Case study Broker trade executions

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
lazyday
Posts: 3094
Joined: Wed Mar 14, 2007 10:27 pm

Re: Case study Broker trade executions

Post by lazyday » Sat Jan 20, 2018 10:03 am

Caduceus wrote:
Sat Jan 20, 2018 8:56 am
I place limit order of 10,000 shares for $30, with stock currently at $31. A pair of transactions takes place at $30 and I don't get filled (the order was placed before mine.)
My challenge trading stocks with large spreads (such as 500 basis points) is not that another bid executes at $30 before mine executes at $30. It's that other trades execute at $30.01 or so. And my bid of $30 would only get executed if the spread moves so that the ask is nearly $30.

For the last several years, I've found limit orders on most very thinly traded stocks (not ETFs) to be useless. Instead I usually wait for a narrow-ish spread and place small market orders. If there's a burst of volume after an earnings announcement, the spread might narrow during parts of the day.

Years ago, I don't recall having this problem, and my limit orders could execute at good prices without the spread moving to my order price. I think I had good luck with limit orders for years after stocks stopped trading in fractions.

I'm not a frequent trader, and don't understand the workings of the market, so please take the above for what little it's worth.

Luckily most of my illiquid shares are gone now.

User avatar
triceratop
Moderator
Posts: 4354
Joined: Tue Aug 04, 2015 8:20 pm
Location: la la land

Re: Case study Broker trade executions

Post by triceratop » Sat Jan 20, 2018 1:03 pm

livesoft wrote:
For the past 2 days I never got fills on some limit orders to buy set at the bid even though many orders were filled at that price. The volume was about a million shares at the end of the day, so not thinly traded. I am thinking collusion is why my orders did not execute.
Were you looking at the order book at the time? That happens to me too and my guess is always that the volume wasn't sufficient to hit me given how low in the book I am.
"To play the stock market is to play musical chairs under the chord progression of a bid-ask spread."

livesoft
Posts: 59191
Joined: Thu Mar 01, 2007 8:00 pm

Re: Case study Broker trade executions

Post by livesoft » Sat Jan 20, 2018 1:45 pm

triceratop wrote:
Sat Jan 20, 2018 1:03 pm
Were you looking at the order book at the time? That happens to me too and my guess is always that the volume wasn't sufficient to hit me given how low in the book I am.
Yes. Here is a screen capture from about 20 to 25 minutes after entering orders:



Image



The thing to note in the above is that trades happen every 30 seconds or so and 5 trades happened at the bid price of 30.36 or about one-third of the trades shown.



Below I show 4 trades I entered. My earliest order (at the bottom) was a market order and filled right away at a price about halfway between the bid/ask. My other three orders were limit to buy at 30.36. The first of those limit orders was executed at 12:58:16 about 4 minutes after it was entered. The other two expired at 4 pm and were never executed. Clearly, the 5 trades that executed at 30.36 shown in the screen capture above were not my orders. Not shown are the trades between when I entered my orders and 13:13:58 first shown trade in the screen capture above. There were many trades at 30.36 which were not my trades. So whose trades were they? Yes, you can say I was low in the book since thousands of shares are offered. I believe all those shares are just computers trading with retail investors who place marketable limit orders. If the retail investor had just placed a market order instead, I would guess they would get a price about halfway between the bid and ask as seen for many of executions shown. I'm sure ftobin or someone will correct me since I am probably wrong. :)



Image
Last edited by livesoft on Sat Jan 20, 2018 1:57 pm, edited 2 times in total.
This signature message sponsored by sscritic: Learn to fish.

livesoft
Posts: 59191
Joined: Thu Mar 01, 2007 8:00 pm

Re: Case study Broker trade executions

Post by livesoft » Sat Jan 20, 2018 1:55 pm

Here is another SPSB example from January 11:



Image



There is one trade that really stands out. That is one reason for the screen capture. I had several orders to buy at the 30.42 that were getting no love. Then all of a sudden, they all executed. I surmise that somebody got their broker to sell with a block trade over 100,000 shares for a penny less than the prevailing bid/ask spread. Their broker knew they had willing buyers at the current bid price and just made a quick thousand dollars for their trouble. OTOH, if the entity that sold the 100K+ shares went to cash, then over the next week, they saved themselves $5,000.
This signature message sponsored by sscritic: Learn to fish.

saver007
Posts: 72
Joined: Fri Nov 07, 2014 9:18 pm

Re: Case study Broker trade executions

Post by saver007 » Sat Jan 20, 2018 2:58 pm

Well, one big problem is if your broker sell your order to a internalizer rather than to route it to a public exchange, your order won't be part of the exchange order book shown in that level 2 quote. The internalizer/market maker who bought your order from the broker have the legal obligation to execute at no worse than nbbo but they are not going to go out of their way to make sure you get the best price possible. Their profit depends on giving you the worst execution legally possible! Since your order might not have went to exchange, consider it to be the last in the orderbook perpetually.

