Case study Broker trade executions

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Case study Broker trade executions

Post by livesoft »

This thread is just a way for me to document some testing of various brokers trade executions. Please feel free to add your own experiences here.

A moment ago, I logged in to two different brokerage accounts and set up a limit order to buy one share of VCSH in each account. I had 2 browser windows open simultaneously and clicked on "Confirm" as close in time as I could. The limit order was not a marketable limit order, so in neither case was it executed. I will wait and see if either one of the identical orders gets executed and at what price.
Last edited by livesoft on Thu Oct 01, 2015 11:56 pm, edited 1 time in total.
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

For reference, here is a screen capture of real-time Level 2 quotes, etc at the time of placing the orders:

Image

Update: Both trades were executed about 40 minutes later at essentially the same time around 1:17 pm.

Image
Wiki This signature message sponsored by sscritic: Learn to fish.
caklim00
Posts: 2420
Joined: Mon May 26, 2008 10:09 am

Re: Case study: Broker trade executions

Post by caklim00 »

Very odd thing happened today... Put in multiple orders at a couple of different brokerages and I noticed that execution at Wells Fargo was MUCH better than TDAM, I was very, very surprised.

As an example, put in Sell Order for the same fund at both WF and TDAM. About 5 minutes later my WF position sold, yet my TDAM position was just sitting there (and I even had the Limit 0.01 less at TDAM, so surely someone or some computer would want to pay 0.01 less right?). I made a few more transactions this morning and I noticed the same thing... WF was much better at execution on both buy and sell than TDAM.

Now, at WF I couldn't buy one of the funds as it was on their "restricted" list, but I could at TDAM (that's an entirely different issue).

If you are wondering why I have the multiple accounts: broker bonuses. I suppose the poor trade execution at TDAM is worth the bonuses.

Side note: I went back and took a look at TDAM and I see there is a routing option:
1) Smart (Default option)
2) NSDQ (INET)
3) ARCA

I just left it on Smart for my orders...
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

Can you please clarify the size of these orders? I wish to try to understand iff round lots (number of shares divisible by 100) get a better execution than small odd lots of under 100 shares or even over 100 hundred shares such as 109 shares.

If I wish to sell 204 shares, I tend to submit 3 orders of 100, 100, and 4 shares since I do not pay commissions for trades. I have noticed that such a split generally seems to get a better execution than a single order for 204 shares.
Wiki This signature message sponsored by sscritic: Learn to fish.
caklim00
Posts: 2420
Joined: Mon May 26, 2008 10:09 am

Re: Case study: Broker trade executions

Post by caklim00 »

livesoft wrote:Can you please clarify the size of these orders? I wish to try to understand iff round lots (number of shares divisible by 100) get a better execution than small odd lots of under 100 shares or even over 100 hundred shares such as 109 shares.

If I wish to sell 204 shares, I tend to submit 3 orders of 100, 100, and 4 shares since I do not pay commissions for trades. I have noticed that such a split generally seems to get a better execution than a single order for 204 shares.
Multiple variations and all appeared to act similarly. One sell at TDAM for around 800 and at the same time a sell at WF for <100. But in another instance I had a buy at both TDAM and WF for <100 shares. WF executed first and it was almost 10 minutes later when the TDAM trade executed. Given it was a down-trending day this didn't turn out to be an issue, but it irks me that TDAM was so poor at execution. Every single trade that I made executed at the limit today. In the past I believe I've had trades at WF that executed better than the limit (even after having to wait on execution).

I think this is the first time I've ever made more than 10 trades at a broker where I moved for a broker bonus, but its incredible the difference. TDAM through in 300 free trades as well.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

Thanks for the feedback.
Wiki This signature message sponsored by sscritic: Learn to fish.
TradingPlaces
Posts: 1245
Joined: Sun Nov 09, 2014 12:19 pm
Location: 30.286029, -97.530011

Re: Case study: Broker trade executions

Post by TradingPlaces »

livesoft wrote:Can you please clarify the size of these orders? I wish to try to understand iff round lots (number of shares divisible by 100) get a better execution than small odd lots of under 100 shares or even over 100 hundred shares such as 109 shares.

If I wish to sell 204 shares, I tend to submit 3 orders of 100, 100, and 4 shares since I do not pay commissions for trades. I have noticed that such a split generally seems to get a better execution than a single order for 204 shares.
Questions / comments:

- what do you call a better execution?
- odd lots (smaller than 100) are exempt from Reg NMS rule 611 (order and quote protection),
- round lots (100, 200, 300) might send a different signal than mixed lots. Odd lots definitely send the signal that the order is uninformed, for the most part,
- someone mentioned routing to ARCA or INET. Those are bad places to get executed when you are sending a what would be a RESTING limit order. The issue here is effective price and rebates commissions.

When the displayed price is the same, e.g., a buy order for $100.00, the effective price is different at different exchanges.

