Sunday, October 30, 2011

To scale, or not to scale out?

After trade reviews of the challenging narrow range day in ES on Friday posted here, I did a "what-if" type analysis on my trades with regards to scaling out.  Renato mentions in his blog post here the advantages of scaling out and an example how.  I also believe this is a great way to potentially reduce risk and increase profits, *IF* you're starting with the adequate number of contracts and scaling out properly. 

My current scaling out strategy
Instead of using units of 4 contracts as Renato recommends, my current trade management of scaling out usually involves starting with 2 contracts, exiting 1 car at +1.50, and letting the other run after bringing up the stop loss to +.25.

As you could imagine, nearly most of the time the 2nd car is stopped out at breakeven.  So from the pure mathematical perspective, unless you occasionally have a monster runner, the #'s just don't seem to add up for the scaling out method I use. 

The reason to dig deeper?  Because I had to climb so far
My gross profit on Friday was $50, and that's after going into aggressive scalp mode about 1.5 hours before the close and climbing myself out of a $600 loss on 12 trades.  That type of P&L curve doesn't sit well with me so it was time to do some further analysis. 

One point was clear, I was not patient enough and entered some non-strategy (i.e. rogue) trades that contributed to the drawdown.  That was was got me into trouble for most of the day.  But there was also something about my scaling out strategy that I wanted to explore.

The what-if scenario I modeled
"What if I did NOT scale out and instead, simply exited my entire position at +1.50 when the 1st target was hit?"

Here are the results
Instead of ending Friday with +1.00 gross, I would have had an incremental:

  • +13.75 if I used this method the entire day
  • +9.25 if I used this method only during the final 1.5 hours of aggressive scalping
Crazy.  The few bad non-strategy/rogue trades I took earlier in the day were NOT filtered out of my analysis, so the results could have been even better. 

During the final 1.5 hours, the range of the ES was about  6.50 points, so these results are difficult for me to believe (but remember, they're simply based on the actual executions I took from scaling out of the 1st contract). 

What's also interesting is that when the trades were profitable, nearly all took only 1-3 ticks of heat.  So there's a possibility that the stop loss could be tightened up -- however, this requires further research and a much bigger sample size.

But can it be replicated?
A part of me wonders whether the aggressive scalping performance was due to luck and therefore, can not be consistently replicated in the future.  On the other hand, I have enough evidence over the past 3 weeks to believe that this type of performance can be more than possible. 

The real challenge is whether I or anyone can have the mental stamina to maintain this rapid-fire method of trading both intraday, as well as over days/weeks/months/quarters/years.  It's difficult being a human trying to keep up with the HFT algo-bots!

The bottom line
Renato mentioned this in his chatroom last week, and the bottom line is that you really need to be able to enter a trade with the appropriate number of contracts to effectively scale out.  So if you can't enter with at least 4 contracts, then you need to chip away and build up your account so that you can.  And only then can you take advantage of the powerful benefits that scaling out can offer.

Therefore, my new strategy starting this week will be to exit out of my entire position at +1.50 or +2.00, based on market conditions.  No more scaling out, since I will not likely be using that method when I start trading live.


Greg said...

Seems like you need to evaluate this in the context of risk. If your stops are 3 points away, and your accuracy rate is say 55% (probably close, from what I remember of the last few posts), then limiting your potential gains to 2 points isn't going to work, unless something else changes...

Even having a single lot run for +10 points could make a big difference in your overall for thought.

Tomross58 said...

Hey Grove,,

I look forward to your posts and progress everyday and this was very interesting. Greg brought up a point that I didnt think of and it makes sense. I am still trading stocks (SPY mainly) but you are getting me more and more interested in trading ES, so my comment is somewhat directed to your last post. I see where Matt has just signed up with DS and wanted to ask you guys which platform you each use to trade the ES?,,,I currently use Lightspeed and cannot trade futures and I have a buddy that has Tradestation and really likes it. I cannot view my charts after 8:00 pm or on weekends so I use Freestockcharts to study but really would like to study the charts I use during the day. I understand Tradestation has the "Matrix" which my friend really likes and has some advantages not currently available to me on my platform. I know this is a little off topic but everyday I am getting more and more interested in what you are doing and may give it a go soon but need to learn another platform in the process. Thanks in advance.


Grove Under said...

Hi Greg,
Great point on incorporating the stop loss figure into the equation. Looks like I really skipped over that part unintentionally!

When tracking some of the more aggressive scalp type trades, the accuracy has ranged from the low 70% to low 80%, with a stop around -2.00 points. So the nearly even risk/reward tends to work out OK in that scenario.

However, I also went over the trades over the past month, and you're absolutely correct. Even though I got stopped out on many of my "runners" at b/e, a few trades really made it big and offset all those b/e trades.

In the end, based on a quick and dirty (emphasis on dirty) analysis, the bottom line P&L wasn't too far apart from one another based on whether I scaled out or went all in/all out. But in reality, I might be trying to compare apples and oranges with the data I have.

One thing is for sure, I still need to do a lot more homework to select what method works best for my personality.

Thanks for the great observation.

Grove Under said...

Hi Tom,
I'm also using TradeStation, but Renato and most of the others in the room are using OEC, along with charts from Sierra

I know about learning new platforms, I'm also an old dog that's not getting as swift of learning new things, so I'm sticking with TS for now. I like TS's matrix, too. But the other futures platforms seem to have it, and in many cases, might be better.

One drawback on TS is that it's very difficult to set a MIT (market if touched) order. I've found that those orders would be very valuable due to the precision of the DS system.

But definitely let me know if you have any other questions!

On another note, I do NOT want to sound like a promoter, but if you're interested at all with Renato's system, it looks like he's going to raise the price up after the next 20 subs. For me, I've found the $345 to be a great risk reward trade.

Tomross58 said...

Hey Grove,,

Do you find that Tradestation is fine for trading DS ?,,,Like you I am an odler dog and have started to at least become somewhat familiar with Tradestation and have a friend that can make the "learning" curve a bit smoother. I looked at openecry and would rather have my charts from Tradestation if they have what I need?,,Also a big THANKYOU for letting me know about the price increase. I am in the process of putting things together so I look forward to shareing my trades also. I will revisit some of your posts on DS but I believe you are most satisfied,,,is that correct?,,,,thanks again


Grove Under said...

Hi Tom:
Yes, I'm a very satisfied customer. And the feedback that I have received from several others in the chatroom is that Renato's service is one of the best services they have joined.

This is my first chatroom, so I can not comment how it compares to others. But based on what I hear, I feel very fortunate.

I do know that Renato is very diligent, helpful, and a great teacher with an easygoing style.
But most importantly, his system has produced consistent profits (assuming you follow it!) during the nearly 4 weeks I've been there.

Some of the methods he uses are similar what I've seen work nearly 20 years ago, so I have no reason to believe it will change 20 years from now.

If you do end up deciding to join, then let me know and I can definitely share with you the parameters I use for TradeStation to speed up your setup. There are some minor differences vs OEC or Sierra, but I would say it's not material to the use of the system.

Hope this helps, but let me know if you have any other questions.