Tuesday, February 23, 2010

Pullback Revelations 101

This might end up being a pseudo novel so I recommend only to continue on if you are really bored or under the influence of something mind altering. That being said, I think there might be something of worth in the jumble that follows.

That massive losing day last week slapped some sense into me (and rightly so). What I was doing was fighting the market like usual. For the past year my entire method had been created around support and resistance - sounds good right? Well actually it was created around trying to catch reversals at support and resistance, double tops and bottoms, etc. All that is awesome save for one very important fact - each time I was fighting market momentum. Price may move into an area of prior S/R but unless it is dead set on reversing on a dime you stand to lose a lot of money very often. Typically there might be some reaction, maybe enough to get to par but rarely did you get a clean V reversal at those levels.

Whats worse was I was using fixed targets for profit so my winners were cut short every time and my losers weren't fixed and never really made up for anything. I was hitting 71% (Win+Par/Loss) on a good day. My account suffered to say the least.

I had been doing a lot of reading on different sites about profit factors and how important they are (much more so than win percentages) and so I did some simple calculations on my results of my S/R method.

Win/Loss (No Pars) %: 56%
Avg Win: $48.10
Loss % (1-56%): 43%
Avg Loss: $-81.37
Profit Factor: .76

Drrr.... why can't I seem to turn a profit? We are now entering Dumbassville, population me. My 71% rate was inflated by useless commission decaying pars. I was an idiot and I just never got around to realizing the fallacy of "But I have a 71% win rate INCLUDING PARS). Win rates mean nothing, especially including your pars. Profit factors are what keep you consistently profitable or not.

If you curious about trade expectancy and profit factors check out this excellent article by Richard here.

So I needed to do a 180 and change everything.

The issues I had to correct were simple and obvious:
  • Achieve a method with a Profit Factor of 1.5+
  • Increase my ACTUAL Win/Loss ratio.
  • Increase my Avg. Wins, while limiting my losses.
  • Limit and have a FIXED risk.
  • Uncap winners.
  • Focus on reducing pars.
  • Getting paid for my par efforts to cover commissions.
First thing I did was switch to range charts. I had traded minute, volume, and tick charts my entire career and thought they were the cats meow. The problem being they still don't filter market noise well. I thought ticks did this great in comparison to minute bars (and they do in my opinion) but in comparison to range bars, well there is no comparison. They truly make the "picture" more clear for me by showing me the IMPORTANT movement of the market.

But the other big thing with range charts is this: They are a constant size. Obvious right? Well that obvious fact allows me to do a number of things. First off, I can have a fixed stop on every trade AND know what it is prior to me getting into the trade. Second, I know what my entry price is going to be before the bar ends, thus I can place my orders well ahead of the fact and get in when the market is making its move without scrambling for a fill only to be passed up.

Finally, it allows me to have a fixed target to get to par and a fixed loss each trade. I have been measuring off the size of my entry bar for par and target levels forever (and I like this method), problem was depending on the size of the entry bar on a tick or minute chart your par might be 4 ticks away or 14. It wasn't consistent. With range bars it is the same each and every time.

So now that I had part of the equation figured out I just had to come up with a simple entry method. I knew what I needed to do. One word - Pullbacks. I'd been slapped by trending markets more often than I cared to talk about so I decided to get on board rather than continue to hand over my stops to the trend traders. But the problem with pullbacks has ALWAYS BEEN this: "The trend is you're friend until the end". Well shit... if it ends that means i'm guaranteed to take a loss at the end of each trend - or does it?

In years passed I had recognized this inherent flaw in the ideology of trending pullback entries but I worked to avoid this problem by trying to utilize indicators and divergences in the markets. Problem was they were just as hit or miss as anything else. Turns out the answer had been in front of me all along - I just didn't SEE it until now.

If the issues with pullbacks is catching the "end of trend" where is the most obvious place to look for the highest probability of avoiding this losing trade? Why at the BEGINNING of the trend of course. By attempting to trade ONLY the first pullback from the start of a new L, H, HL, HH or H, L, LH, LL sequence you could avoid the situation almost entirely. Not only that but I found that typically these "first pullbacks" were the strongest and had the tendency to run the farthest. I know its rudimentary and blatantly obvious to all of you, but I hadn't had this "aha" moment until right now.

The following should illustrate what i'm talking about quite well:

The example is from a typical day in the currencies, not particularly trending at any one time... but the idea remains the consistent. The first pullback trades are the biggest runners, the most likely to run far enough to get to par, and the most likely to work in favor. Is it perfect? Nope. But its a good logical analysis of price action that seems to have a practical application.

But in the back of my mind I continued to think "but doesn't all of that extra profit from the smaller continuation pullbacks make up for the one "end of trend loser"? So we tested and tested and tested and came up with some pretty interesting results. **AND YES I RECOGNIZE THAT THESE RESULTS ARE NOT STATISTICALLY SIGNIFICANT**

ALL TRADE Results (5 Days, 9AM-2PMCST):
Gross Ticks: 106
Gross P/L: $1,088.75
Commissions (@5.45/RT): $288.90
Net P/L: $859.85
Win/Loss: 74%
Win+Par/Loss: 86%
Avg $/Trade: 20.47
Total Trades: 42
Profit Factor: 5.6

FIRST PULLBACK Results (5 Days, 9AM-2PMCST):
Gross Ticks: 43
Gross P/L: $513.75
Commissions (@5.45/RT): $81.75
Net P/L: $432.00
Win/Loss: 86%
Win+Par/Loss: 93%
Avg $/Trade: $28.80
Total Trades: 15
Profit Factor: 5.1

First things first... BOTH MADE MONEY. That in and of itself is a pretty big objective. And the second biggest thing to note is the profit factors on both. WELL ABOVE 1.5. This was in no small way due to the fact that I got BIG winners and had a small fixed risk. Actually you could make more profit by trading EVERY pullback but look at what you give up...

The efficiency of your system takes a huge dive. Taking only the first pullback requires nearly 3X's less trades, and increases the efficiency of each trade up 41%! Basically the added profit was due to a slight increase in winners but those winners were smaller overall compared to the 1st pullback trades, and you had more losses (6 vs. 1) and a LOT more pars which confirms our theory on first pullbacks. They simply run farther and are more consistent.

Frankly, i'd rather take less trades and have more consistent results even if it means less net profit at the end of the week. 15 trades over 5 days is still 3 a day and that's right about where I want to be.

So that's the story, that's what i'm doing, that's my revelation. The results this week speak for themselves:

Results Mon (2/22):
4 Trades, 1 Par, 3 Winners.
Gross P/L: $260.00/per.

Results Tues (2/23):
6 Trades, 3 Pars, 3 Winners.
Gross P/L: $402.50/per.

1 comment:

E-Mini Player said...

You need to listen to me more often!



Good luck dude! Wish you the best, as always!