In the order of lowest to highest, this is the list:

- DIRECT EDGE, NASDAQ (INET), ARCA, BATS-Z, NYSE - roughly 0.3 cents below $100.00: $99.997
- DIRECT EDGE-A $100.0001
- BAT-Y and NASDAQ-BX roughly 0.15 above $100.00: $100.0015
- some other unimportant exchanges like PSX, AMEX, Chicago, etc.

Thus, someone, e.g., an automated execution algorithm, looking to SELL, would route in the opposite order on that list:
- first, try BAT-Y and NASDAQ-BX,
- next, try DIRECT EDGE-A,
- next, try one of NASDAQ, ARCA, BATS-Z, NYSE.

Notice that from your point of view, you get $100.00. However, broker routing to an exchange in the first list vs the last list makes this much difference:

800 * ($0.003 + $0.0015) = $3.6

which, to the broker, might be a lot of money if they do it a lot and to a lot of customers.

On a $100 stock, that's just 0.45 basis points. However, on a $10 stock, it is huge: 4.5 basis points. And on a $3 stock, that is absolutely horrendous at 15 basis points.
I should clarify: it is not bad to send the order to NASDAQ (INET) or ARCA, provided you receive the EFFECTIVE price. However, if you are going to receive the displayed price, it is bad to send to those exchanges.
Last edited by TradingPlaces on Thu Aug 20, 2015 11:34 pm, edited 2 times in total.
TradingPlaces
Posts: 1245
Joined: Sun Nov 09, 2014 12:19 pm
Location: 30.286029, -97.530011

Re: Case study: Broker trade executions

Post by TradingPlaces »

FYI, these are all the US Equities exchanges:

https://www.batstrading.com/market_summary/

Change the date to any past date to see the volumes.

FINRA and TRF show volume from dark pools, internal matching programs, wholesaling, and trades done OTC, then reported to the tape.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

Thanks very much for the comments.
TradingPlaces wrote: - what do you call a better execution?
For me better execution involves things like:

1. Price improvement. A better price than my limit price.
2. A purchase price below the reported low price for the day.
3. A sale price above the reported high price for the day.
4. I am watching real-time level II quotes and I get an execution when there are several orders ahead of mine listed at the same price.
5. I get execution in one window before I can click in another window to see my order in a real-time level II display.
6. I submit same order at same time at two different brokerages. For me, the first order to execute had better execution.
7. On a market order to sell, I get a better price than the current bid. On a market order to buy, I get a better price than the ask. This might be called price improvement on a market order. In the figure posted previously in this thread, the lower left quadrant would show my trade between the bid and ask as seen by a few of the dots in that figure (or if I sell, I get the current ask (better than the current bid) and so on).
8. I get someone else to pay the spread. See: http://www.bogleheads.org/forum/viewtop ... 9#p1488409

What do you call better execution?
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
nedsaid
Posts: 19249
Joined: Fri Nov 23, 2012 11:33 am

Re: Case study: Broker trade executions

Post by nedsaid »

I remember Johnny McKay when he was coach of the hapless Tampa Bay Buccaneers during their hapless 0-14 season. Reporters asked him what he thought about his team's execution. "I'm in favor of it", he quipped.

I am sure some Bogleheads are in favor of broker execution.
A fool and his money are good for business.
Tanelorn
Posts: 2366
Joined: Thu May 01, 2014 9:35 pm

Re: Case study: Broker trade executions

Post by Tanelorn »

livesoft wrote:8. I get someone else to pay the spread. See: http://www.bogleheads.org/forum/viewtop ... 9#p1488409
How do you know the spread you earn isn't effectively negative, ie a loss? If the market falls fast, the algos will pick off your stale order and you'll wish you hadn't gotten that fill.

In some efficient market sense, a limit order represents an option sold ("I am willing to buy at this price regardless of market conditions") and half the spread is the premium you earn for it. In that world, it wouldn't matter if you placed market orders or limit orders, you get to the same place eventually up to noise/luck. But if you aren't as fast as the market makers, which you assuredly are not, you won't earn the full option premium of your limit order due to adverse selection and you may end up paying rather than getting paid.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

Tanelorn wrote:
livesoft wrote:8. I get someone else to pay the spread. See: http://www.bogleheads.org/forum/viewtop ... 9#p1488409
How do you know the spread you earn isn't effectively negative, ie a loss? If the market falls fast, the algos will pick off your stale order and you'll wish you hadn't gotten that fill.
I look at the order fills a few minutes before and after my execution. The lower-left quadrant of the figure I posted shows what I mean.
Wiki This signature message sponsored by sscritic: Learn to fish.
TradingPlaces
Posts: 1245
Joined: Sun Nov 09, 2014 12:19 pm
Location: 30.286029, -97.530011

Re: Case study: Broker trade executions

Post by TradingPlaces »

livesoft wrote:Thanks very much for the comments.
TradingPlaces wrote: - what do you call a better execution?
For me better execution involves things like:

