (and exactly the opposite of what we have now, which is a requirement to publish and meet other published prices).

> They don't tend to work out as liquidity is much lower on them and bid/ask spreads are wider.

If you're a big institutional investor though, and you know that buying a bunch of shares will move the market for sure, well, it might be worth suffering a wider spread. As long as that spread looks small relative to how much you think you'll move the market, it's not a bad deal. And if you already have a subscription to the pool you might as well try it before you get the algorithms involved to buy on the open market.

Absolutely if you are a large block trader your dream scenario is a very liquid market with very inefficient price discovery.

Dark Pools were meant to provide that, but it turns out that in practice it didn't work (because no one wants to provide liquidity in that environment) and Dark Pools ended up resorting to either letting liquidity providers in, going out of business, or extremely scammy things to keep up the ruse.

So there are no institutional investors who ever need to sell? Or maybe it's that people aren't willing to wait at all when they want to sell, so putting it on the dark pool for a few hours or days isn't really an option. Where as if you want to buy you might be willing to wait a day or two.

You are under the incorrect impression that providing liquidity means selling shares. Providing liquidity happens on either side of a trade. Market makers are full time liquidity providers offering to both buy and sell (usually at the same time). They make their money by bridging the temporal gaps in the supply/demand of an instrument.

So if you are using a liquidity providers it is evidence that you do not want to wait. If you did want to wait, you yourself can just put the order out and provide the liquidity to others.

Nasdaq has an excessive messaging policy and charges the exact fees you suggest.

http://www.nasdaqtrader.com/Trader.aspx?id=bx_pricing

Fees only kick in if you cancel 100x the orders than you actually execute. And then you only pay half a cent per order over that limit.

I was suggesting something much more like a txn fee per order closed above a much tighter threshold (like 3x)

And none of those limits currently apply to registered market makers... And I agree with their exemption from the harsher rule too, but they should bear increased scrutiny and guaranteed liquidity rules for the privilege.

Market makers are obligated to keep lots of orders way outside the NBBO in order to maintain an "orderly market".

If a trade occurs at $10.00 and then the next trade is $12.00, that's considered "not orderly". So a market maker must keep orders at $10.00, $10.20, $10.40, ..., $11.80, $12.00. This way if someone comes in and buys everything then the sequence of trades will be $10.00, $10.20, $10.40, etc.

HN should require a signed comment when you downvote.

My comment is getting hammered with downvotes now.... And I don't know why... I'd hate to assume that people just want to bury a simple, fair solution to the problem without any justification.

Because it is not a simple or fair solution to the problem; it is a sophomoric attempt at solving a human problem that simultaneously guts the actual mechanics of exchanges.

You want to make market-making unprofitable and to turn the stock market into your town's real estate market. How easy is it to sell a house? What is an accurate price for your house? At this moment? To the dollar?

Protecting market making is excuse-making since registered market makers are already exempt from "over-messaging" rules like this. Plus It would be possible to enter time-expiring orders to accommodate market making without allowing unbounded spoofing. The expiries are extra info you could use to disregard price flashers who have no intention of actually transacting. (filtering out ms expiries for example). Time-unbounded orders should cost something to cancel. (heck... Give institutions 3x the cancels as actual orders for free... It would still help).

Real Estate isnt comparable. There are no market makers in real estate because the product isnt standardized and you can't make an equivalence market in one-off products... Houses aren't securities and the illusions propagated by securitizing the loans around them clearly has wild historical market risk attached.

"it is a sophomoric attempt at solving a human problem that simultaneously guts the actual mechanics of exchanges."

Markets are human constructs...the pure mechanics of exhanges aren't more important than the humans they serve. Fixing the human problems should be the priority, right?

Again, you are confusing the different uses of spoofing. People that are spoofing put orders in and then leave them for a long time. The way they use those orders is that they pull them all at once.

There are lots of legitimate reasons for behavior that looks just like this. The problem is not one of technology, it is purely the intention that causes the problem.

No confusion at all kasey_junk, I think those cancels should cost something in all those spoofing cases you mentioned (except for registered market makers who are usually exempted from messaging rules like this in exchange for real rules).

Those canceled orders had to be listed, the bid-ask system had to transmit them, matching engines had to consider them, cancel machinery had to back them out, and price discovery is affected. Why shouldn't one pay for the costs incurred?

Cancels either are explicitly built into the fees or are captured currently as fill ratios. Spoofers (and everyone else) are happy to pay this. It won't prevent spoofing which was your original position.

I've had to do the math to calc this stuff... what you are saying is simply false. If I execute a single order with a full fill, and you execute the same fully filled order with 99 canceled order behind behind it, show me how the execution costs are "built in" to reflect the difference.

They are built into the same way extra buttons you get with dress shirts are built in. Not everybody needs them, theoretically the people who don't use them are subsidizing the people that do, but no one cares. It is cheaper to deal with missing buttons as the norm than to build custom supply chains to optimize for correct distribution. If I called up the shirt manufacture and demanded 10K buttons on the other hand, they'd probably make me pay.

Finally, none of this has anything to do with spoofing, because spoofing doesn't need high cancel rates. If anything, making cancels more expensive will encourage the behavior, because it will make traditional market making more expensive (either explicitly or implicitly by requiring membership in a cartel to be a market maker). Meanwhile, the spoofers trade is much higher margin and can absorb the new extra cost more readily.

Please don't complain about downvotes...

Deleted comment

I was doing as you asked. See how fruitful it was?