I recommend using a broker who don't sell your order flow and have a good order routing algorithm
.

ftobin
Posts: 1011
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin » Sat Jan 20, 2018 3:50 pm

Tanelorn wrote:
Sat Jan 20, 2018 8:27 am
ftobin wrote:
Sat Jan 20, 2018 6:29 am
If you have a buy order with Fidelity, they cannot buy at a price equal or lower than your price without giving you the shares.

But they can certainly sell to you, which is generally to your benefit.

In your story, your limit orders at $30 would executed anyways if the stock was at $29.
Could this be an example of a situation where an AON order might get filled at a worse price than $30 when the broker or market maker bought the shares for $29?

In non-NMS markets like bonds, I think it's also possible that the broker may keep inventory or buy up orders at $29 only to flip them to customers at $30.
Yeah, I was not including AON orders with my statement because you shouldn't be sending AON orders :-). However, your scenario is, I'd say, 50% true for AON orders (different market makers' compliance departments rules can differ here). AON orders are "not held" and subject to laxer rules and different interpretations/implementations.
In non-NMS markets like bonds, I think it's also possible that the broker may keep inventory or buy up orders at $29 only to flip them to customers at $30.
I'm not familiar with the bond market, but it's still true for non-NMS equities markets like OTC Markets.

ftobin
Posts: 1011
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin » Sat Jan 20, 2018 3:57 pm

Caduceus wrote:
Sat Jan 20, 2018 8:56 am
Thanks Ftobin. Yes, it makes sense that brokers with their own inventory can always sell to clients. But I am thinking specifically of thinly traded securities. For example, I place limit order of 10,000 shares for $30, with stock currently at $31. A pair of transactions takes place at $30 and I don't get filled (the order was placed before mine.) Next spread is $29-$30, and with 4,000 ask paired with my 10,000 bid. If I'm reading you right, Fidelity is not allowed to buy the next block of 4,000 shares at $29, and then turn around the next second and sell the same block of 4,000 shares to me at $30.

If that's the case, then how do brokers "front-run" their clients? I've read various finance articles which say that placing large limit orders (esp. for institutions) runs the risk of front-running by brokers. How would that happen then?
A couple of possibilities:
  • The order was not subject trade ahead protections (which is why you shouldn't trade AON orders). E.g., institutional orders have less protections.
  • If the order was large enough to create a support level, the market maker, if they had enough alpha, could buy at 29.01 and encourage the stock upwards, knowing they flip to you at 29.00 for, at most, a 1-penny loss.

ftobin
Posts: 1011
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin » Sat Jan 20, 2018 4:00 pm

saver007 wrote:
Sat Jan 20, 2018 2:58 pm
Well, one big problem is if your broker sell your order to a internalizer rather than to route it to a public exchange, your order won't be part of the exchange order book shown in that level 2 quote. The internalizer/market maker who bought your order from the broker have the legal obligation to execute at no worse than nbbo but they are not going to go out of their way to make sure you get the best price possible. Their profit depends on giving you the worst execution legally possible! Since your order might not have went to exchange, consider it to be the last in the orderbook perpetually.

I recommend using a broker who don't sell your order flow and have a good order routing algorithm.
You generally do *not* want your order routed to the open market. An internalizer is usually going to generally provide 40% of the spread as price improvement because of best-ex rules and price-improvement competition among market makers. You aren't getting that in the open market.

saver007
Posts: 72
Joined: Fri Nov 07, 2014 9:18 pm

Re: Case study Broker trade executions

Post by saver007 » Sat Jan 20, 2018 4:58 pm

Open market has many more participants than a handful of internalizers typical retail broker brokers exclusively route order to. When I last looked into it, I saw some brokers using just one internalizer or market maker. I don't know how much price competition happen in those scenarios.

If you want best price execution, you need a broker that can smartly route to as many market centers as possible both public and dark.

