In an older article I showed you a simple trading model that took advantage of the overnight edge on the S&P market. The article was entitled, “S&P Overnight Trading Model“, and it traded a single futures contract. The final results of my tests produced a model with the following results:
S&P Overnight Model Results
Results | |
|---|---|
Net Profit | $10,590 |
Profit Factor | 1.95 |
Total Trades | 167 |
%Winners | 65% |
Avg.Trade Net Profit | $64.41 |
Annual Rate of Return | 2.50% |
Max Drawdown(Intraday) | $1,333 |
Expectancy | 0.33 |
Expectancy Score | 3.92 |
Can I Trade This System?
While the results are decent and the trading rules are simple, this is not a trading system. The question of, can I trade this system, is one I often receive and respond with a simple, no. It’s true the results holds promise that a trading system could be developed from this basic model, but we don’t have a trading system as it currently stands. This is true for many of the articles I write. Often these articles contain trading models, ideas or market edges that can lead to a profitable trading system. But as they currently stand, they are a far cry from a thoroughly tested and developed trading system. The intent for these articles is to help inspire and point you in a direction. This site has always been a great source of ideas for those wishing to create profitable trading systems – not necessarily a site where you are given a complete trading system.
There are many reasons why complete trading systems are not presented. Outside of liability, trading systems are often very personal creatures. What might work well for one person may be far too risky for another. Each trading system must match the personality, psychological profile and goals of the trader. Getting this from a pre-build system often does not work. But I’m really getting off topic here.
What’s The Difference Between “Trading Model” and “Trading System”?
This brings me to my first point I would like to make. I would like to point out that over the last few weeks I’m making an attempt to be more clear in my vocabulary. When I say “trading model” I’m referring to a set of rules that determine buy/sell/stop conditions. The model is designed to explore a potential market edge on historical data. It does not necessarily take into account money management or positions sizing. In many cases, stops are not included. In short, it’s not meant to be traded with real money on a live system. It’s a learning tool that could be used as a genesis for a complete trading system.
A “trading system” is a trading model (a set of buy/sell/stop conditions) wrapped in a much larger set of criteria. First, a trading system has a set trading symbol or set of symbols to execute against. On the other hand, a trading model may be general enough to work on the S&P futures and the Euro currency futures. A trading system has been designed for a specific symbol or set of symbols. Furthermore, the trading system has been thoroughly back tested and forward tested. It has passed numerous tests to help determine its robustness and ability to survive into the future. The question of, has the system been over optimized on the historical data, has been reasonably demonstrated not to be a major issue. These test could have included Monty Carlo simulations, walk-forward optimizations, and testing on live market data. Finally, the trading system has been executed on the live market. It has placed live orders with real money. This is done to help discover potential coding bugs or logic errors that can only be detected on the live market. In short, the trading system has survived a whole host of testing to earn it the badge of being called a trading system.
As you can see, a trading system is a far cry from a trading model. In the future, I’m going to be using these terms with articles that I write. It will take me a while to modify my historical articles to reflect this distinction between a trading model and a trading system. So unitl that day happens, there may still be some confusion.
Why Are The Returns So Low?
I often receive emails asking me why my trading model has such a low annual rate of return. Some will ask, is it worth trading? The confusion has to do with believing the trading rules, as stated for the trading model, are all that determines rate of return. This is not true. At first glance this may seem not to make sense. If the buy/sell rules don’t determine return, what does? In short, position sizing.
What’s missing from the trading models is a position sizing model. A position sizing model can dramatically improve the return of the model. It’s completely possible to have two identical trading systems, taking the same trades on the same market yet, produce different returns.
