## Friday, July 29, 2005

### Part II: Scale Fish not Trades

In part I, I mentioned how scaling into trades ends up increasing the complexity of your system while reducing the edge and thus your profitability. Now, on to scaling out of trades.

Scaling out of trades is pure magic. Yes, magic. The more I think about scaling out of trades...the more I realize how powerful this technique is to your system. Think of it this way. Scaling out of your trades gives you the chance to create trading systems that are 100% profitable. How in the world can you beat that? :)

Let's get to work. Here's the 1st example we'll work with:
Full Deck Method
Starting Equity: $35,000.00 Symbol: XYZ Shares: 1,000 Entry Price:$35.00
Exit Price: $52.50 Gain/Loss%: 50.00% Gain/Loss$: $17,500.00 Results: New Adjusted Equity:$52,500.00
Net Profit%: 50.00%
Net Profit$:$17,500.00
Win Ratio: 100.00%

Not too shabby for an old dog like yourself, eh? Keeping your deck fully loaded throughout the trade thus capitalizing on the big 50% move.

Now, what happens when you attempt to scale out instead? Let's keep it simple, we'll break the scales into quarters and scale out at the 6%, 18%, 32% profit mark, and then let the remaining quarter ride. So, here's what the trade looks like:
Scale Out Method
Starting Equity: $35,000.00 Symbol: XYZ Shares: 250 Entry Price:$35.00
Exit Price: $37.10 Gain/Loss%: 6.00% Gain/Loss$: $525.00 Symbol: XYZ Shares: 250 Entry Price:$35.00
Exit Price: $41.30 Gain/Loss%: 18.00% Gain/Loss$: $1,575.00 Symbol: XYZ Shares: 250 Entry Price:$35.00
Exit Price: $46.20 Gain/Loss%: 32.00% Gain/Loss$: $2,800.00 Symbol: XYZ Shares: 250 Entry Price:$35.00
Exit Price: $52.50 Gain/Loss%: 50.00% Gain/Loss$: $4,375.00 Results: New Adjusted Equity:$44,275.00
Net Profit%: 26.50%
Net Profit$:$9,275.00
Win Ratio: 100.00%

Wow! That's rough. The Scaling Out Method cut our profits almost in half when compared to the Full Deck Method. What kind of magic dragon you've been puffing Taylor? But, there's another layer hidden behind these numbers. This hidden layer is a roller coaster filled with ups, downs, and a whole lotta nothins that earning 50% on your investment requires.

What if I told you that in order to achieve that 50% return you had to witness a run-up to $46.20 (32% profit) then a fall back down to$37.10 (6% profit)? That's a 20% drawdown in the trade or $9,100.00 in paper profits just wiped away. Now, you're probably thinking...no way! I have trail stops buddy! I'd never trail my stops 20% or more from the high/close/whatever. If that's true then you'd never achieve that 50% profit for this particular trade example. And since trail stops are a whole nutha beast let's move on. So, back to the example. If you scaled just like we discussed before...instead of$9,100.00 paper profits vanishing along with your hopes for the trade...only $2,275.00 is lost (4th scale still open) while$4900 (3 scales sold) in actual profits are locked-in. This Scaling Out Method works to lock in your profits but more importantly reduce the drawdown in the trade.

Let's keep testing. What if in the example above we had another trade after the XYZ stock. This new trade was in the ZZZ stock and believe it or not lighting does strike twice in the same portfolio because this trade worked just like the XYZ stock. Hit every mark...but instead of climbing back from that 20% drawdown from the peak of 32% profit...it never came back? And instead you were forced to close out with the paltry 6% profit. What kind of details would that give us...

Full Deck Method
Starting Equity: $35,000.00 Symbol: XYZ Shares: 1,000 Entry Price:$35.00
Exit Price: $52.50 Gain/Loss%: 50.00% Gain/Loss$: $17,500.00 New Adjusted Equity:$52,500.00
Symbol: ZZZ
Shares: 1,000
Entry Price: $35.00 Exit Price:$37.10
Gain/Loss%: 6.00%
Gain/Loss$:$2,100.00

Results:
New Adjusted Equity: $54,600.00 Net Profit%: 56.00% Net Profit$: $19,600.00 Win Ratio: 100.00% Number of Trades: 2 Max Equity Drawdown: -19.69% Scale Out Method Starting Equity:$35,000.00
Symbol: XYZ
Shares: 250
Entry Price: $35.00 Exit Price:$37.10
Gain/Loss%: 6.00%
Gain/Loss$:$525.00

