I have designed a countertrend trading system that makes entries at Close of the day and holds positions for a few days. The exits are also at Close. The average profit is about 2% and the frequency of trading is about 5 entries (in different stocks) at a given day. The stocks I am concidering are mainly from Nasdaq and a smaller portion is listed in Nyse. I am a bit confused as to how I should trade such a system in real time. The first idea is to wait for the closing prices and then make the entries during the afterhours session. The second choice is to consider the prices of 15:50 as "closing prices" and place "MOC" orders before the close. With this scenario I would "lock" the actual closing price, however I would also initiate some invalid trades because of the amateur "closing prices" in the calculations of my system. So, I am conidering to go with afterhours trading, and here I need your advice! I am thinking of placing afterhours limit orders hoping that I would eventually capture the same price as this of a MOC. Note that this is a countertrend system, so most of the times my orders will be in opposite than the trend of the day (and I only enter when there is a trend). Do you consider that I would be able to find the Closing price, in afterhours trading, with a success rate of above 90% ? Besides the price, I also am thinking of liquidity difficulties trading afterhours. Some of the selected stocks have daily volumes of just 100k. The good thing is that I trade each stock with just 100-500 lots. Do you think that I can initiate such trades afterhours, and find the MOC price at about 90% of the times ? Many thanks! RapidEye
Can you reverse the logic in your system so that you can put a limit order into the market as a Good Til Canceled order? That way if the closing print matches your system's criteria, you should be entitled (at least on the NYSE) to the fill. It's a lot of orders, and may not be possible given Buying Power limitations. In that case, you might have to limit your orders to only those whose prices are close. As to your question about getting filled on the MOC price in the aftermarket, well that's a hard one. If your trend that you are countering is a strong one, then there's a better chance that the aftermarket will go through your limit. In many stocks with low volume, you may hold the market back with your limit order, especially in the slower days of summer. The best way you can get a feel for that is to watch some of those 100k stocks trade. They'll vary, but you should get a good overall idea.
For stocks with daily volumes of 100k, I doubt you will get hit 90% of the time. I would say 50% max. For more liquid stocks, your rate would probably go up but probably not to 90%. Why don't you try it and see? You could also try sending your market orders in at 3:59:59pm EST for Nasdaq stocks and sending MOC orders for NYSE stocks. For 100k volume stocks you are going to give up a big spread on Naz stocks with market orders.
I now have read everything I could find about afterhours trading.. and I understand that this would be extremely difficult. So I am considering the preclose solution again. Actually, with nasdaq I can send simulated MOC or LOC orders through IB. *The problem is with Nyse*, since MOC orders for listed stocks have to be sent before 3:40, wich is too early for me. Most of the stocks I am talking about have very big volumes. There is about a 2% of the total stocks that have volumes bellow 100k, which afterall I could exclude them if the spreads will be so bad. However, there is also a percentage of 5-10% with volumes around 300k. In my simulations I have considered an average spread of 2 cents per side as slippage costs, having in mind that these would be trades at the closing, where liquidity is going up. Could someone provide an expectation of spread costs for market orders around 3:55 in Naz and Nyse stocks, taking into account the average volume? Thanks! RapidEye
If you send a market order a few seconds before the close, you'll probably get the closing price, and you'll know pretty closely what it will be. This would be a problem for shorts though because you won't get a fill if they close on a downtick.
Seconds before the close may not get you a fill on NYSE/AMEX. I'd give the spec at least a minute. Do your scan around 15:59 and send your orders then. You'll get something pretty close to the closing price, though stocks with really large MOC imbalances can sometimes surprise you. It's hard to imagine that the system only works if you get the exact closing price.
Thanks for the replies. I did some simulations on intraday data and I can be sure that the system works even with 15:55 "closings". Can someone provide some tips about my other question: "In my simulations I have considered an average spread of 2 cents per side as slippage costs, having in mind that these would be trades at the closing, where liquidity is going up. Could someone provide an expectation of spread costs for market orders around 3:55 in Naz and Nyse stocks, taking into account the average volume? " What I ask for is something like: "1) for 100k-300k stocks expect to pay a spread of about X cents at 15:55 2) for 300k-1m stocks expect to pay a spread of about Y cents at 15:55 ..." Many thanks
This is a very similar problem I was tackling when I was doing tests. The best estimates I can come up with: if your average volume 100 - 300k for the day the spread may be as high as 3 - 10 cents. for 300 - 1m stocks it may get a bit smaller 3-8 also, more expensive the stock is - the larger the spread. My best recommendation to you is to take about 20 - 30 potential stocks you are going to trade and watch them end of day on your Level II. You also have to take into account the size factor. For example, there may be a 2 cent spread with 100x100 size and you are planning to use 1000 lots. then look at a more realistic fill - look down through level to adjust for size. I would also recommend doing some forward tests trading 100 shares - but I am not sure if 100 shares would be too much for a forward test in your case. I think I spent and spend most of my time not designing strategy but reducing spread costs. I hope I helped.