Themis Trading: The HFT Whistleblower
- Thread starter NY0BScalper
- Start date
I'm not against HFT as a general rule, but exchanges need to charge for cancelled orders above a reasonable level. You shouldn't be able to send 100,000 orders then cancel them with no intent to actually buy or sell.
It puts a strain on exchange infrastructure which everyone else has to pay for through brokerage fees.
It puts a strain on exchange infrastructure which everyone else has to pay for through brokerage fees.
Brokerage fees don't pay for exchange infrastructure. Only the spread between the rebate and the charge on ECN fees, which are extremely low and competitive - in some cases, like with BYX, the exchange actually loses money on every trade. (To take liquidty, you get a rebate of 30 cents per thousand; to add liquidity you're charged 20 cents, so the exchange eats 10 cents per thousand on every transaction).
Cancellation fees are OK but they can't be levied on every trade, that would suck.
I believe right now NASDAQ charges for orders above 300:1. I.E. 1 out of every 300 orders needs to get a fill.
If they changed that to, say, 50:1, that would be okay. The problem with doing that is in some instruments, like SPY, and even more so the ADRs where the stock is dually listed, ALL the liquidity is arbing robots. This is also HFT, but this is good liquidity; you'd have a hard time executing 5000 shares, least of all 50,000, in SPY if the algos couldn't cancel every time ES ticked.
If cancellation ratios are changed, it's imperative that certain products - ETFs and ADRs, at least, be exempt.
My contention with the current cancellation rules is that a firm can put in 30,000 BS manipulative orders for tremendous size, say, 50,000 shares and all they have to do to not get billed by NASDAQ for excessive messaging is execute 10,000 shares 100 shares at a time. So a firm could have posted and then canceled 30,000 orders for 50,000 shares - a total of 1,500,000,000 (1.5 BILLION shares), and as long as during that month they execute 1/300th of 30,000 orders, I.E. 100 orders, even if those orders are for only 100 shares each, they're fine. (Technically in my example they'd be slightly under, because now they've posted 30,100 orders and only 100 have executed, so they'd need to do 1 more order for 100 shares).
As absurd as it is -posting 1.5 BILLION fake shares and only having to trade 10,100 REAL shares, this makes sense for message traffic; sending an order for 50,000 shares takes up just as much messaging traffic as an order for 100 shares. The consequence of this, however, is that all HFT machines send their orders 100 shares at a time.
I support a rule that makes the messaging ratio requirements be based on size. Even if it kept at 1/300, that would be OKAY, if it didn't mean you can flash 50,000 shares 299 times and then print 100 shares ONCE and now you're okay (the way it is now). I propose that the total volume of your orders is counted; so if someone flashes 50,000 shares 299 times, they have to PRINT, aka EXECUTE, 50,000 shares. This would have the effect of decreasing spoofing, generating the exchanges revenue that wouldn't effect most of us traders, increasing the size of transactions on the tape, and disabling the most manipulative of strategies.
---------
As far as my original question. Would anyone here actually do business with Themis and trust their order execution? They're a broker who offers execute services, yet 99% of their site is talking about how unfair the market is instead of how great they are at working within the system we have to achieve results. Does anyone else get the same "these guys are too whiny to convince me they're good at exceuting" feel I do?
I was more put off by the 3 page "white paper."
This is a non-factor....HFTs knows your stops are in-client and will trigger it anyway....
Good question, what's their angle. They get a lot of attention from these "papers" and judging from the comments on this site and others like it whenever one of Themis's rants is posted, many people seem to think Themis is actually discovering, proving and expert on something. Maybe that helps their business.
I think most people who have knowledge on Themis's subjects recognize that they are sprinkling a tiny bit of true information or data and then leaping directly to a conclusion that many people want to hear. I see little connection in their papers that I've read between their evidence and their conclusion. But it doesn't matter because the conclusion is so comforting to the audience. This tactic works against all the bogeymen in our world (or boogedymen to Elaine): evil HFT, evil Goldman, evil teacher's unions, evil democrats, evil republicans, etc.... As long as the last sentence is agreeable, people won't question how you got there.
The question is, are they aware these papers prove nothing and churning out these diatribes just to get attention, hoping that attention leads to business and prestige? Or are they confused about their own subject, railing against perceived enemies and the attention feeds only their egos, not their bottom line?
Judging from the content, I'm guessing the later.
...that and apparently all HFT transactions/participants are considered "frontrunning". Soon, we'll all be frontrunning ourselves!
A very articulate post.
Things get blown out of proportion when you get your ignorant 13y.o son to bash HFT:
http://www.youtube.com/watch?v=9ddoha-34gg&feature=player_embedded
^Son of a co-head of Themis. Awkward....
http://www.youtube.com/watch?v=9ddoha-34gg&feature=player_embedded
^Son of a co-head of Themis. Awkward....
Nice find. I got a good laugh out of that.
Just like his father, he anecdotally explains why he believes HFT is bad without citing any proof or explaining it with any depth. Then again, I probably shouldn't be counting on 13 year olds to understand market structure.