Forums  > Trading  > Execution on Thick-Book/Wide-Tick Equities/ETFs  
     
Page 1 of 1
Display using:  

EspressoLover


Total Posts: 334
Joined: Jan 2015
 
Posted: 2015-11-09 04:03
Anyone have insight here on what's the best approach for getting good execution on thick-book US equities? I.e. lower-priced, high-volume stocks that nearly always sit at one penny NBBO spreads with large size on the touch. BAC or VXX would be good examples. The baseline would obviously be hitting the ask, which eats $.005 in spread cost and another $.003 in exchange fees. But what kind of improvement is feasible from a relatively simple execution system or off-the-shelf algo?

The alphas I'm working with are relatively patient (time horizons of 5-30 minutes) and the execution isn't bound by any fixed schedule. I can join the bid or even 1 level away and probably still get filled with regularity. I vaguely remember from ITG's curves that 1/4-1/3 spread is a good rule of thumb for low-participation, high patience execution that tries to mostly execute passively. Don't know how accurate that is. Also I'm thinking that since the 1-tick spread is so wide there's a good amount of opportunity for dark price improvement at the mid or better.

I'd appreciate any insight phorumers might have around this issue. Thanks.

Good questions outrank easy answers. -Paul Samuelson

gaj


Total Posts: 25
Joined: Apr 2018
 
Posted: 2018-07-07 13:53
Reviving this as I've been thinking about the same issue. Have you found any good solution?

The first thing I'd try is make use of exchange order types. There's all kinds of tricks with ALO, ISO, etc to get priority on new price levels. You can also try to post midpoint. Depending on how often your trade opportunities occur, it may be worth it to keep orders at levels behind to gain some priority.

The tricky thing about passive fills is adverse selection. Although your time horizon is longer, I'm guessing your alpha is at least partially correlated with short-term alpha that HFTs use. That means you'll often get filled by those HFTs when your alpha has decayed.

As for liquidity taking, you probably want to have a short-term alpha model as well and try to hit the bid/ask just before it goes away. Since you have a long-term alpha, you can afford to be more aggressive than other HFTs. It also helps if you can find other "fungible" instruments, UVXY and VXX for example, so you have more alternatives to snipe.

EspressoLover


Total Posts: 334
Joined: Jan 2015
 
Posted: 2018-07-09 23:08
Thanks, @gaj. Those are all great suggestions. Agree with all the points.

Never really found a silver bullet. There seems to be a long-tail of optimizations you can make, where you keep doubling time and effort for increasingly marginal gains. But there's not really a clear point, where you can say "okay, looks like all the low-hanging fruit is exhausted". As you said, it's also hard to disentangle execution strategy from the underlying signal. So, the "right" behavior is pretty dependent on the specific strategy and resources available.

Beyond what you mention, I've also found it worthwhile to keep trying different execution algos from different brokers. There does seem to be a pretty-wide discrepancy in quality. Another thing that helped, especially if you're not large size, is utilizing taker-maker exchanges. Either to get the rebate if you're aggressing. Or to get fast fills if you're passive, avoiding the need to manage long-lived orders.

Finally, I've found that it helps if the parent strategy gives the execution system some flexibility. It's conceptually easier to keep execution modularized, and to mandate it to complete orders in X amount of time. However those times that it has to fill at bad prices to meet it last minute obligation can contribute a seriously disproportionate amount of costs.

Again, this doesn't seem to be an easy problem. So I'd love to hear anything else you find.

Good questions outrank easy answers. -Paul Samuelson

anonq


Total Posts: 12
Joined: Aug 2018
 
Posted: 2018-08-10 18:32
If you're ok with using broker algos and you haven't tried Goldman that might be worth a shot particularly cause they do a good job of internalizing between their algos with their principal liquidity book. If you're flow isn't toxic to them can also ping their principal liquidity book directly although they do require a min amount of flow to do that.

EspressoLover


Total Posts: 334
Joined: Jan 2015
 
Posted: 2018-08-14 21:55
@anonq

Thanks! Will definitely look into that.

Good questions outrank easy answers. -Paul Samuelson
Previous Thread :: Next Thread 
Page 1 of 1