Sunday, September 06, 2009

An Intraday Look at the Transition Trading System

It struck me after writing the recent post regarding a trading system based on transition patterns in the stock market that an intraday version of the same system would trade false breakouts by fading short-term strength (weakness) that showed divergences relative to an earlier momentum high (low). A common setup would be a stronger (weaker) open that moves above (below) the previous day's high (low), only to reverse and trade back to the prior day's pivot level.

Let's take a look at the numbers that make this a promising basis for a system-assisted discretionary trade. All figures go back to the start of 2000:

* About 51% of all trading days trade above their previous day's high; about 27% of all trading days close above their prior day's high;

* About 48% of all trading days trade below their prior day's low; about 25% of all trading days close below their previous day's low;

* About 88% of all trading days trade either above their previous day's high or below their previous day's low; about 52% of all trading days close either above their previous day's high or below their previous day's low;

* That means that 48% of all trading days will close within yesterday's price range;

For the intraday transition setup, you would look to yesterday's high (low) as a potential momentum high (low) and today's upside (downside) break of yesterday's high (low) as a potential price high (low) to fade *if* today's high (low) is accompanied by non-confirmations from three or more sectors. This includes the eight S&P 500 sectors that I typically follow each week, as well as the NASDAQ and Russell 2000 indexes, which I'm also treating as sectors.

The trade is to fade today's break of yesterday's high (low), with yesterday's pivot price as an initial target. In Market Profile terms, the trade is betting that the market rejects value above (below) yesterday's high (low) and returns to the value area of the prior trading day.

I'm further filtering the pattern by only taking a signal if the move above (below) yesterday's high (low) occurs during the first hour of trade. This is because I want to benefit from rejection of value above (below) the prior day's high (low) when institutions are most active in the marketplace (i.e., I want to benefit from large traders being trapped long or short and having to cover their positions).

Ideal entry would be on a rise (drop) in NYSE TICK that cannot make new price highs (lows), confirming that a high (low) has been put in place. Any move to new price highs (lows) for the day would stop out the position. Beyond the initial price target at the prior day's pivot, a piece of the position could be kept in place for a further move to S1 (R1), depending on the emerging strength of the rejection of new highs (lows).

As with the swing version of the Transition System, I'll be firming up these rules over time. I'll also be adding a Momentum System, which I will outline in a post tomorrow. Once these are built out, I will post all signals to the blog/Twitter. I am also dedicating a separate $100K test account to the discretionary trading of these systems, where the systems give the basic setup parameters and I use discretion to execute, size, and manage the positions.

More on the blending of discretionary and system trading to come; in the interim, check out Henry Carstens' insightful take on this topic.
.

4 comments:

RAW said...

(light bulb comes on)

Gro said...

How about using the overnight high/low ? From what I've seen, many ranged days generally can't break these limits and return quickly inside the range (my target is usually the VWAP).
I'm sure you have already thought about it, just giving my 0.02$ here :)

Joslin Lolo said...

Hi Dr.Brett! Great post ...yet another thats pulls together a lot of different concepts...keep it up! I assume an additional filter would be if early trade satisfies the criteria for a range day vs trend day.

Joslin Lolo said...

hi again, more numbers that make this work is when we open between R1 and S1, we will hit the previous day's pivot level 76% of the time