• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

Large Volume Trading

Joined
8/6/08
Messages
19
Points
11
Do any firms or people do a lot of large volume trading in a day? I guess I am getting at day trading but selling when the price only goes up a small amount (like 5 cents). Do companies or day traders do this? Is there a reason you wouldn't want to use this strategy? It seems to me if you set your threshold low enough you would come out on top. Maybe this is exactly what day traders do -- I don't know anything about day trading. Sorry if this is an annoying question!

Jeff
 
Those are high frequency statistical strategies, and they are out there. However, they will have some additional constraints (mean reversion, momentum, resistance, etc.). One thing that becomes very important here is speed, trading costs, and the spread. Even if a strategy exists, it may not be feasible given those restrictions. To lower the latency and cost threshold, this is restricted to people who can make a big investment in equipment (I believe Citadel, for example, has their trading machines on the exchange floor.)
 
So are you saying the average investor can't move fast enough to use this strategy? Let's say I buy GM stock at 10.02 at 10AM and at 2PM I see the price is now 10.09. If I were to sell at this price what kind of error can I expect to see in the final price? Is +/- $.01 reasonable or is it more along the lines of +/- $.05? Thanks

Jeff
 
If you buy GM to make 7 cents profit after 5 hours, you are in trouble and should not consider trading.

A 7 cent move would take minutes or seconds even with a 10 dollar stock like GM.

People employ strategies like this where they try to make pennies. You only need to make around 1.5 cent a trade depending on your arrangement with the prime broker for commission fees to cover your trading costs and ECN fees taking into consideration whether you add to liquidity or take away from liquidity (put in a bid on ISLD as opposed to taking the offer, or putting an offer on ISLD as opposed to hitting the bid).

There are some day traders who purchase for example 5K to 10K to even 20K shares or more and look to make pennies.

You can only do this on stocks where the spreads are very tight and there is large size on bids and offers at every level of the price, and of course the volume is very high and can absorb your size easily.

It's a viable strategy but it can hurt you bad if you're wrong on the trade because of the size you're dealing with and most likely, once you get accustomed to trading such large blocks, you'll find your commissions are gonna get very high.

I would rather trade, JMO, a more illiquid stock with larger spreads, a higher price, and trade maybe a few hundred to at most a couple thousand shares and make 50 cents or dollars on a trade. This is easier IMO because the smaller size lots are easier to get out of if you're wrong.

For example, a stock like Google which is like 500 dollars you can easily make points at a time trading a lot size of 100 to 300 shares.

You would not want to trade more than 1K shares of Google unless you are a really big gambler or high roller because you could easily lose a few thousand on a move that takes seconds.

I don't think most funds (the "pros") day trade although I have seen buy side traders give orders to on the same day when a stock just so happens to have made a large move right after they bought. Usually they give out market orders or limit orders at a certain price and park the trade for months if not weeks. It's really difficult and doesn't make sense to day trade when you're running 100's of millions or billions.
 
I have a few questions for you. What do you mean by "large spread" or "tight spread" ? Maybe you answered my question but I don't see it. When you sell at say 10.07 what kind of error can one expect to see in the final price?

I am also confused by you saying "If you buy GM to make 7 cents profit after 5 hours, you are in trouble and should not consider trading." but then you say " It's a viable strategy..." I am not trying to argue with you but am just trying to better understand your points.

Do most brokers charge fees that change with the size of the deal instead of one constant fee?

Would this strategy be better suited in the FX arena because there is more liquidity there?
 
I have a few questions for you. What do you mean by "large spread" or "tight spread" ?

Bid-ask spread. Tight spread means that there is a a lot of liquidity, and it is very cheap to get in and out of the position

I am also confused by you saying "If you buy GM to make 7 cents profit after 5 hours, you are in trouble and should not consider trading." but then you say " It's a viable strategy..." I am not trying to argue with you but am just trying to better understand your points.

Making 7 bps in the entire trading day is a poor return on your money given the risk.

Do most brokers charge fees that change with the size of the deal instead of one constant fee?

Institutional brokers charge per-share in the US. I believe in Europe (and elsewhere?) they charge a fixed percentage of the trade amount.
 
When you sell at say 10.07 what kind of error can one expect to see in the final price?

If you get filled at 10.07, the only difference in your P&L is the cost of executing the trade.

If you put the order in when you see 10.07 go across the tape, that's another question, and it depends on how quickly the price is moving.
 
I guess I did not make my point clear. I of course realize 7 cents is not a good return for one day but if you are dealing with 10-20k stocks that adds up.