1. Price improvement. A better price than my limit price.
2. A purchase price below the reported low price for the day.
3. A sale price above the reported high price for the day.
4. I am watching real-time level II quotes and I get an execution when there are several orders ahead of mine listed at the same price.
5. I get execution in one window before I can click in another window to see my order in a real-time level II display.
6. I submit same order at same time at two different brokerages. For me, the first order to execute had better execution.
7. On a market order to sell, I get a better price than the current bid. On a market order to buy, I get a better price than the ask. This might be called price improvement on a market order. In the figure posted previously in this thread, the lower left quadrant would show my trade between the bid and ask as seen by a few of the dots in that figure (or if I sell, I get the current ask (better than the current bid) and so on).
8. I get someone else to pay the spread. See: http://www.bogleheads.org/forum/viewtop ... 9#p1488409

What do you call better execution?
I generally understand your preferences, but I think there is a lot of preferences in different directions:

1. Price improvement: definitely agreed. This is an excellent criterion. Your best chance to get a price improvement is to have your broker to route to the best wholesaler. Essentially, a better wholesaler should give you a price inside the spread. E.g., price is $100.00 x $100.01, and maybe you get a midpoint execution: $100.005. Or if the spread is wider than a penny, you most likely get a price strictly inside the spread.

Do notice that price improvement makes a lot of sense for a market order or a marketable limit order. But for a resting limit order, it is hard to get a price improvement.

2/3. Getting a buy price below low price for the day, or getting a sale price above the high price for the day. Statistically very improbable. Your best chance is if recent price is very close to the day's max / min. Otherwise, it is basically a matter of random luck. In such timescales as minutes or hours, price is random. Say, current price is P_c, P_max is day's high so far; your chances of execution a sell order at P_max can be very precisely mathematically expressed in terms of P_max - P_c, and the recent realized volatility of the market.

You can not possibly measure broker quality because randomness is just too much.

4. I did not quite get this. But let me try to interpret it.

When you say ahead, do you literally mean ahead on the exchanges real time order book, but at the same level? If so, you have a zero chance to get executed, unless your order is held by the broker, and an event resulting in the Manning Rule occurs (https://en.wikipedia.org/wiki/Manning_rule and held is meant in a very strict way). In this case, I agree. A better broker will increase the probability of you getting Manning Rule advantage, but there are no guarantees.

5. This is an entirely subjective criterion. I fail to see how it translates to better execution.

6. Notice that you can define your objective to be either fast execution or better price. Immediacy has a cost.
But of course, it is possible to give a fast execution, and a really crappy one. E.g., market is $100.00 x $100.01. You send a buy market order. Broker fills you at $100.01 (half a spread cost), and the market immediately moves down one penny to $99.99 x $100.00.

Of course, this can happen. But a broker doing this consistently would get poor Execution Quality scores. I think SEC is getting better at articulating criteria for better execution.

7. Agreed. Even though Reg NMS mandates a penny spread, in a lot of stocks the efficient (some notion of a fair price that is closer to the "true" price) price is not necessarily at the midpoint. So you might get a price closer to that efficient price, strictly within the spread. Also, various exchanges have price improvement programs. I don't know the details, but you can read some: https://www.nyse.com/markets/liquidity-programs

8. You get someone else to pay the spread. There is no such thing. Even highly sophisticated algos can barely shave percentage off the spread, and some of the best algos might beat the spread just a tiny amount.

Say, market is $99.99 x $100.00. And you send a resting limit buy order at $99.99. Your order is not held, and you sat there until you got executed. Changes are almost 100% that you got executed because the bid level got cleaned, and after your execution, the price is $99.98 x $99.99.

You might still think you got a good deal, but in reality, you lost half a spread.

From my point of view, execution quality needs to differentiate one of two criteria: (a) either price, (b) immediacy, or (c) possibly both, but it is harder.

If price is $99.99 x $100.00, and I want to buy, I don't want the market to slip away. If I post a resting limit buy order at $99.99, there is a 50% chance that price goes up, and my order will not get executed (price slippage), and 50% chance that if my order got executed, I will show a loss of 0.5 cents in the immediate aftermath.

Do you see how in the first case, price slippage, immediacy is not satisfied, and in the second case, price is not satisfied.

In this order, I would prefer:

- to trade at mid-point immediately,
- to pay a small spread that is less than half-tick, e.g., quarter tick, or even 1/10th tick, but again get executed immedately.

And lastly, if I was trading in size, I would prefer to have the entire size executed.

Notice that if you are sending a 1 lot order to a thick book (e.g., 10,000 shares at bid and 10,000 at ask, 1 penny spread), it is very easy to get executed at mid-point, the full size.

But if order book is showing 300 at bid and 200 at ask, and you want to buy 2000 shares, then good luck. In a situation like that, I would appreciate paying higher than the midpoint (e.g., ask, the higher price), and getting the entire size executed at once.