Let’s start with a thought experiment. Let’s say we have two traders who are each given identical trading systems to execute on the SPDR S&P 500 ETF (SPY). Because both systems are identical copies of a trading system, both trading systems generate the same buy/sell signals, use identical stop loss and trailing stop parameters. These two traders are also going to start trading on the same day, with the same starting account size. In essence let’s pretend we have two traders that have identical trading circumstances. At the end of the trading period, we would expect them to have the same account balance, right? But at the end of the trading period, reality has made fools of us. One trader generated nearly twice as much net profit. How can this be? Reviewing the trading results for both traders you can see both traders took the same signals, yet both have different account balances. What’s going on? The secret is one trader used a simple mathematical formula to determine how many shares to buy. This formula is a position sizing model.
So there you have it, a position sizing model is a simple formula that tells you (or the system) how many shares/contracts to buy. This is a far cry from our trading model which simply traded a fixed lot. The position sizing model also reinvests the profits in order to buy more share/contracts thus, compounding the returns. The topic of position sizing is beyond this article, but you can start to learn a bit more about it by reading the following article:
A Position Sizing Model Example
Let’s add a very simply position sizing model to the overnight S&P system. Let’s say for every $10,000 in trading capital we trade a single contact. Recall we started out with a $25,000 trading account and we’re trading the E-mini S&P. Thus, we would trade 2 contracts until our total account value rises to $30,000 or above. If our equity would ever fall below $20,000 we would trade one contact. The following table demonstrates our position size based upon our account balance.
Account Size | Contracts To Trade |
|---|---|
$25,000-$29,999 | 2 |
$30,000,-$39,999 | 3 |
$40,000,-$49,999 | 4 |
$50,000,-$59,999 | 5 |
$60,000,-$69,999 | 6 |
$70,000,-$79,999 | 7 |
Now, this is a very simple position sizing model example and I just picked it off the top of my head, but it should serve to help make my point. Now what will this due for our annual returns? Let’s find out by reviewing the results table below.
S&P Overnight Model Results
Baseline | Position Size (1Cntrct?$10k) | |
|---|---|---|
Net Profit | $10,590 | $28,902 |
Profit Factor | 1.95 | 1.87 |
Avg.Trade Net Profit | $64.41 | $173.07 |
Annual Rate of Return | 2.50% | 5.61% |
The returns are higher because we are reinvesting our profits and thus, buying more contracts as the system climbs in net profit. Even if the baseline started off trading two contracts (instead of one), it would only generate about $21,180 of net profit. The position sizing model generated more profit with $28,902. Keep in mind, this is just one example and it might not even be a very good example. Yet, you can see how position sizing can change the results when compared to a trading model which simply uses a fixed lot. In some cases, you will find a dramatic difference when adding a position sizing model.
There are many types of position sizing models that you could test for any given trading system. Optimizing your position sizing model can really change the dynamic of your system. A good place to start learning more is by doing some research online. I would also recommend the following books:
- Trade Your Way To Financial Freedom
- Super Trader
Of course another way to increase your returns is to start with a greater amount of trading capital. Let’s use the same position sizing model setup as above except we will start with a $50,000 trading account instead of our $25,000 with the baseline model. The results are below.
S&P Overnight Model Results
Baseline | Position Size (1Cntrct?$10k) | $50 K Account With Pos Size | |
|---|---|---|---|
Net Profit | $10,590 | $28,902 | $80,445 |
Profit Factor | 1.95 | 1.87 | 1.99 |
Avg.Trade Net Profit | $64.41 | $173.07 | $481.71 |
Annual Rate of Return | 2.50% | 5.61% | 7.05% |
You would think by doubling your account size you would double your net profit. Yet, you don’t! You more than double it. Looking at these results I can’t help but think of the question, what’s the fastest way to make a million dollars in trading? Answer: Start with a $10 million dollar trading account!
Conclusion
When building a trading system it’s important to start with a trading model with no position sizing. Why? During the early stages of developing a model you want to strictly measure the performance of that model without interference of position sizing. You don’t want to use a position sizing model to make a poor system look acceptable! It’s only after a trading model has passed all the rigors of testing would I think of adding a position sizing model. It is the final step in the development process.