Symbol: XYZ
Shares: 250
Entry Price: $35.00 Exit Price:$41.30
Gain/Loss%: 18.00%
Gain/Loss$:$1,575.00

Symbol: XYZ
Shares: 250
Entry Price: $35.00 Exit Price:$46.20
Gain/Loss%: 32.00%
Gain/Loss$:$2,800.00

Symbol: XYZ
Shares: 250
Entry Price: $35.00 Exit Price:$52.50
Gain/Loss%: 50.00%
Gain/Loss$:$4,375.00

New Adjusted Equity: $44,275.00 Symbol: ZZZ Shares: 250 Entry Price:$35.00
Exit Price: $37.10 Gain/Loss%: 6.00% Gain/Loss$: $525.00 Symbol: ZZZ Shares: 250 Entry Price:$35.00
Exit Price: $41.30 Gain/Loss%: 18.00% Gain/Loss$: $1,575.00 Symbol: ZZZ Shares: 250 Entry Price:$35.00
Exit Price: $46.20 Gain/Loss%: 32.00% Gain/Loss$: $2,800.00 Symbol: ZZZ Shares: 250 Entry Price:$35.00
Exit Price: $37.10 Gain/Loss%: 6.00% Gain/Loss$: $525.00 Results: New Adjusted Equity:$49,700.00
Net Profit%: 42.00%
Net Profit$:$14,700.00
Win Ratio: 100.00%
Max Equity Drawdown: -5.33%

Now where getting somewhere. The Full Deck had 2 winning trades, $19,600 profit, and a -19.69% system drawdown. Scaling Out had 8 wins,$14,700 profit, and -5.33% system drawdown. So, who's the winner?

Well, the Full Deck Method still earned almost 10% more in profits than the Scaling Out Method. But, with 300% more in system drawdown I think it's fair to say the Scaling Out Method provides the better return for the amount risked. Throw in just one more trade that is a loser to boot and the whole ball games changes. Currently, the win ratio on both systems are 100%. But, with a loser added to the mix the Full Deck Method would drop down to a 67% win ratio. While the Scaling Out method at a worst case scenario would still achieve an 89% win ratio. And could be better if just one of those scales were made before the trade turned into a loss.

What do you think? A way to lock in your profits, free up your money quicker, let some profits ride, reduce your system drawdown, and increase your win ratio to boot. Not too shabby, eh? But, what I like even better is again...where else do you get the chance to create new trading systems with guaranteed profits?

Now, please don't get picky on me. I'm not counting slippage and commission costs into the mix. Feel free to do that when analyzing your own trades. And the 6%, 18%, 32%, and let it ride scale is just an example. By all means, test out different combination of scales in your own trades. Should be enough to keep you busy for a good long while. :)

Rants:
I've been a programmer for 10+ years now. This has involved working with many other programmers both old and new. Know what the most common fault I have witnessed over these 10 years? The failure to test ideas. The failure to test programs. We think it should work, so it must. Ha!

Just what do we have against testing? Why do we feel the need to pontificate on things that can easily be tested? It's a really strange phenomenon that even I'm guilty of from time to time. And I know better!

I think the key is this: Avoid trying to prove an idea with logic and speculation. Use logic and speculation to create ideas and use testing to prove whether those ideas are indeed valid. Maybe logic is not the correct term here. But, I hope you get my meaning.

Side Note:
If you've had a good day, week, month, or year...do you reward yourself for trading well? I haven't really done this before but I'm beginning to think this is a good idea. If all you do is roll the profits back into the business then where's the reward for all your efforts? Usually, my reward or gratification comes from the creation of new systems or learning of new tools/ideas/methods. And of course, writing this blog helps to communicate my thoughts to an audience that can understand and appreciate what I do. That's a reward in itself. But, maybe Rod Tidwell is seeping into my brain and a little part of me is shouting...SHOW ME THE MONEY! Ha ha.

Anyway, upon this theme...I'm thinking about getting a Kodak Digital Camera (EasyShare DX7440 4MP with 4x Optical Zoom for a little reward. I'm a photo-taking newbie so if anyone has any comments on this camera or any other good digital cameras in the $200 -$300 range, please let me know.

