TRADE LIKE A GUERILLA with the ATEEZ trading strategy (Patreon)

The ATEEZ trading strategy

Hi there, and welcome to this post where I will present you one of my own trading strategies. It is the ATEEZ strategy and only Atinys know what this name means :-)

The strategy

The ATEEZ trading strategy is an acronym for the indicators I use this time. I intended this for trend trading and it consists of the following indicators:

• A of APO or absolute price oscillator, which is a technical indicator used by traders and analysts to measure the difference between two moving averages of a security’s price.
• T of the TSI or True Strength Index which is used in financial analysis to measure the strength of a trend over time.
• Two Exponential moving averages. I will use the EMA 21 for shorter trends and the EMA 100 as the baseline to detect bullish or bearish conditions.
• Z of the Z-score that are also used by traders and analysts, where they can gain insights into market trends and potential opportunities.

Trading rules

Now what about my trading rules. Considering all these indicators I will…

Go LONG when:

• The Absolute Price Oscillator is above it’s zero line then check if the
• The close price is above EMA 21 and
• the close price is above EMA 100 – This EMA 100 is a bull bear market divider by the way
• THe EMA21 should be above the EMA100
• The TSI should be above the signal line and above the 0 line
• And finally, the Z-score over 50 periods should be above 0

Exit the trade when the True Strenght Index goes belows it’s signal line.

The go SHORT signals are the opposite as the long signals here:

• The Absolute Price Oscillator is below it’s zero line
• The close price is below EMA 21 and
• The close price is also below the EMA 100
• Here the EMA21 should be below the EMA 100
• The TSI should be below the signal line and below the 0 line
• And finally, the Z-score over 50 periods should be below 0

And again exit when the TSI crosses its signal line above.

Jupyter notebook with the strategy files code

In the files section below this post I will add the html file of the jupyter notebook for this strategy. There you will find all the information you will need to recreate this strategy or make an adaptation of it.

As you will notice, this post is an extract of the text that comes originally from that Jupyter notebook file.

Trading rules

I will create rules for long and short trades here so that I can also find out if futures trading is profitable too.

Remember the trading rules here.

LONG rules:

• APO > 0
• Close > EMA21
• Close > EMA100
• EMA21 > EMA100
• TSI > TSI_signal & TSI > 0
• Zscore(50) > 0.5

Exit the trade: TSI < TSI signal

SHORT rules:

• APO < 0
• Close < EMA21
• Close < EMA100
• EMA21 < EMA100
• TSI < TSI_signal & TSI < 0
• Zscore(50) < 0.5

Exit the trade: TSI > TSI signal

Signals function

Exiting trades

In the notebook are also two dataframe columns that provide information that helps me get out of trades. As said, the exit signal is when:

• In LONG trades the TSI gets below the TSI_signal line
• In SHORT trades the TSI gets above the TSI_signal line

Backtest results for the ATEEZ strategy

After I have created the actual strategy file that can be used by the Freqtrade bot to trade with. I made it so that this file can also be used for optimization of the two EMA's and Zscore.

Optimizing more parameters will only lead to curve fitting and unrealistic future results. So I think this will suffice for now.

Now, After I have backtested this strategy against my usual spot and futures dataset and pairs and got the following results:

Spot backtest results

After I have backtested with the initial parameter settings for all the indicators I get the following results:

And right of the bat it already has supergreat results. All the ratios are at least above 1 and 50% of the pairs react really well to this strategy. Also the drawdown is quite lot for a trend trading strategy. The only dissapointing number is its winrate, but that could be improved though.

The profit plot also shows that this is a strategy that catches the bullish side of the market cycles here. The bot is quite active during the upward market movemenst and stays relatively still when there are more bearish conditions. It also catched some major pumps on some coins and that is pretty inevitable if you backtest over crypto.

So let's see how the strategy performs if it also makes use of bearish market conditions and go short on these.

Futures backtest results

The execution of the futures variant of this strategy proved to be somewhat of a problem. The 1 day timeframe could not be tested for example and I could not find out on which pair the error was happening because all dataframes were nicely created. So I did not bother to do a big investigation in this but simply backtested all the other timeframes to see if they could do any better than the already highly profitable 4 hour spot version.

And after backtesting the futures option I got the following results back:

De eerste twee regels moeten er vanaf gehaald worden. Verder klopt het wel.

Again the 4 hour timeframe seems to have the best papers here. I'm not sure if the 1 day timeframe could top this. Also it is pretty comparable with the spot results, although the win percentage is not that high, it does have better ratios for your trades.

Taking a look at the profit chart shows us the following:

Parameter optimization

I also made the strategy file so that I could optimize the strategies EMA and Zscore level parameters. I chose to not optimize each and every parameter because otherwise I would then made the strategies results suffer too much from curve fitting. And that would cause problems with future performance.

So, for this video I will only optimize the best performing ATEEZ variant.

And that version is the spot trading version. It only has 1 point difference in the final score 269 against 268. Which is the closes finish in my books by a strategy. Very admirable for a strategy that is inspired by a K-pop band.

Let's see what this strategy does after it has been optimized on both the buy space as well as the ROI and stoploss setting.

In the end I did two optimization runs of 500 epochs because the initial hyperopt run did not prove to be very succesfull in terms of profitability.

And the second optimzation and backtest run one also did not get the score I was hoping for. So, actually I am thinking that this strategy is as good as it gets within this construction as it is.

The profit chart is also not that different of the initial backtesting version. It almost has the same drawdown start and end point.

There might be some small differences but the overall charts look the same. Altought the drawdown periode is a little bit smaller in the optimized one. Still nothing really woth of mentioning though.

Also, the trade count chart is a little bit different, but not really mentionworthy. So I think this is the best I can do with this trading strategy at the moment.

Strategy league

And at the end of this post we come to the conclusion. How does this ATEEZ strategy perform?

If I ask my strategy database the performance of all the strategies. Based on their scores, they eventually come up as the second and third best trading strategy. Giving the Clucmay a run for its money.

Eventually it is hard to compare the performance of a 5 minute strategy against a 4 hour timeframe strategy since they both have different objectives. Also the easyness of using these strategies with a bot or manually have to take into consideration.

Being that the ATEEZ strategy can also be used for manual strategy because of the way the indicators work. And also how they can provide you with good enough signals to trade makes this an algorithm that is not only versatile, but also easy enough to use under different cirtumstances.

End

Many thanks to my Patreons that support me with a small donation. If you become one, you will also enjoy the benefits of getting access to the notebooks and actual Freqtrade trading strategy code for your own usage.

Goodbye...!