Building Your Own Methodology: Now Comes the Hard Part

Congratulations! You're almost there! You've stuck it out through every one of my Methodology columns, and you've finally come to ... the really hard work segment.

By now, though, I'm confident you realized there wasn't going to be any free lunch. No, this trading stuff is hard work, no matter what style you embark on. But, if you want to be good and you want to last a while in trading, you'll need to work hard all the time. Today is just step one.

And basically that step entails a four-step process of backtesting, paper-trading, small-scale trading and then real trading. I know, I know: Most of you started with real trading. You probably wished you hadn't, but you did. So, perhaps a few of you now have to paper-trade. Well, next time you'll know. Everyone else out there who wants to jump right in before testing things out, consider yourself forewarned.

Now, I started with backtesting because I really believe that if you're going to trade with a rigorous methodology, your methods should be able to work in the past as well as in the future. So backtesting -- or applying your proposed methods against historic data -- should be your starting point.

But here's the catch: Despite the wealth of historic data and despite the raw MIPs of computing power available, it is nearly impossible to backtest any methodology across a broad range of stocks! I know, flabbergasting, isn't it?

As an example, with the Omega TradeStation package, you can backtest just about any method you choose against years of historic data. The problem: You can only do it one stock at a time. In other words, say my method is to just buy the next day after a 52-week high, and then sell for either a 1-point gain or 1-point loss, whichever comes first. Well, I can't test that against all 10,000 stocks for the past eight years. No, I can only test that against IBM (IBM Quote) for the past eight years. And then, if I'd like, Intel (INTC Quote) and Microsoft (MSFT Quote) and Dell (DELL Quote) and MCI WorldCom (WCOM Quote). Frustrating, right? (Vendors, if I'm wrong, PLEASE correct me!)

In addition, it is extremely difficult to test various money-management parameters and position-sizing. In Van Tharp's book Trade Your Way to Financial Freedom, he talks about scaling your position size based on total equity, volatility, risk, etc. It's an important concept that can make a significant impact on your trading. However, it's very difficult to test. (Sidebar: In his book, he does talk about a software package that can test various money-management scenarios. I've not looked at it, but if you want to investigate it, it's called Athena, and there's information at www.iitm.com.)

So, backtesting is a problem, but you have to start somewhere. How did I do it? I tediously went through a few thousand charts on TC/2000, handpicking the ones that met my buy criteria and then figuring out if my entry and exit points made sense. Not fun by any measure, but extremely helpful.

Of course, the other way to test your methods is in real time with real data. And that brings me to paper-trading.

Paper-trading, if done properly, is an excellent way to bulk up the amount of data you collect on your methods, as well as get you just a bit closer to real trading.

But before you jump in and paper-trade, the first thing to do is write down the details of your trading plan. This will help you stick to a routine structure each day, and it will eventually act as a reference point when you drift off course. Your next step is to figure out a way to keep track of your trades, simulating your P&L statement. Early last week, I received a helpful email from reader Greg Starks:

Gary,

I've found MarketPlayer.com to be an excellent resource for paper-trading. You can set up portfolios with realistic amounts of capital to simulate your real buying power. It keeps track of commissions (again, a value you set), margin interest, etc.

You place an order and get filled. Whether or not the fill is always realistic is not really the point. The point is that it prevents the paper-trader from even subconsciously working things in his or her favor. You play the hand you're dealt. And you set your stops and limits accordingly (unfortunately, you can only set one at a time, so you have to hang loose until you get an idea of which direction the stock is going. And you have to keep watching in case it changes its mind).

It also very nicely keeps track of gains and losses, and plots your performance against the S&P 500.

I've been paper-trading the GBS classic since the first of the year on MarketPlayer, and it has proven to be an invaluable resource. I'm able to go back and look at historical trades and analyze good trades vs. bad trades vs. just plain bad luck.

I highly recommend it.

OK, that details one effective way to go about paper-trading. In addition, though, or as a substitute, I've always used a spreadsheet.

After you have your record-keeping structure intact, it's time to do some trading. This is the tricky part and will involve a little bit of guesswork.

The guesswork comes in the area of fills and concerns something that 99.9% of traders miss: slippage. What's slippage? It's having a target price of $27.50 to sell but only getting filled at that price on 200 shares of your 2,000-share lot, with the remaining 1,800 shares filled at $27.25. Or, it's having a stop of $50 and having the stock gap down and open at $47 with you getting filled at $46.75. Oh, but you say slippage isn't a big concern for you because you have a good broker, right?

Well, I think I have a good broker too, but let me give you an example how even minor slippage can damage a great system.

A few weeks ago, Wesson came up with a set of scans and money-management parameters that returned 100% per year. And that wasn't 100% average, that was 100% compounded! Wow, right? No, double wow!! We couldn't wait to try it out.

Ha! The joke was on us, though: If we added in about 0.4% of slippage (not 4%, but 0.4%!), the returns dropped all the way below 20% per year!

But, you're thinking, "Wow, 20% a year compounded, that's fantastic! Sign me up." Whoops, now add on commissions and any margin charges you might incur. You want the code? You can have it!

So, trust me, you're going to get slippage, and it will negatively affect your system. Of course, the less trading you do, the less of an impact it becomes. But, it affects everyone. Yes, even if you use limit orders, you're affected. How so? Gaps that go right by your limit price. Their effect? You're then either missing the trade or getting tagged with a large loss when the stock goes by your stop. Sorry, there's no way around it: Slippage is an insidious creature.

Therefore, when I paper-trade, I'll normally go through and eyeball a reasonable ask price where there's enough volume for me to get filled. Going long near the open, there's generally not much slippage, but of course, your risk of a crummy fill is probably greater. Using buy stops is fine, but even then, you'll get some slippage as you're always the next trade to go off after your stop is hit.

However, going short is a different animal, and it's complicated by the uptick rule. Essentially, if a stock opens down and keeps going down, you can't go short until the first uptick. This morning, in fact, a stock I shorted opened at $36.50, but I wasn't able to get in until it hit $36 even. That was a nice 1.3% slippage and over time can turn almost any "high turn" model to trash.

So, when I simulate my short trades, I usually look at a one-minute intraday chart and, again, try to eyeball the first uptick. Usually, I can come pretty close to paper-trading what my actual fill would be.

The key, though, with paper-trading is to take it seriously. Don't fudge your numbers or go back and erase stupid trades. I've done both, but in real life it just never works out that way. No, treat this type of trading just like the real thing. Of course, it isn't, and to actually put you and your system through the grinder, you'll have to start using real money. So, next week, we'll go live, and I absolutely promise to wrap up everything in that final column!

---

Attention, attention: all GBS fans and all non-GBS fans: Think I'm full of it? Think I walk on water? Think Cramer is 'da Man, and I am 'da Bum? Well, you're absolutely right, so come one, come all and voice your opinions on my Yahoo! Chat Tuesday evening at 5 p.m. EST. As a special treat, I'll have my lovely daughter Katherine at the helm, plugging in charts so I'll actually know what I'm talking about! Hope you can join me!