HFT robots seeing orders before the exchange ?

This thread raises the interesting question...
Whether "HFT robots" have caused a drop in volume/liquidity...
In "illiquid" stocks which I will define as trading 50,000 shares/day.

My gut feeling was that liquidity is about the same...
But after looking at > 10 stocks trading in the 50,000/day range...
That have had consistent floats and no significant ETF action for last 10 years...
None of them have lower volume...
And several have doubled volume in last 2-3 years.

Conclusion

It's not even a close call...
"HFT robots" have significantly increased liquidity (often by 100%)...
In NYSE Bottom Third by Volume...
Assuming one can access liquidity using Basic Trading 101.
 
----That have had consistent floats and no significant ETF action for last 10 years...
None of them have lower volume...
And several have doubled volume in last 2-3 years.


It's not even a close call...
"HFT robots" have significantly increased liquidity (often by 100%)...--

amen to that..i guess we all dreaming..

http://finance.yahoo.com/q/hp?a=04&b=26&c=2000&d=11&e=17&f=2012&g=m&s=qqq&ql=1

note that current QQQ volume is a fraction of what we use to have in 2007 and even fraction of what we use to have back DECADE AGO, 2002.

let me ask you this-is it possible today to trade without paying commission? can big houses trade and not paying at all?

can some big houses just shovel the shares? i mean-buy and sell to themselves? can members of some dark pool trade between each other and make "arranged" trades?
 
---But I trade HUNDREDS of stocks of varying liquidity...
So it all adds up.

Again, the Dude just doesn't know how to trade illiquid stocks...
It has almost nothing to do with HFT...----


i would appreciate,if you share your wisdom with us and give us an example, along with logic behind your actions.just like OP did. show us how to trade those stocks. show us,how it's done on professional level

Thank you!
 
Quote from Bob111:

---But I trade HUNDREDS of stocks of varying liquidity...
So it all adds up.

Again, the Dude just doesn't know how to trade illiquid stocks...
It has almost nothing to do with HFT...----


i would appreciate,if you share your wisdom with us and give us an example, along with logic behind your actions.just like OP did. show us how to trade those stocks. show us,how it's done on professional level

Thank you!
More...

The problem is: I could write a 50 page manual...
Describing my trading operation in extreme detail...
And you are just back to the Zero Sum Game reality.

The 2% that can compete don't need my manual...
The 98% that cannot will be unable to implement it.
 
i'm not interested in 50 pages. my question was very straight( i can make it even more simple)-what would you do and how would you buy those shares in OP's example.
 
Quote from Bob111:

i'm not interested in 50 pages. my question was very straight( i can make it even more simple)-what would you do and how would you buy those shares in OP's example.
More...

OK, I only watch NBBO bid-ask.

This would only interest me after a $0.30 to $0.50 intraday decline...
Depending on other factors as well.

So bid-ask is 26.21-26.29 after decline...
I would send 2 orders via IB SMART...
Buy 300 at 26.34 and buy 300 at 26.28.

I would almost certainly fill 200-300 shares...
And be willing to average down somewhat to about 1500 shares max...
That would be my limit for a stock this illiquid.

Then I would immediately put in sell order(s) at $0.08 to $0.10 higher...
And be very happy to scalp the stock here for a few days.

Every fill on either side sends an Automated scalping order...
It's just a tiny position out of maybe 60 long positions vs 60 short positions...
I have a 100,000 line code base on 5 PCs that analyze market versus my Portfolio...
It's not rocket science... but extremely well organized.

I could just as easily be short this in the reverse scenario.

The common is rock solid...
So there can't be any fundamental reason for a decline out of the blue.

Two weeks ago I could not even have told you if I'm up on this one...
But it turns out I've netted $1,400 on 125,000 shares after all costs...
So that's a $0.011 profit margin.

In a low vol year I might trade 40,000,000 shares to net net $250,000 = $0.006/share...
In a high vol year like 2008 I've netted > $1,000,000...
All Market Makers cleaned up in 2008 because of the wide spreads...
And easy scalping in panicky markets.

I don't feel "HFT robots" have affected me all that much...
My view is that HFT is a SEPARATE Zero Sum Game...
That doesn't affect Limit Orders all that much.

But I have to constantly upgrade my Systems...
I've gone from 1,000 orders/day to 2,000 orders/day...
And this winter I'm doubling again to 4,000 orders/day.

In the mid-90s I used to phone in about 50 orders/day.
 
Quote from DeeDeeTwo:

OK, I only...
More...

Interesting post. How are you staying market neutral if you only put bids in after an intraday sell-off and vice versa? In other words, don't you accumulate a large net long position in a broad market sell off and the opposite in a rally?
 
Quote from DeeDeeTwo:

"It's not rocket science... but extremely well organized."
More...
Nominated for quote of the year. :D :D :D

In so many ways, this summarizes the vast majority of algo trading, at all frequencies, imho.
 
I don't think the guy on zerohedge is all that crazy. I have had similar problems taking/hitting especially when the offer/bid is on multiple exchanges. Kind of surprised he gets 0 fills though. For instance, I'll see 3000 shares of XYZ on the offer split up between NYSE, ACRA and BATS. I'll go to take the 3k using a smart route and get filled on 1-2k meanwhile all 3k prints. Somebody else isn't getting the first fill, they are getting the fills after the first exchange prints the first trade. I think the HFT's have gotten so fast that somehow they can get in between peoples orders that are broken up on different exchanges. Seems impossible if the orders are sent in parallel, but I find it happening all the time.

I trade through Lightspeed and have not found any smartroute/ecn that can get all the shares on the bid or offer consistently. I've tried NDAQ, INCA, EDGX, ARCA and Lightspeed's smart and supersmart routing and none have been totally satisfactory. I think HFT's have found ways to game people's orders in situations where they should not be able to. If you see 3k on the offer and you're the first guy to hit the buy button, you deserve 3k shares. Anything less is a quasi disfunctional market, and I think thats where we are. Not a rookie trader, 13 year professional.
 
Quote from DeeDeeTwo:I would send 2 orders via IB SMART...
More...
I stopped reading here - because sending "IB SMART" orders allows your broker to sell/flash your orders before they see an exchange.

If you really wanted to buy you would send in a direct marketable limit order and it would fill at the NBBO.

Forget about commissions; routing "smart" gives permission for your order flow to be sold. A few years ago I used to rant at the brokers for selling order flow. These days it's all common knowledge what's going on so it's all on your shoulders.

If you want your order flow to be sold, sell it. If you don't, don't.

Whatever you do, don't b*tch that some HFT bot is screwing you when you are showing them the order flow to begin with.
 
You must log in or register to reply here.
Back
Top