There is a lot of literature in market impact, which describes what a fair price would be in such cases.

In my opinion, retail investors are getting extremely good prices these days, primarily because of technology and proliferation of electronic trading participants. Not all brokers are top notch, and some of them are stuck in crappy deals with certain wholesalers, and that's where improvement can happen.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

Thank you once again for the comments. I appreciate them.
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

The other side of better execution is poor execution. One can read books such as Scott Patterson's Dark Pools and Michael Lewis's Flash Boys for lots of info about poor execution.
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

I am doing another test this morning. I have submitted the same ETF market order BEFORE the open at two different brokers at the same time. I will report on the executed trades in a later post this morning. Normally, I would not submit an ETF market order when the market is not open, but my curiosity has gotten the better of me.

Update @ 9:38 am Eastern. Both orders executed at the same price. I suppose this is to be expected with opening cross of orders.
http://www.investopedia.com/articles/in ... re-set.asp
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

Wow. A more than 4.5% drop in VTI since the market open.*

I have changed my mind about submitting market orders outside of market open hours. If one can see pre-market trading to guess on the opening cross trade price, then perhaps one can safely submit some market orders in the few minutes before the market opens.

*This is something that I think RodC's RBD study did not take into account.
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
alec
Posts: 3178
Joined: Fri Mar 02, 2007 1:15 pm

Re: Case study: Broker trade executions

Post by alec »

livesoft wrote:Wow. A more than 4.5% drop in VTI since the market open.*

I have changed my mind about submitting market orders outside of market open hours. If one can see pre-market trading to guess on the opening cross trade price, then perhaps one can safely submit some market orders in the few minutes before the market opens.

*This is something that I think RodC's RBD study did not take into account.
Isn't this the opening imbalance info that exhanges put out?
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

alec wrote:Isn't this the opening imbalance info that exhanges put out?
I'm not sure what you are getting at, but I did look up "opening imbalance" and I don't think so.

It appears to me that if one's order is executed at the opening cross, then there is no bid/ask spread. You get what everybody else is getting.

I can tell you that 381,400 shares of VTI traded today at the opening cross which was about 4% of today's reported volume.
Wiki This signature message sponsored by sscritic: Learn to fish.
Karamatsu
Posts: 1447
Joined: Mon Oct 27, 2008 2:42 am

Re: Case study: Broker trade executions

Post by Karamatsu »

Something I've noticed at Schwab is that there's a delay between when the actual time of a trade and the moment when it becomes visible through the web interface. For example, China and associated fallout dropped my foreign stock allocations below the rebalancing line the other day, so I put in some (limit) buy orders before market open, then watched for quite some time before they were reported filled. I could see, in fact, that my orders should have been filled because the market had drifted below my limits, but still the "order status" page still showed them as open. However, if I look at the order details, I find that all of the orders were filled in fact within the first three minutes. So there's a delay, and I'm not sure what accounts for that, but if you want to measure execution time it's probably important to look at the actual, rather than apparent, trade times.

Of course, the fact that there is a noticeable lag is a bit disturbing, but since I'm not a day trader or anything like that it's ultimately not a big deal. Also I frequently get price improvement, so whatever they're doing in the interval does seem to have benefits.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

I have not noticed any lag* in reporting executions at WellsTrade nor at TDAmeritrade. Vanguard is another story. It is even hard to find if you have an order at Vanguard after you have placed one.

*where "lag" is more than a few seconds.
Wiki This signature message sponsored by sscritic: Learn to fish.
TradingPlaces
Posts: 1245
Joined: Sun Nov 09, 2014 12:19 pm
Location: 30.286029, -97.530011

Re: Case study: Broker trade executions

Post by TradingPlaces »

livesoft wrote:I am doing another test this morning. I have submitted the same ETF market order BEFORE the open at two different brokers at the same time. I will report on the executed trades in a later post this morning. Normally, I would not submit an ETF market order when the market is not open, but my curiosity has gotten the better of me.

Update @ 9:38 am Eastern. Both orders executed at the same price. I suppose this is to be expected with opening cross of orders.
http://www.investopedia.com/articles/in ... re-set.asp
I just saw this post. I wish I had seen this earlier.

What you just did is an extremely bad idea.

I don't know how the brokers interpreted your pre-market dispatch, regular hours, market order. But let's say they interpreted this as "market-on-open".

Market-on-open (and market-on-close) are very specific, technical order types primarily supported by the listing exchanges: NYSE and Nasdaq. In both open and close, there is an auction.

It is widely known that trading on close has the following benefits:
- less private information,
- more robust volume,
- tighter spreads,
- supported by significant price discovery during the past 6.5 hours.

At the open, on the other hand, prices can be anywhere. A lot of mutual funds, hedge funds, market intermediaries, do not even turn their algos until the opening cross is known. Thus, you are getting a less robust price.

