Monthly Archives: February 2015

Time length from PTZ to TZ

A simple statistic that I never considered doing until just earlier today when I was staring at the chart thinking about my next move.

Capture

The time difference between the first PTZ and the real TZ. It does confirm something that I noticed, which involves when you can start to anticipate a true TZ occurring. I was specifically looking at the 2 day range (48), and 80% seems to be about what I was guesstimating based on the charts. I may have to try custom time frames after all…

Advertisements

Another post on filters, stacking indicators

I keep talking about it because I like reminding myself about what’s important and what my focus is.

Here’s a snapshot of E/U 30m from today. I took a long at a pretty good spot last night and made a winner.

fromto

However, looking at this chart, I really see no reason to long. In fact shorts look quite good, due to the logic of trend following principles and probabilities. I took the trade due to some stats and figures that I had on the H1 frame, where the long looks more appealing. The difference between having the “winning” information and not is simply the difference between filters. If I gathered information about 30m stats and wave patterns I think I would end up with slight bias towards a short here. But, the important take-away here is that the cause of failure in the 30m is due to a successful follow through of the 1hr, or some model that is not known to someone purely looking at this 30m chart. 

If a patterns success rate is 50%, we say there is no edge. However, if 20% of those failures are explained by another indicator, then we have a pretty good edge actually. This is the premise of why new traders stack indicators like stoch, macd, ma, etc in hopes that they can combine them to make a winning system. What I do now is really the same thing, except the success lies in creating indicators that “talk” to each other well, in a way that it can be shown.

 

Hmm….

hmmm

Pivot work, benefits to trade tracking

Haven’t been posting much because I haven’t had as much time in a week (and likely won’t unless something changes) so I want to make my posts longer! I hope that compensates a little.

A lot of times, I’m looking for a consistent pattern or some number to pop out to me. It’s generally best to do things in a general way if possible, because it leaves more room for improvements and alterations. I did this a lot with the previous work, but recently I tried to something similar to a trade simulator and learned something new.

big

In this snip here, I have actual data, followed by the pivot’s current direction, the current bars PTZ (aka, this bar is the highest bar of the past x bars), and some other directional filters. On the right side is my attempt at custom TFs. I created a “signal” that occurs when circumstance “x” is met. When this is met, I wanted to track the past 20 days. BUT this was assuming that the signal bar was the day’s close. In other words, I had to make use of all 24 possible “days”. Pain in the butt. Really. Then I ran something as close to a live simulation as I could. I took the signal, put a TP/SL on it, and tried to find my win rate. The result? +/- 5% from break even. No good.

After checking, rechecking, and digging through the code a LOT to make sure it was functioning correctly (and discovering a lot of bugs) I actually took the time to look at the results. The actual results.

streak

Both 1:3 and 1:1 results end up close to break even, but there’s a huge difference in what I came to see as a BIG deal. (if correct. Now I need to double check some more hehe).

Each histogram is counting the wins and losses as streaks. As an example when trading using the 1:3 RR, 39% of my losses occur in a streak of 1. That is, my trade history may look like:

L, W, W
or
L, W, L
or
L, W, W, W, etc

but the string of L, L, x.. occurs about 60% of the time. If I lose, I’m more likely than not to lose the very next trade. 60-40 is one thing, but here’s another. The 22% of streaks lasting longer than 10. This is MUCH more pronounced in 1:3 ratios than 1:1 ratios. Simply cutting trades after losing 6, 8, or 10 trades in a row and WAITING until a winner has passed seems to have some merit to it! As an added bonus, streaks of wins seem to happen as well.