One day not so long ago, I could be found checking out the commission column on my backtested trades. I wanted to determine if I was paying more for commissions than I had during that backtested period.
Did I suspect that my trading platform has sneakily overcharged me for my trades? No, I know one of my faults as a trader is that I tend to overtrade. That tendency isn't present when I'm backtesting. Checking to see if I'm paying a lot more in commissions for my live trades than showed up during the backtesting period would give me a quick heads-up that I might be overtrading.
What would lead to more overtrading during live trading than during the backtested period? Let's look at one backtested trade from 2011 as an example.
Backtested Trade on January 27, 2011, 50 Days to Expiration:
The trade guidelines allow for the price to move outside the upside breakeven when the trade still has more than thirty days to expiration. When I was backtesting, I wouldn't have been tempted to adjust this trade. Delta looks well controlled. I would have followed the guidelines, looking at the trade at the end of each day and making a decision.
In live trading, matters would have been much different. This was the second day in a row that the RUT's price had been outside the upside expiration breakeven, and I would have been well aware of that fact through two long days of trading. In addition, I don't have a clue when backtesting what important economic reports might have been due the next day before the open unless I have some special reason to remember one particular date. What was happening in Asia or Europe? Don't know. I do when I'm trading live. It would have been easy to convince myself after those long two days that it was time to adjust if I knew there was a potentially market-moving report the next day.
What would it have hurt if I'd made a small adjustment, moving some of the flies or adding some credit spreads? Perhaps such adjustments wouldn't have caused any harm, but they would have cost extra commissions. Moreover, if I'd chosen to move some of those flies, I would have been selling the original flies for less than I paid for them, so I would have been locking in a realized loss for that transaction. Fortunately, the potential profit available to the butterfly trade I'd set up was large enough that it probably could have absorbed the extra costs. Remember, however, that the extra costs are not only monetary: they would have included a probable longer time in the trade before profit built enough to pay for those extra costs and approach my profit target.
Same Trade, Next Day:
The RUT moved back under the tent the next day. The idea of setting up the trade this way is to allow a one- or two-day excursion outside the tent if deltas are well controlled and there are more than 30 days to expiration before jumping into a big adjustment. Then, if the price does drop back, no harm has been done.
Of course, what did I prove by showing you one instance in which the RUT turned around and it was a good thing that I followed my guidelines in the backtest? I could also have located instances in which it would have been a very good thing that if I had ignored my guidelines and moved some or all of those flies that previous day. The point is that I didn't know anything other than where price was and my guidelines. Because I was backtesting, no fears about what would happen the next day intruded on my decision. No overtrading occurred. The tests showed that, despite some periods when I would experience several losing months in a row, the trade made money.
I could have confidence that the trade would work and put it on month after month. If I were overtrading, however, there was no guarantee that the results would be anything like what I'd found when I was backtesting the trade. Certainly during the course of the backtested trades, I found many times when I bought debit spreads because the delta was too negative or sold them because the delta was too positive and then had to reverse the trade the next day. I try to keep the deltas at +/- 5 deltas per butterfly contract. Then perhaps I might have to reverse the reversal the next day. That's part of what I tested.
However, I did not test keeping deltas at +/- 2 deltas per contract because I was getting too antsy when the deltas were any distance at all away from zero or when I knew there was some economic release the next day. Suddenly switching away from my plan that way would have meant I was adding more commissions and also buying back and selling debit spreads or moving flies too often, locking in more realized losses than I had done during the backtesting period. My live results couldn't be expected to track my backtested ones in that case.
Let's be honest. I don't always follow my guidelines exactly. I try. My general guideline is to adjust once, at the end of the day, but if I see the likelihood that there's going to be a huge move, and if the advance/decline data shows that such a move is favored, I might make a small protective adjustment earlier in the day. That has to be counted as overtrading, since it's not a strict following of the backtested guidelines.
If it's a day when I know I'm going to have to move flies--perhaps it's less than 30 days to expiration, for example, and there's little chance that the underlying will change direction enough to move back under the tent--I'm not going to wait until the very end of the day to move flies. I'm going to look for a quiet time to move those flies, enough ahead of the close that I can avoid that end-of-day scramble and the way that prices go all wonky at that time.
Let's also acknowledge that backtesting doesn't always mimic live trading even if you can religiously follow all guidelines. You might follow them, but that doesn't mean the markets agree with your choices. On 2/24/14, I tried for 21 minutes to buy two flies to replace flies I had just sold, in an effort to move those two flies higher. The RUT was just about to tumble lower, and I was having trouble getting anyone to take my offer. I had to keep upping my offer. When backtesting, I would have just put in the order, maybe built in some slippage, and then "committed" the trade, in the nomenclature of ONE, the testing platform illustrated earlier. In real life, I chased that order as the mid went up, and the fill was a whopping $0.60 more than my original offer. I'm glad it was just two of my flies and not the whole kit and caboodle.
Do these realities of live trading mean that we should not bother to backtest? No, certainly not. That's why that late February day, I was checking my typical commission costs when backtesting against what I had been paying the last several months. That check gave me a needed and quick reference as to whether I was overtrading and venturing too far away from my trading guidelines. As it turned out, those commission costs did not appear excessive when compared to the backtests I had run, so I relaxed and just did what needed to be done with the trade. It was going to turn out that the trade I was in that late February day was not going to be a successful trade, as you'll see in next week's article, but those backtests still give me confidence to keep trading, even when it's not so much fun to look at the PnL each day, or even when I have a losing trade . . . or two losing trades in a row.
The moral of the story? Backtest. If you don't have backtesting capabilities, forward test. Put your simulated trade up and run lots of "what if" scenarios, running implied volatilities up and down and underlying price every which way and moving dates forward. Follow simulated trades through for several months. Trade very small, perhaps in IWM for an intended RUT trade or SPY for an intended SPX trade, although commission costs will take a bigger burden on those ETFs. You can't replicate all the hurdles of live trading with backtests or forward tests and can't measure the same emotional import with a very small trade when compared to a much larger one, but it all helps! It's about building confidence in your trade so you can follow the guidelines and enter your trade on schedule, even when you've just experienced a losing trade.