My Weird Strategy

I was backtesting some freaky strategy I stitched together and out of 29 trades, 1 was a loss
image

I’m going to try it out today on a real account and see what happens
I currently have 2 setups on GBPUSD and EURUSD

I hope the setups are right :slight_smile:

they weren’t

1 Like

It’s not that the strategy trade didn’t work, I accidentally traded against it
The final confirmation for me to open a trade is that the stochastic oscillator has to be heading down but it didn’t

When I was watching the chart in real time, the stochastic crossed the signal line and was heading downwards. Immediately I opened my trade, it started heading in the opposite direction which is why the stochastic oscillator never crossed

Now, anytime I start seeing the signals to enter a trade but the stochastic hasn’t crossed, I’ll just wait for that candle to close on M30 and if it has crossed when the new one opens, I enter :upside_down_face:

From the data, it appears the best time to trade is from 7AM - 5PM (GMT+1)
image

Just made 2 cents, it could have been 15 but I entered late
image

If I had let it run, I could have caught a move of 120 pips :face_with_diagonal_mouth:
image

In case anyone is wondering, it is a trend-reversal strategy
The strategy involves determining the trend of the market on the daily time frame, then going to the one-minute timeframe and waiting for a trend that is in the opposite direction
So if the market is in an uptrend, I will look for a downtrend on the 1M.
When I find this downtrend, I’ll find the bottom of the trend so I can profit from the reversal
I do this by using indicators and some kinda freaky hybrid between S&D zones and break of structure

The main idea is that since the market is currently trending up, the downtrend I see on the 1M will basically be like the trend taking a break and then the trend will continue after the 1M downtrend reverses

So far, I’ve only tested it for pairs that have USD in them (GBP/USD and EUR/USD) since they tend to not have that many single candle spikes when there is no news happening

2 Likes

This is true for every strategy, it’s one thing to know the statistical p/l of your strategy through backtesting, it’s another thing to unlock its full potential of every gains and loss of same strategy through forwardtesting.
To do this, one have to be equilibrant of both the human & the computer mind which is for all a weird task to accomplish.

Goodluck on your strategy, do you mind to share with us what you think would made the difference as regards your live trade & backtested trades?

Warm Regards,
Ayọ

2 Likes

I think on the backtests, some entries may be a result of market noise and not market behaviour.

Since my strategy relies on the stochastic oscillator, the lines in it react in real time to price as it changes, which may lead to false entries since they crossed as the market was moving but then suddenly uncrossed when the market moves in the opposite direction
And some times when I’m backtesting, the stochastic crosses on one candle but then immediately cross the other way because the next candle moved in the opposite direction so on the candle that it crossed in the direction you want, it is an entry but if it was a live trade, the entry may have appeared after the market had already moved in your desired direction and is already losing, making your entry late

To anybody who has found the BOMB strategy, please trade it on a demo first before using it and if it doesn’t work at all, then it is probably curve-fitted(It only worked under certain specific circumstances which don’t repeat themselves in live trading). That’s what I did

I backtested the strategy on EUR/USD and got surprising results
Out of 29 trades, 9 were losses
image
I discovered that on the day that there were losses, if the candle candle on the daily timeframe is green, the good trades on that day surpass the bad ones, thereby turning the day profitable but if the candle on the daily timeframe is red, even if there are good trades, the bad trades will surpass them and make that day a losing one
image

According to the test, if you had an account of $700 and traded a lot size of 1, you would make a total of $809 over the course of 5 days. If you hadn’t traded on that last day, it would have been $971.
You may be wondering how 1 pip equals $1. If you trade a pair where the currency of your account is the same as the quote currency (the second currency in the pair), 1 pip be equal to 1 unit currency movement in your account

I have learned a lot from a strategy that was Frankensteined together and just seems to reasonably work, including how to backtest using spreadsheets

A daily trend reversal on GBP/USD may have happened
image

And also on EUR/USD
image

What is your entry and exit criteria?

Your backtest is way too small to be meaningful. The reason green candles are better is because it’s been largely in an uptrend for the last few months. Over a long time, that won’t hold up in forex

1 Like

Your backtest is way too small to be meaningful. The reason green candles are better is because it’s been largely in an uptrend for the last few months. Over a long time, that won’t hold up in forex

Interesting
So how long do you advise I backtest for (how many trades)

Also, my TP is determined by MA crossovers and my SL is determined by a S&D zone with a birth defect

A backtest needs to be for years if you want to know the true potential. The market changes between trends in either direction and ranges across every time frame. You need to know that your strategy works in all of them. Or if it doesn’t then how to identify when to wait.

Covid blew away several strategies I had because it threw volatility about. Now I have different strategies that work in current conditions but weren’t as good in 2018 or 2019 when backtesting. So my backtests are 5 years, which is many thousands of trades on that strategy.

1 Like

DAMN!!

5 years; that’s a long time
Did you do all those trades manually or did you automate the backtest?

I understand the part about COVID and volatility since really big events shake up a lot of things about the market
Now I understand why people do backtests that can reach as far back as 1990; it’s to test its profitability in all market conditions

it all makes sense now

1 Like

I recently wrote what I thought was a great strategy that was performing brilliantly in backtesting, that is until I included March 2020 in the results, of which no matter what lot size I used it would have wiped out the entire account or been in drawdown for the best part of 2 years. Needless to say that one went in the bin. I know you can’t factor in all black swan events but there needs to be a certain robustness to any strategy.

Just because it didn’t work 5 years ago doesn’t mean you shouldn’t trade it. You just need to figure out how to know it’s not going to work and stop. Easier said than done.

Unfortunately I didn’t take a proper screenshot or save the results, I only took a photo on my phone to send to a friend, really it was pretty impressive losses though.

I was looking into a mean reversion type strategy that had been bugging me for a while, unfortunately in March there was no reversion it was just all mean :rofl:

Needless to say, I only use volatility and volume these days to trade :+1:

2 Likes

Apologies, didn’t mean to crash your thread here, anyway, as @chesterjohn said, backtesting should be over years not a few trades, definitely need a good sample size

Yeah
But how do I automate it

You need to use some kind of software, apologies I’m not sure what level you’re at so I don’t want to patronise you. I either use MT4 or MT5 and code an EA that you can then backtest or if it’s a fairly simple strategy I use ProRealTime and either code direct or the beauty of PRT is you can use more of a gui type interface and just click on the chart/indicator and use custom commands with no coding needed.