Also, from the point of view of market makers, open flows have a lot of private information (I don't mean illegal private information, but analyzed and predicted private information). Thus, in the morning spreads are wider, and by using a market order, you are paying more than you could pay compared to trading later in the day.

Furthermore, on volatile days such as the last few, the opening trades might have erratic prints. And if you don't get executed on open, your market order might just get an even worse print.

On this forum, a lot of attention is devoted to asset allocation, indexing, and passive investing.

However, one point of discussion that is very underserved is the role of market intermediaries and market liquidity.

How does Vanguard provide low tracking error to the index? Well, they execute well. But who do they execute against? Against market intermediaries.

Why are index prices determined by the underlying prices at the close, and not the open? Because close occurs after 6.5 hours of trading, allowing time for prices to be discovered, spreads to tighten, and more participants to be sure that there is enough liquidity to exchange large number of shares. The more shares are exchanged in a narrow band of prices, the better the liquidity and the more confidence that the price is good.

You executed VTI in the open, and got a worse price than if you had executed at close. There are two issues here: (1) did you get a good price at the open, and (2) was it a good idea to execute at the open, from market timing point of view.

Well, it is still possible that you got a good price on the open.

But the fact the close was 4.5% lower than the open has nothing to do with quality of execution.

It has everything to do with the fact that volatility is high, and such swings are expected. You could have easily gotten a 4.5% price better, e.g., if you had done that the previous day (Monday, Aug 24 2015).

Large number of stocks opened nearly 10% down on Monday, and during the day, went up quite a bit (although might have still closed down from Friday's closing price).

IMO, this shows that prices are difficult to predict, especially when volatility is high.

Separately, if you load VTI and SPY in the same chart on yahoo or google finance, you will notice that they traded very closely, which indicates that you did not get a bad price on VTI. You just timed it poorly.
Tanelorn
Posts: 2366
Joined: Thu May 01, 2014 9:35 pm

Re: Case study: Broker trade executions

Post by Tanelorn »

I agree with TradingPlaces - it's not likely to go wrong, but you were asking for trouble. What would your broker do if the primary exchange delayed their open for half an hour, like NYSE has done this week? Maybe they would send your order to the first exchange that was open at 9:30am, no opening auction, and you'd get filled right when spreads are the worst. Depending on the exchange and the size of the order, you might not get to participate in the opening cross, or your broker might not send your order there (which comes to the same thing).
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

Once again thank you for the comments. I think I am well aware of everything you wrote already, but it is good to see it again.
TradingPlaces wrote:Separately, if you load VTI and SPY in the same chart on yahoo or google finance, you will notice that they traded very closely, which indicates that you did not get a bad price on VTI. You just timed it poorly.
I have to chuckle about "You timed it poorly" because any other timing would have giving me a worse price for the day including prices 4.5% LOWER, so in fact, I got a GREAT PRICE on VTI since I was selling.
You executed VTI in the open, and got a worse price than if you had executed at close.
It seems that perhaps some readers erroneously think that the 4.5% drop was at/near the open. It was at the end of the day which was many hours after I had sold my shares. (I now see that you probably meant I got a "worse execution" and not a worse price.)

As for price discovery, I was watching the pre-market-open trades of VTI and IVV (like SPY) simultaneously before I placed my order. I had a very good idea of the prices that these shares were trading at.

This kind of trade is not something that I would do routinely, but I no longer fear it and would not get any anxiety from the comments to my trades. However, I do appreciate the comments, so please keep checking this thread out from time to time and continue posting here. Thanks!
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

TradingPlaces wrote:It has everything to do with the fact that volatility is high, and such swings are expected. You could have easily gotten a 4.5% price better, e.g., if you had done that the previous day (Monday, Aug 24 2015).

Large number of stocks opened nearly 10% down on Monday, and during the day, went up quite a bit (although might have still closed down from Friday's closing price).
But I did easily get a better price on Monday when I bought these shares. I didn't own the shares Monday morning, so there was no way I was selling at that time. :)
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study: Broker trade executions

Post by livesoft »

I just had another thought about this statement:
TradingPlaces wrote:A lot of mutual funds, hedge funds, market intermediaries, do not even turn their algos until the opening cross is known.
Perhaps this is good for the small retail investor because if their algos are not turned on, then their algos cannot be used to take advantage of the retail investor. Yes, I know these places claim they provide liquidity to the market, but they do not do it out of the goodness of their hearts. When 4% of daily volume trades in the opening cross, I think there is probably enough liquidity at that time for that stock or ETF.

Clearly, one should not be doing this without some idea of what the price will be for the opening cross.
Wiki This signature message sponsored by sscritic: Learn to fish.
TradingPlaces
Posts: 1245
Joined: Sun Nov 09, 2014 12:19 pm
Location: 30.286029, -97.530011

Re: Case study: Broker trade executions

Post by TradingPlaces »

livesoft wrote:
TradingPlaces wrote:It has everything to do with the fact that volatility is high, and such swings are expected. You could have easily gotten a 4.5% price better, e.g., if you had done that the previous day (Monday, Aug 24 2015).