If I understand this correctly, many brokers charge a 1.5 cent fee per stock? In other words if I made 7cents per stock I would really walk away with 5.5 cents.

Thanks for all the responses. I am just trying to learn more about how all this stuff works.

Jeff
 
if you are dealing with 10-20k stocks that adds up.

I don't follow your reasoning. A poor return on 1 share is the same as a poor return on 10,000 shares, unless you are leveraged.

I didn't think you were trading just 1 share; I am thinking percentage-wise.
 
I don't follow your reasoning. A poor return on 1 share is the same as a poor return on 10,000 shares, unless you are leveraged.

I didn't think you were trading just 1 share; I am thinking percentage-wise.

You are right in that its a small percentage increase; however, since the growth is geometric that small percent grows pretty fast.

Is there a problem in my logic? I feel like there has to be somewhere but I can't find it.

Jeff
 
You are right in that its a small percentage increase; however, since the growth is geometric that small percent grows pretty fast.

Is there a problem in my logic? I feel like there has to be somewhere but I can't find it.

Jeff

It sort of sounds like you're trying view the market as guaranteed profits as long as you play within certain confines such as profit limits of 7 cents and expect those profits to be recurring.

I'm sure you can create your models using those assumptions. As to whether they would work recurringly for your model to be effective is something I don't altogether believe in although I've never seen it done.

What I'm trying to say is, in school the professors seem to sort of teach their students as if the market can be quantified and then create formulas/models that would create your profit expectations realistically. With portfolio theory and expected returns using stuff like beta/alpha I agree on. On the whole, I believe they wouldn't be there especially many during market hours teaching a class if they could make money themselves unless they just love to teach.
 
You are right in that its a small percentage increase; however, since the growth is geometric that small percent grows pretty fast.

Is there a problem in my logic? I feel like there has to be somewhere but I can't find it.

Jeff

5 to 9 cent profit on a stock is pretty easy to do.

That's because that's like betting money on which direction a dandelion will swing in a light breeze.

Stocks move 5 to 9 cents all day long one way or the other.

The problem with this strategy, IMO, is that you'll be right statistically at most half the time. Maybe a little bit better than half but not by much.

So at best, you'll break even or lose money on your total P&L because of the commissions.

Now if a trading model can be created that can accurately be on the right side of these small little moves, I guess it can be done.

For example, a quick day trader can do this kind of stuff based on instinct and what you see on the bid and offer.

Usually when you see blocks of trades go off really quickly on the offer, they are in size, and at the same time see bids stacked up with large size while the offers are thin, that usually means the stock is about to make a move up or already moving up. At the same time, you have to judge the price on where it is at with respect to how it has been trading. Discretionarily, the trader has to calculate all of these things and take a quick (usually gut/instinct/feel) decision whether to enter a trade or not. The model would have to do all of these things as well in order to be effective, imo, which would be difficult since humans still get close to 50% right at best.
 
I traded equities for 3 1/2 years. Your biggest hurdle in the strategy you describe will be transaction costs (and if you're trading Listed, a fair fill -- or at least it used to be).

A prop desk at a legitimate member firm is the only way to pull it off. A smaller prop shop will eat you alive.
 
What about an online place like TradeKing? It looks like no matter what the size of the trade it is always 4.95. Can they not handle trades with 50k shares?
 
What about an online place like TradeKing? It looks like no matter what the size of the trade it is always 4.95. Can they not handle trades with 50k shares?

I can't say because I've never heard of them. And things have changed dramatically since I left the business in '04.

I would have given anything for a flat rate back then. I'd be concerned about quality of access to various ECN's and exchanges. The fine print is probably interesting. They're not running a charity.

A member will always have the edge in cost.
 
you need to buy a seat in the exchange... and, nowadays, it will cost you a pretty penny
 
Here is historical membership price on NYSE

seatsalechart1.jpg
 
Just trade the OTC NASDAQ. No seats and even MM's use the ECN"s nowadays to hide their cards. You can become the ECN and add to liquidity. That's the most democratic market available. It really is a trading forum like this forum quantnet. It's just that the "server happens to be in NASDAQ NYC, but it is not a centrally located market where a great bulk of the volume is done by the seat as in the NYSE. But even in the NYSE, you can still trade with an ECN. Subscribe to an island book feed and you'll see what the orders look like on every price level way out of the market on both sides anyways. Everything is going OTC nowadays. With computers and networks, what's the point of buying seats. The mercantile exchanges is going the same route.
 
Thanks for all the replies. They have been extremely informative. Would dealing with shares in excess of 50k give me liquidity issues in the NASDAQ?
 
Back
Top