Anonymous said...

hey this is Greg from The Trader Log I was checking out your blog, great work! I am adding you to me bloglines today.

Mike Taylor said...

Thanks! Two things of interest I saw on your blog. The python COM work. Something I will begin to look into in order to get email alerts working with the Trade-Ideas software.

I also like your recent mention of The Mathematics of Technical Analysis by Sherry. That book is the next in my queue to read.

MT

Michael said...

Great post Michael! I'm about to put up a link to it on my blog...

Michael said...

oh, on the rewarding yoursef thing -- I've started to do that recently. It may be something as simple as going to my favorite restaurant for dinner after a good day. I think it's critically important to do stuff like that as well as taking some profits out of one's trading account and do something with it. Otherwise what's the point of making the money?

Damian said...

Isn't there a tax aspect of this that you need to take into account? I agee, in general, with what you're saying, but if the full deck trade can be held for over a year, then it has significant tax benefits in terms of long term cap gains verses short term cap gains. Buffett was obviously a big believer in this - what would the model look like, in terms of returns of both systems, if you took into account taxes?

funkysmell said...

interesting

Mike Taylor said...

And thanks for the dinner idea. I'll have to try that sometime soon.

You know, it seems so much effort is used to earn money...hard to turn that off and enjoy it. But, I promise...I'll learn. :)

MT

Anonymous said...

Re: scaling out I really think you have to analyze the benefits versus some trailing stop scheme, because both methods get you liquid and into another trade and you need to establish for yourself which way makes more money presuming you re-invest your stopped out cash once you get it.

Re: rewards: "You can't have your cake and eat it too". When I was a kid I never understood this saying; who'd want to have a cake? The POINT is to EAT the cake ... with profits it's a little different. If you're working with less than 250,000 my vote would go to putting your profits back into your system. You don't have enough cash to start sapping it out of your account on every good day. What's the point of shopping around for the lowest commissions only to blow that commission and more on a reward each day? Of course, at some point the idea is to make enough money to spend it, because like the cake ... what's the point of having the money if you can't spend it.

In summary, I guess I'd save the rewards for when the account is able to toss the money out on a regular basis. For many people reading your blog, they won't be at that point. They should use rewards sparingly at best - perhaps after an up year or an up quarter. I'm up over several years of trading but the account sure isn't generating profits at the rate of a dinner per success ... at least, not a _nice_ dinner :-)

Mike Taylor said...

Damian, I have not analyzed the tax aspect of scaling since the majority of my positions held are less than a year. But, for readers who hold a year or longer, this is something to consider along with commission costs and slippage when testing your own portfolio's history of trades.

MT

Mike Taylor said...

Anonymous, thanks for the camera review site. Will help in my research.

Regarding trailing stops. I briefly mentioned them in the post mainly because trailing stops are a topic in themselves. And, yes, you are right to consider them. But, not in comparing one versus the other. In my study, the type of exit I was using on the original system was a trailing stop. Scaling out actually improves upon trailing stops.

You see, for years I've always had a gripe with my long-term systems and their exits. Why? Because if I use a profit stop then there is a possibility the stock will continue moving up leaving me without a position.

Trail stops work in reverse. You had profits but due to an adverse move...now you don't have as much.

Time exits? The stock didn't behave like you thought so you get out. But, was there some point in the hold where the stock did okay? Just not okay enough to kick in one of the other exit types.

By combining trailing stops, time stops, and scaling out...I've found a happy medium.

Now, instead of scaling out due to profits, you can get real fancy and choose a volatility-based movement to scale out instead. This is something Tom Basso mentioned in one of the Market Wizards book. In fact, about 5 years ago I contacted Mr. Basso about his logic from the interview and asked if I got it right. He was nice enough to contact me back and confirm my logic was true and wished me the best. Kinda cool considering I was just a budding system trader at the time.

Thanks!

MT

Dan McCarthy said...

Hey Mike,
While I've had a ton of experience with data manipulation in R, up to this point I've been inputting the data manually into a root directory. This has obvious downsides. What do you think of the fBasics package and do you have any recommendations on how I can interactively pull data from the web into R in a more seamless fashion?

btw, as I'm sure you know, S+FinMetrics contains basically all you'll ever need when it comes to weekend/weekday conventions in whatever time zone you happen to be in (I believe fCalendar performs a similar function in R).
-Dan