Large number of stocks opened nearly 10% down on Monday, and during the day, went up quite a bit (although might have still closed down from Friday's closing price).
But I did easily get a better price on Monday when I bought these shares. I didn't own the shares Monday morning, so there was no way I was selling at that time. :)

I thought that your trade on Tuesday morning was a buy. If it was a sell, then you did time it very well, as VTI (and everything else in US Equities) dropped 4.5% from open to close.
TradingPlaces
Posts: 1245
Joined: Sun Nov 09, 2014 12:19 pm
Location: 30.286029, -97.530011

Re: Case study: Broker trade executions

Post by TradingPlaces »

livesoft wrote:I just had another thought about this statement:
TradingPlaces wrote:A lot of mutual funds, hedge funds, market intermediaries, do not even turn their algos until the opening cross is known.
Perhaps this is good for the small retail investor because if their algos are not turned on, then their algos cannot be used to take advantage of the retail investor. Yes, I know these places claim they provide liquidity to the market, but they do not do it out of the goodness of their hearts. When 4% of daily volume trades in the opening cross, I think there is probably enough liquidity at that time for that stock or ETF.

Clearly, one should not be doing this without some idea of what the price will be for the opening cross.
I am in favor of trading in my personal account when there are more market intermediaries in the market. The reason is that the true price of the security is more likely to be within the bid / ask price.

If you have private information, and your private information says that current price is ABOVE the ask, then yes, maybe it is better to trade / BUY when there are less market intermediaries.
boglerocks
Posts: 675
Joined: Sun Apr 03, 2011 9:08 pm

Re: Case study Broker trade executions

Post by boglerocks »

In my Scottrade account, I regularly see "last price" quotes above my ask limit or below my bid limit for OTC stocks. I brought a specific example of this to Scottrade's attention and I got this back:

"Although there were prints at your stated limit price or better, the market center holding your order did not participate in any of those transactions. The market center holding your order did not print anything at or better than your stated limit price. As such, no fill was ever due on your order."

Do I need a new brokerage?
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

@boglerocks, Thanks for posting in this thread. I know you discussed much of this in the other thread that you started: viewtopic.php?f=10&t=175827

As noted, the primary issue is probably that these are orders of OTC stocks which I am not familiar wth anymore. It is good for people to learn that you can't always get what you want.
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

I am going to place a market order before the market opens in a moment to buy some shares at the opening cross. I will report on the results in about 10 minutes. This post just serves as a timestamp for my intentions.
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

Ooops, my dog distracted me, so I missed the opening cross. Good thing, too, as market dropped a little bit lower. Anyways, I went ahead and bought some shares with a market order about a minute after the open. The execution price for the market order was 2 cents less than trades reported on either side of timestamp of my trade.
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

I am going to try to buy something near the open today, but not at the opening cross. I have submitted a limit order for a price that is 0.3% higher than yesterday's closing price of something that I expect to go up today.

Why 0.3% higher? Mostly because it will use up all the cash that I have leftover in that account from selling some BND shares yesterday AND because in another account I will sell the same number of shares of the same investment later in the day at I hope an even higher price.

Why sell then buy the same thing? I am making my portfolio more tax-efficient. Selling VSS in taxable (at a gain) and buying in tax-advantaged. Since VSS traded at a low price yesterday, then I will realized the lowest possible capital gain in quite a long time which helps perserve my carryover losses.

Why not just buy at the opening cross where there is no bid/ask spread? I think the opening cross may turn out to be higher than my limit price and I do not have enough cash to buy that number of shares at a higher price. Hmmm, just typing this out gave me an idea: I should place two orders: One for the opening cross for 80% of the shares I want to buy and one later after I see what the quote becomes. Nah! I'll just let this limit order stand and may have to change it if it is not executed.

I put this note here in this thread just for future reference to myself.
Wiki This signature message sponsored by sscritic: Learn to fish.
swl
Posts: 186
Joined: Wed Dec 09, 2015 1:27 pm

Re: Case study Broker trade executions

Post by swl »

You can submit a limit on open order that gives you the opening cross only if it uncrossed below your price.

Though ETF auctions tend to be pretty illiquid in general so not recommended for any sizeable trade.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

Thanks. The order was filled at the opening cross since my buy limit was higher than the opening cross. One can see that VSS is not particularly liquid so far today. I am not displeased with the result.

And so far, my "going up" expectation has not been met, so I will remain doubled-up on this amount of VSS until after that happens.

