Technician's Take
Testing the Cooper Day Trading Method
Gary B. Smith
01/20/98 - 08:24 AM EST
Editor's Note: This is the third installment of a three-part series on day trading. Click here for part one and here for part two. However, Gary is sufficiently intrigued that there may well be more columns on the topic to come. ***** Well, the Cooper system worked. End of column. Have a nice day. Boy, if only things were that straightforward. Of course, they're not, but -- DARN IT! -- couldn't they just be ONCE in our lives?? That being said, let me spend the rest of the time explaining the opening sentence. My test started on Dec. 1, and concluded on Jan. 13. As I mentioned in last week's column, I had read both of day trader Jeff Cooper's books and if I had had the right software (TradeStation or Supercharts), I could have bought an add-on package and generated the same "hit list" he works off. But I much preferred to get Jeff's picks directly from Jeff, and fortunately he has a fax service where he sends you the ideal candidates for the next day. Actually, he has three fax services: a list of small-cap stocks; large-cap; and a list of stocks based on his 5 Day Momentum book. How the candidates are selected, though, is basically the same. Jeff scans the universe of stocks and, according to the chart patterns he likes, culls out the probable winners. He then narrows the list by looking at stocks that usually show a good deal of "movement." As a result, each day I was presented with a list of 15 to 20 candidates. These were generally divided equally among longs and shorts, although at times -- reflecting market conditions -- they were heavily weighted on one side or another. In early January, for example, there were a good deal more shorts than longs. Ideally, I would have acted on every one of Jeff's recommendations, gathering a fairly large sample set to judge final results. From the start, though, while I agreed with most of his picks, some just didn't look right. Now, I admit, "look right" is not something easily subjected to rigorous scientific testing. All I can say is look at about 200 charts a day. You'll know what I mean. So from Jeff's list, I'd select about eight or 10 and put orders in for those based on Jeff's methodology: 1/8th above the previous high for longs, 1/8th below the previous low for shorts. In addition, I also threw in a couple of my own selections, but based on similar patterns to what Jeff looks for. I did adhere to his trailing stop methodology, which called for me to start with a maximum 1 point stop-loss, but move that stop-loss to break-even as soon as the position showed a 5/8th- to 3/4th-point profit. After that, if the stock continued to rise, I'd follow along with my trailing stop. Finally, if the stock closed near its high of the day (for longs), or low of the day (for shorts), I'd carry it over to the next day. This occurred about 20% of the time. Okay, enough stage-setting -- even my eyes are starting to glaze over. Here were the results: Total number of closed trades (Cooper and Smith picks combined): 99 Total number of wins: 58 (58.59%) Total number of losses: 41 (41.41%) Average point gain per win: 1.21 Average point loss per loss: 0.71 Average point profit per trade: 0.41 Total net points during test: 40.59 Total number of closed trades (Cooper only): 34 Total number of wins: 22 (64.71%) Total number of losses: 12 (35.29%) Average point gain per win: 1.11 Average point loss per loss: 0.64 Average profit per trade 0.5 Total net points during test: 17 Okay, not too bad. As you can see, Jeff's picks were slightly better than mine, but not significantly so. (Note: Jeff was on vacation for the last two weeks of the year, so he generated no picks during that time.) More winners than losers, which is good. Even better: The average win was significantly higher than the average loss. Bottom line: On 1,000-share lots, this system would have generated $40K net with our combined picks, and $16K with Jeff's picks only. (Commissions, though, are NOT included.) I can certainly live with those kind of results. Okay, all this is good news. Now, some gotchas and random thoughts. Sample size You could argue that with only 99 trades, I didn't have a large enough sample size to really test the system. I can't argue with that. However, I received some expert assistance from reader Marsten Parker, who was able to backtest Jeff's methodology using MetaStock. His results were fairly consistent with mine: a higher win rate (68%), and a win size about 50% higher than the loss size. Fills Just paper trading the system would not have yielded accurate results. On the other hand, in testing a system, I didn't exactly feel comfortable trading 1,000-share lots. Therefore, I made all trades with either 250- or 500-share lots. Realistically, I might not have received the same exact fills with 1,000-share lots, but in looking back at time and sales on a number of trades, I don't think it