This weekly trade review is part 4 of my progress on rebuilding my trading system to be more systematic and semi-automated.
See My Weekly Trade Review Part 1 and Part 2 and Part 3 for previous updates.
What have I been doing this week?
The focus this week, has been on comparing the sim entries for the auto triggers against my manual entries, taken at the time I hear the submit order announcement, which on the conservative entries, is at bar close.
What have I learnt
1) Manual placed trades did not get filled as much as the sim placed trades. Which was not a surprise, even with the Sim on realistic settings.
2) All of the trades that I did get filled on, the manual trades where always better entries by 1 to 2 ticks.
I have manually placing the trades at the bid for the longs and at the ask for the shorts, on the close of the bar.
Which I assumed what the auto triggers was doing, as that is the same settings as the back testing.
As we all know, assumption is the mother of all fuck ups.
Having decided to look at the actual setting at the end of the week, the auto triggers place a limit order at the last traded place.
So my testing on that side has been a bit of a waste of time, as I am comparing apples against oranges. But at least I now know why the auto triggers are placing the orders in slightly different place to where I was expecting.
I have changed this to the correct settings. It may be after the real life testing that I change the setting back to last price, but I need to see what the actual fill percentage before I deviate from what I tested.
The triggers are not that sensitive to low volume and high volume conditions. In the low volume, it doesn’t trigger always and in high volume it triggers a bit too often. At the moment, the settings are fixed, but further down the line, I can look at having this auto adjust based on low volume and high volume conditions.
But at the moment, I need to focus on the testing of the current setup.
The next step is to test it live.
I will commence this next week. One of the jobs this weekend, will be to write a risk profile and plan for the week.
Questions to be answered
What is my aim for the week?
I am looking at where it actually place orders and what the fills are like.
How practical is using the auto trigger?
What is the management like after entry?
What is the typical stop?
What is the typical target?
What is the R/R ?
What is the frequency of trades?
The last four needs to be compared against the back testing and the sim testing to see if I am getting similar sort of stats.
Of course, I will not get enough sample data from one week to confirm either way, but it should give me an idea if it is way off base.
Risk will only be one lot. I intend to reduce risk to a minimum whilst testing, to limit any damage from auto error or human error.
What do I if I get 2 different triggers in a row?
What if I get opposing trigger whilst I am in trade before my opposing area / target?
What do I do if I get opposing trigger just before a blind order at a strong area?
I would like to reduce as much of the grey areas I can think of before going live.
This is all based on the assumption that the platform will actually send orders to the live account without problems. That is the first hurdle.
My Semi Auto Trading – What it is and what it isn’t
I have had a few questions via social media and email on what I am doing. I thought it best to address this here, so that others can see what I am trying to achieve.
Let’s start with what it isn’t
It is not set and forget.
I cannot turn this on at the start of the session and just walk away.
You will not find me out sailing, anytime soon, whilst the auto trigger prints money for me.
It does not set any stops or targets, this still has to be done manually.
It does not decide what the context is, whether the market is trending, balancing / ranging, etc. This is still decided by me manually according to my procedure.
What it is
All it is, when certain criteria are met, it sends an order to buy or sell. This is based on what I look for in triggers, when I was discretionary placing trigger orders.
I decide when the auto trigger should be looking for an entry and whether that should be a buy or sell.
The market is trending up, and my expectations is for price to make a higher high and test the next opposing level higher on the chart.
Therefore I will want to buy a pullback. I will look for areas that if price did pullback I would want to take a long in. And depending on certain criteria, I will decide if the level should be taken blind, i.e. I will place a limit order a couple of ticks in front of that level. Or I do want to be long at that level, but only if I get a trigger there.
If I decide that I want to get long at this level on a trigger.
I will activate the buy trigger, and once price reaches the level, the auto trigger will send a long, if the trigger criteria is met. But price has to be at that level and the criteria has to be met for it to send an order.
I have only automated one part of the trading, the part that I found that produced the most inconsistency in my real life trading. The stalking of trigger trades.
All the rest is still down to me.