[Update} VSS is now trading up more 1%, so I think I will be able to sell those shares of VSS in the other account. "going up" is working. I'm going to be greedy and see how high VSS will go today before I pull the trigger on the sale.
Last edited by livesoft on Thu Jan 21, 2016 10:58 am, edited 1 time in total.
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

But this does suggest a way for ETF investors to transfer ETF shares between accounts in a frictionless transaction: Just do it at the opening cross. One is guaranteed the same sell/buy price and no bid/ask spread. :)
Wiki This signature message sponsored by sscritic: Learn to fish.
swl
Posts: 186
Joined: Wed Dec 09, 2015 1:27 pm

Re: Case study Broker trade executions

Post by swl »

That would probably be considered manipulative activity by the SEC for a professional trader. Don't know if that applies to retail but it's probably not a 100% safe idea :)

Can't you browbeat WellsTrade to give you free in-kind securities transfers? At least TDA lets you do free partial outbound transfers.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

WellsTrade does give me free in-kind partial outbound transfers without me asking for them.

These trades are being done at TDAmeritrade. I am trying to make a profit from these transactions and would not be satisfied with an even-steven transaction.

Also it probably matters that one account is a taxable account and the other is a 401(k). The VSS in taxable has got to go since only 49% of dividends were qualified last year. I replaced it with VEA (though I have yet to sell VSS in that account).
Wiki This signature message sponsored by sscritic: Learn to fish.
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: Case study Broker trade executions

Post by stlutz »

But this does suggest a way for ETF investors to transfer ETF shares between accounts in a frictionless transaction: Just do it at the opening cross. One is guaranteed the same sell/buy price and no bid/ask spread.
How does one know for sure that their brokerage firm is actually executing a Market-on-Open order and just a regular plain old boring market order a few seconds after the open? I know trading platforms like Interactive Brokers or (I'm assuming) Think or Swim allow you to specify a market-on-open order. One can't do that places like Fido, Schwab or VG, so how to you know that they are actually submitted a MOO order?
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

I only have today's data point that my order executed at the opening cross because I watched in real-time what was going on.
1. My trade was unambiguously large enough that it was not executed after the opening cross.
2. My trade price was the "Open" price quoted for VSS.
3. My trade time-stamp was " 01/21/2016 09:30:00 " My previous opening cross order back in August also has a 09:30:00 time stamp.

If I had waited to submit an order or if I had a lower limit order, then I could probably had a lower buy price since VSS traded lower for about 30 minutes or so. I do know the number of shares traded in the first few minutes after 9:30:00 were lower than the number of shares I bought. At the end of the day, I made a profit of about 1% and learned something in the process, so I am not complaining.
Wiki This signature message sponsored by sscritic: Learn to fish.
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: Case study Broker trade executions

Post by stlutz »

I only have today's data point that my order executed at the opening cross because I watched in real-time what was going on.
1. My trade was unambiguously large enough that it was not executed after the opening cross.
2. My trade price was the "Open" price quoted for VSS.
3. My trade time-stamp was " 01/21/2016 09:30:00 " My previous opening cross order back in August also has a 09:30:00 time stamp.
Thanks. I might have to try this in my IRA in the near future and see if the results are the same. Although I will unfortunately be transacting in smaller dollar amounts. :(
User avatar
DonCamillo
Posts: 1050
Joined: Tue Nov 26, 2013 9:27 pm
Location: Northern New Jersey

Re: Case study Broker trade executions

Post by DonCamillo »

After previous experience with other brokers, I have been happy with my executions at Fidelity.

I always use limit orders, and usually get the price I want within a few hours, often below my limit price for a buy and above it for the sell. Large orders are broken up into several executions. At times, I have gotten the best price for the day, and am often within a few cents. I normally adjusted my price by pennies to compensate for my $7.95 brokerage fee. Normal trades were round lots, 100 to 500 shares at a time, using bigger lots to save on fees. Fido did give me a lot of free trades at one time, but they were time limited and have now gone away. Since I have switched from being a dividend growth investor to a index fund investor, I have gone from about 20 paid trades a year to about 5. Now, I am mostly reinvesting my dividends in Fidelity Spartan Funds.
Les vieillards aiment à donner de bons préceptes, pour se consoler de n'être plus en état de donner de mauvais exemples. | (François, duc de La Rochefoucauld, maxim 93)
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

stlutz, this is for you.

While making trades in the past few days I had two interesting experiences.

1. I submitted a limit order to sell at TDAmeritrade. I was watching real-time quotes while eating ice cream and noticed that prices had marched up and trades were executing above my limit by a few cents. WTF? Why hadn't my shares been sold? I'm thinking "stuck in a dark pool". So I quickly changed the order to a market order and the shares get sold immediately at a few cents higher than my original limit price. Executions after mine were at lower prices.