would have made a difference. In fact, some of my fills on the long side, and particularly on the short side, were terrible. So, unlike a paper-traded system, I think the results are accurate. Online trading One of Jeff's rules is that if the stock gaps more than 3/4ths of a point above your entry price, then pass on that stock. And that's a good rule. To adhere to that, then, I needed to enter an order with two parts: a buy stop and a buy limit. As an example, say IBM was a long candidate and it closed the previous day at 100. The buy stop would then be 100.125 and the buy limit would be 100.875. Therefore, once IBM traded at 100.125, my order was executable at any price below 100.875. On the other hand, if IBM opened at 101, then I wouldn't have to pay for the "gap" opening. But here's the problem: A lot of online brokers don't allow you to enter an order that way. Ameritrade (which I use) doesn't, and neither does E*Trade, Fidelity, DLJ and many others. That's a problem and if your online system can't handle it, you then have to watch the opening prices. And with 10 candidates in play, that takes concentration. But, let's say you could keep an eye on all 10, and enter orders only for the ones that don't gap. Well, problem two: Even the swiftest of the traditional online brokers are dreadfully slow. This speed, or lack thereof, probably cost me 1/8th point per trade. It doesn't seem like a lot, but on the 99-trade scenario, that's over $12K. That's a BIG difference. The only solution: If you're going to day trade, and really want to maximize your performance, you have to either find a broker that takes both stop and limit orders on the same position, or find a broker that can link you directly (well, almost directly) to the floor. Do they exist? They're a niche, but yes. Try AB Watley (www.abwatley.com). Fills are within seconds, and they display bid/asks of all the market makers for the Nasdaq stocks (i.e. Level II quotes) that allow you to get a slighter better price at times. Watley isn't as cheap as an Ameritrade or Suretrade, but they're less expensive per trade than Schwab and some of the others. (Note: they do have higher monthly charges, however.) Is the speed really necessary for a day trader? Let's put it this way: I'm definitely considering a switch. The Market December and January weren't exactly the easiest months to trade, but that's when you really want to test a system. In general, then, I thought Jeff's methodology held up fine. The only time I questioned it was during the gap/whipsaw openings that we saw in early January. My picks would execute and seemingly within minutes, the stock would reverse and hit my stop. Believe me, it's past frustrating to lose $1K in the space of 20 minutes. On the other hand, I don't know any way around it, other than to avoid the open when the S&P futures are way up or way down prior to trading. Shorting Shorting is always a pain. You have to wait for an uptick or the stock isn't available to short. I missed some real winners on both counts. You could get around this with some fancy married option strategies, but you probably don't want to go there. No, it's one of the harder parts of trading, and I don't know anyone who's cracked the code. Time Still my No. 1 complaint about day trading. Yes, there's some downtime between 10:30 and 3, but basically you're chained to the tube. On the days when I was killing it, it didn't bother me in the slightest. I really miss my noon workouts, though. (Note: Jeff's "5 Day Momentum Method" does not require you to be in front of the tube all day. It's certainly an option.) Being Flat Maybe the biggest advantage -- period -- to day trading. There is no better feeling than waking up, finding that the market will open down a billion points and having zippo positions in play. Conclusion I'd have to say Jeff's come up with a winning methodology. But, if you think about it, it almost has to work: He only takes a position if momentum reasserts itself, and while the maximum loss he incurs is 1 point, maximum profit is unlimited. (That said, my biggest gain was only 5 1/2 points.) Stack those odds in your favor and you should get exactly what I found: more wins than losses; and the average win being greater than the average loss. But Jeff was right, there aren't a lot of home runs here, and you spend an awful lot of time some days scratching out a measly 3/4 of a point. On the other hand, the biggest point loss I had in any one day was 1.125. All in all, a reliable, solid way of trading. Just keep showing up, and executing, and yes, you should be able to pound out a pretty good year. But, are there better strategies out there? Perhaps. Is this style of trading for everyone? Of course not. And after all this, is Jeff's day trading method, or anyone else's, for GBS? Stay tuned: I thought this saga was over, but the journey is turning out to be more educational (and hopefully profitable) than I first imagined! * * * * *