User avatar
alec
Posts: 2888
Joined: Fri Mar 02, 2007 2:15 pm

Re: Case study Broker trade executions

Post by alec » Sat Jan 20, 2018 4:59 pm

livesoft wrote:
Sat Jan 20, 2018 1:55 pm


There is one trade that really stands out. That is one reason for the screen capture. I had several orders to buy at the 30.42 that were getting no love. Then all of a sudden, they all executed. I surmise that somebody got their broker to sell with a block trade over 100,000 shares for a penny less than the prevailing bid/ask spread. Their broker knew they had willing buyers at the current bid price and just made a quick thousand dollars for their trouble. OTOH, if the entity that sold the 100K+ shares went to cash, then over the next week, they saved themselves $5,000.
Livesoft,

Chances are your orders to buy @ 30.42 got filled because (1) they were displayed on an exchange and (2) to be able to execute that big block (off exchange) 1 penny below the exchanges' bids of 30.42, by rule, the firm that executed at 30.41 was required to take out all the displayed liquidity on the exchanges, of which your orders were some.
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair

livesoft
Posts: 59191
Joined: Thu Mar 01, 2007 8:00 pm

Re: Case study Broker trade executions

Post by livesoft » Sat Jan 20, 2018 5:02 pm

@alec, I agree. I often see my orders displayed which can be bad because it means my limit order is probably mis-priced and everybody else knows it. :twisted:
This signature message sponsored by sscritic: Learn to fish.

ftobin
Posts: 1011
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin » Sat Jan 20, 2018 5:05 pm

saver007 wrote:
Sat Jan 20, 2018 4:58 pm
Open market has many more participants than a handful of internalizers typical retail broker brokers exclusively route order to. When I last looked into it, I saw some brokers using just one internalizer or market maker. I don't know how much price competition happen in those scenarios.

If you want best price execution, you need a broker that can smartly route to as many market centers as possible both public and dark.
For the most part, market centers don't price-improve. I assure you, orders that we, as a market maker, handled that got routed to the market had MUCH worse execution prices than those we internalized. There is *intense* competition between market makers on price improvement.

User avatar
alec
Posts: 2888
Joined: Fri Mar 02, 2007 2:15 pm

Re: Case study Broker trade executions

Post by alec » Sat Jan 20, 2018 5:38 pm

ftobin wrote:
Sat Jan 20, 2018 5:05 pm
saver007 wrote:
Sat Jan 20, 2018 4:58 pm
Open market has many more participants than a handful of internalizers typical retail broker brokers exclusively route order to. When I last looked into it, I saw some brokers using just one internalizer or market maker. I don't know how much price competition happen in those scenarios.

If you want best price execution, you need a broker that can smartly route to as many market centers as possible both public and dark.
For the most part, market centers don't price-improve. I assure you, orders that we, as a market maker, handled that got routed to the market had MUCH worse execution prices than those we internalized. There is *intense* competition between market makers on price improvement.
:thumbsup
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair

aaronl
Posts: 325
Joined: Thu Apr 03, 2008 8:54 pm

Re: Case study Broker trade executions

Post by aaronl » Mon Feb 05, 2018 3:04 am

I just read this whole thread and the information contained here is fascinating.

I had always assumed that if I'm not in a rush for an order to get executed (and why would I be?), it's best to use a non-marketable limit order. If it's priced inside the spread, someone might trade against it. And even if it's slightly outside the spread, the random jitters in prices give me a decent chance that it will be filled within a few minutes. If not, I can try again later at a different limit price, but statistically I won't be worse off. Intuitively this makes sense because a non-marketable order provides liquidity, and I would be compensated for that through a statistically slightly better execution price than a market order.

But this thread makes me doubt that reasoning. By using a non-marketable order, you give up the possibility of price improvement and effectively pay the full spread after the market has moved towards your order. It might be preferable to take advantage of price improvement at the current price than to put in a non-marketable order and gamble that the price moves in your favor. The notion of being compensated for liquidity is something that happens with exchange-level fees and rebates, and doesn't extend to retail clients.

Also, the caveats for odd lots are interesting. I don't pay any attention to creating round lots for trades. I'll generally have a certain dollar amount I want to buy or sell and the share count will follow from that, and it's very likely to either be an odd lot or mixed lot. Combining non-marketable limit orders with odd/mixed lots may explain some of the weird execution behavior I've seen in the past.

I don't think any of this matters much to a buy-and-hold investor. But it is definitely interesting to understand the inner workings.

livesoft
Posts: 59191
Joined: Thu Mar 01, 2007 8:00 pm

Re: Case study Broker trade executions

Post by livesoft » Mon Feb 05, 2018 6:30 am

aaronl wrote:
Mon Feb 05, 2018 3:04 am
[...]
Combining non-marketable limit orders with odd/mixed lots may explain some of the weird execution behavior I've seen in the past.
This is the thread to post the "weird execution behavior" when you see some. I always enjoy seeing such trades.
This signature message sponsored by sscritic: Learn to fish.

Post Reply