2. I wanted to sell shares of VSS. This is my lowest volume holding and with the highest spreads. Yesterday the spread was rather high at about 15 cents. i had some time on my hands waiting for my family, so I submitted a sale order with limit price 1 cent above the lowest ask (not a marketable limit order). I watched real-time quotes and nothing much happened for 5 minutes, then the best bid went up 1 cent and the lowest ask went up. Now my limit price was the lowest ask, so I changed my limit price upwards to once again be one cent above the other best ask. After a minute, the best bid went up 1 cent and the other asks went up as well thus maintaining the spread while waiting for someone to submit a market order. This repeated a couple more times and each time I raised my ask. Then I got a partial fill, so I raised my limit price once again and the rest of the shares sold. Then I saw the trades on either side of my last fill to be lower by about 5 cents. So who paid the spreads?
Wiki This signature message sponsored by sscritic: Learn to fish.
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: Case study Broker trade executions

Post by stlutz »

Interesting. In my orders at least I've found that when I put in an order that isn't marketable to start, but the price drops (in the case of a buy), it will execute immediately. Actually when I bought some IXUS earlier in the week (has spreads similar to VSS), once the order became marketable it executed inside the spread (i.e. I got a price better than my limit) and then the price went up shortly thereafter.

I got lucky on that one as prices never dropped back to my original limit the rest of the week.

In general I can't say I've ever seen anything unusual in my trades at Fidelity.
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

livesoft wrote:1. I submitted a limit order to sell at TDAmeritrade. I was watching real-time quotes while eating ice cream and noticed that prices had marched up and trades were executing above my limit by a few cents. WTF? Why hadn't my shares been sold? I'm thinking "stuck in a dark pool".
Regulations require that your order be placed on a bright exchange. You should have called TD for an explanation. The only non-exceptional reason I can think of is that your order was an oddlot (<100 shares), and in those situations people aren't required to access your sell order (in fact, the most basic quote feed doesn't show them).
So who paid the spreads?
I'm not sure what the question is here..obviously a buyer was paying the spread in your situation.
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

stlutz wrote:How does one know for sure that their brokerage firm is actually executing a Market-on-Open order and just a regular plain old boring market order a few seconds after the open? I know trading platforms like Interactive Brokers or (I'm assuming) Think or Swim allow you to specify a market-on-open order. One can't do that places like Fido, Schwab or VG, so how to you know that they are actually submitted a MOO order?
As a client you are guaranteed the opening price. Generally, there's too much risk for a broker to execute your order after the opening, instead of in the cross, since if it goes the wrong way, they're out the difference. Generally, the broker is sending your order to the primary cross.

In general, any market order submitted before the open is eligible for the cross (some timing considerations come into play). MOO orders, as opposed to a regular pre-open market order, just means that if for some reason you don't execute in the opening cross, your order will be kicked back. One example of this could be if the stock tripped a breaker of some kind and the opening price would have been too far out of whack.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

ftobin wrote:Regulations require that your order be placed on a bright exchange. You should have called TD for an explanation. The only non-exceptional reason I can think of is that your order was an oddlot (<100 shares), and in those situations people aren't required to access your sell order (in fact, the most basic quote feed doesn't show them).
Thanks for the comments.

Order was more than 400 shares in VEA so highly liquid, but not a rounded hundred. Often, I will break up an order of something like 630 shares into a round lot (say 600 shares) and a sub-100 odd lot (say 30 shares).

As for calling TDAmeritrade to complain, I think I would only do that if I got a worse price than I wanted. :twisted:
Wiki This signature message sponsored by sscritic: Learn to fish.
Longtermgrowth
Posts: 731
Joined: Thu Nov 26, 2015 12:59 pm

Re: Case study Broker trade executions

Post by Longtermgrowth »

I bought some VXUS today. Decided to use Java to run MarketPro. My first limit order was at 44.81, but it shot up a few cents as I placed the order, so after an hour I decided to raise it to 44.85 since that was within the recent price fluctuations. Within a few minutes it dipped down to there and executed. It came down just a hair further, so I bought 400 more shares at 44.84. The 2nd order was broken up into five different amounts, two less than 100 shares and one of the smaller amounts bought for 44.835. It took a few minutes after placing the order for it to fully execute.

Anyway, seems to have worked pretty good! I like seeing the semi real time bid/ask and trade execution amounts in MarketPro. Hope this information is of some use to you, livesoft.
gips
Posts: 1752
Joined: Mon May 13, 2013 5:42 pm

Re: Case study Broker trade executions

Post by gips »

livesoft, late to this thread. exchanges post time and sales information and some brokers provide the information to their clients. Based on this information, you should be able to understand why your did or did not execute. Also, for moo and moc orders, brokers will get fined by the exchange for not handling them properly so there's a big incentive to get it right. Obviously, if you can't denote the order as moc/moo via the broker interface, it's simply a limit order. at least that's how we handled it when we created an interface for a broker.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

Just a note to myself:

Some lower volume small-cap emerging markets ETFs have only had one transaction this morning -- namely at the opening cross -- and no subsequent trades 9 minutes hence. I don't believe "price discover" has really happened.
Wiki This signature message sponsored by sscritic: Learn to fish.
Post Reply