The last 3 posts on Magnitude/Direction/Timing are really the cornerstones of building an edge. Beyond having an idea of which direction price moves, in how much magnitude, and at what time, the Key Action of trading is fundamentally entry and exit. I’ve often struggled not with paralysis by analysis per say, but that the analysis doesn’t yield a single answer. It often says “here’s what is likely” but breaking that down in a definitive Probability has always been elusive. I’ve found it almost akin to breaking down physical materials into their particles to anticipate their interactions, but at a certain point the ability to grasp and use that knowledge practically is gone. It could be, as with science, that my tools are not strong enough, but until I know a) what specific problem the tool is supposed to solve and b) how to build the tool, it is not of much use.
So what have I accomplished in the past year? Actually a decent amount.
Direction
What I initially discovered when I built my first wave model was a rough “continuation probability” given the a swings current trajectory.When we track waves that begin with Trend (i.e the traditional ABC move), the probability that the next Trend will be in the same direction as the first Trend is about 60%. Trends and Swings Springs make up a majority of wave formations. This is all review. What is new is that these probabilities are quite robust – When building waves in a variety of ways or looking at the 15m Time Frame, the probabilities are very similar. That’s not to say that all wave models will look like this, but it gives me some peace to know that the model I built is both robust and not special. This market structure, as far as I’m concerned, is fundamental.
Another area that I’ve found to be consistent across models is the retracement percentages by wave type. A retracement of at least 30% occurs in Spring waves close to 90-95% of the time and don’t go past 80% more than 75% of the time. The problem that needs to be solved with spring waves is that I can’t tell the difference between Spring waves and Expand waves until the Expand wave breaks the High/Low. This led to an investigation into the counter moves that occur within these spring/expand waves, or the wave within the wave.
Somewhat surprisingly, the minimum number of pips in the counter swing here is quite sizeable; in most cases at least 30 pips (80%). Finding a way to target these seems like a good way to “test” entries and move then to BE, or else verify that if the 30 pip counter swing hasn’t occurred yet, there is still time to add to the position.
Magnitude
Most of my research into Magnitude this year went into a concept I call the bar wave. Initially this came from an idea (code wise) that would save me time in backtesting given that I made entry/exit decisions based on a single bar’s range. For example on a long trade entering on a bar with a 12 pip range I would have TP at 24 pips and SL at 12 pips. I mapped out, for every bar, the next 1 day/1 week movement given the following rules:
1) Wait for price to break out either above or below (this is the initial direction)
2) Once the initial direction is decided (ex up), wait for price to break below the lowest low (aka the initial bar’s low)
3) Now wait for price to break higher than the current high
4) Repeat breaking highs and lows until end of period, and summarize price movement in terms on price movement/initial bar pip range
This in same ways resembles very early wave models I was building, but it has some interesting properties that I missed/wasn’t aware of when I built the early models. From a wave perspective, these movements are always expanding, but I was curious to know by hour much, and if they differed by how sequential the break occurred. Perhaps the first break or two are modest (1-3 bar ranges) and the later ones were more explosive after getting some steam. There is quite a bit of noise because the first couple bars can break back to back, causing the third break to occur very early, whereas bars that start off in one direction can maintain a lower count. The results I found were not super unexpected, but did bring up an idea worth thinking about. First the results by hour:
Notes: the very left is showing the median break for all hours, or baseline. So the median move in range bars on the first break is just 1 bar, the second being 2, then it jumps to 4.5. What this says to me is that the third break tends to be the strongest. With this in mind, the “starting” hours that are best for this are hours 5-8, with median third breaks closer to 6 range bars instead of 4.5. Not the most intuitive or easiest to understand, but the idea worth thinking about for me is this: When it comes to static indicators that refresh daily like pivot points or daily range, they always start on hour 0, but perhaps in some cases the edge is greater when starting at a different hour.
Not a ton of headway in this department, but a couple of extra tidbits I’ll keep in my mind.
Timing
Last but not least, I made a couple of discoveries worth sharing in the timing department. This mostly comes about because while I expect market patterns to be reflexive across time frames, there some things that are worth digging into that only work in one time scale. For example, the increase average momentum during London and NY sessions is undeniable, as well as the day of week in some cases (which I posted about in the original Timing post. What I studied this year is in regards to the intraday formation.
There are 3 parts to a day’s movement:
1) The initial/head move, or the movement (generally) from open to the losing wick
2. The main/major move, or the move that contains the high/low of the day
3. The tail move, or the movement (generally) from the winners wick to close
Some questions I wanted answered: How many hours does each leg usually last for?
Is there a certain hour of the day that the entry move/main move is usually over? How many bars usually make up the main move?
Key takeaways:
# of bars contained in the major move: The number of bars contained within the main move is a fairly flat bell curve, which isn’t the most helpful, but makes sense in a model where the main move occurs as an ABC move that (sometimes) needs a couple of hours to retrace before finishing the move.
Length of Initial move: 70% of the time, the initial move doesn’t last longer than 11 hours. This means that hour 12 is a part of the main move 70% of the time, which can be helpful, but not always. Given that it is currently hour 12, the last bar in the chart, and given that we assume the current bar is part of the major move, there are two possibilities.
The first assumes the initial move has just finished, and the bar we are looking at now is at the very beginning of the major move
The second assumes the initial move has already finished, and the bar we are looking at now is just in the middle of the major move which should continue for a couple more hours before retracing slightly to end the day.
And finally what actually occurred:
Oh yea, there’s also that extra 30% where the bar is still apart of the initial move. Darn.
Hour of the day Tail starts:
This one is helpful for the Rat Zone trading, where the tail end doesn’t start until at least hour 15 80% of the time. By hour 18, the tail has started a bit over half the time. What this means is that when it comes to scalping the wick of the day, or to trade assuming the day won’t make any additional highs/lows, I should wait at least until hour 17. Hours 15 and 16 are good continuation scalps.
2021 goals:
My goal for 2021 is to start trading more. I think trading my models direction model with tighter stops is a good start. It IS discretionary currently – I have not successfully found the perfect (aka automated) way to enter and exit profitably. Historically I’ve taken this to mean that what I haven’t isn’t enough to be profitable. It’s felt for a while that I had enough information, enough edges to tip the scales in my favor, but I’ve always stopped short of attempting live trading because I really wanted to do it with a rock hard statistical edge in both entry and exit. I think though that I can make some additional gains by doing some live trading to help show the areas I want to improve on next. I’m going back to the basic trading styles with the knowledge that I have now:
- Weekly swing trading (mainly on 1hr frame) across multiple pairs to trade with the trend; Except now I can definitively define what a trend is, where price is likely to stop, where price is likely to continue to, and being more aggressive moving stops to break even.
- Scalping the NY session (now that I can actually be awake for it). Trade in continuation where it makes sense, scalping the daily wick. Keep the weekly flows in mind and try to capture areas where price is best suited to not return (daily/weekly opens)
LGtrades 