Works even better if you look at LAST price > LAST DAY HIGH.
Here are results from 2024-09-08 until 2024-10-07.
122 trades, 0 losing trades. 100% win rate. 5.79% profit. If you trade 1 lot per trade, that’s 5.790 USD in profit over 2 months. If my math is not incorrect here (often is).
It’s always been my dream to find a simple mechanical trading strategy like this. The thing is, even if you have a very high win rate, you cannot depend on it being always 100%: so a stop-loss rule must be available. Where do you set the stop-loss?
Well done - you seem, through your own research, to have stumbled on price action trading.
Day traders and scalpers will work this kind of system.
Waiting for a pullback then enter when first candle exceeds the previous candle and exit on a double top or double bottom if shorting.
Excellent for scalping on 5mins timeframe.
I have not figured out it yet. Problem with 1D charts you don’t know which direction goes first. So theoretically if set stop loss at yesterday low, this price point might hit first before it reaches the yesterday high. I will do some more testing today, probably implement 1H into the 1D price action.
I been thinking about another approach, short the yesterday highs or long the yesterday lows. There are a lot of opportunities of course. Its just sticking with a system for long-term instead of always trying to improve it is my biggest challange.
Yesterday I lost the original code/idea because I was trying to improve it and now I have to go back and re-create the same scenario as yesterday.
There are a couple of things of course not that bad experience. I pull the data from Metatrader 5 with the Python API, do the RSI/MACD calculation in PHP etc. Another script that read from the database and executes the trades. My system is a lot better today than it was 2 years ago.
Another thing I realized with forex, whichever direction you take SHORT or LONG on 1D chart. Eventually at some point of the day you will be in profit. That’s the crazy thing with forex. The high and lows seem to be equal. You can take a wild guess, set the take profit at $50-100 and whithin 1 day you will have that money in your account.
If you just want to make $100 everyday, short the yesterday high, take profit at $100. 99% of the time it will work. But you will miss out on trades that generate $1200 or $500 in 1 day
If money can buy time and you are living in a low-cost country like Phillipines or Thailand, $100 per day can buy you enough time to figure out how to make $5000 per day with forex.
Update:
Short the yesterday high. Do mean reversion on 1H charts to figure out when RSI hits bottom. Take profit when RSI crossover or MACD goes back into buying power.
It’s like Warren Buffet says (not that I am too interested in him). You buy stocks at bargain. If the stocks not make profit right away, eventually the prices will return.
An example: I bought AAL.NAS at 10.97. Within the day it goes up to 11.18. So now I am down -0.21. But looking at the data, look below.
Oooh. Very interesting approach and impressive results. It’s very straightforward and easy to understand. I was also going to ask about stop losses, but I see you’ve already answered them on your latest responses. I just wonder, have you noticed any specific pair where this works particularly well? Any specific pair that it doesn’t work on?
The new strategy is this. Buy at RSI top/low, sell at the high/low. If price is not found immediately wait for 1-2 days. Everytime it says “Sell posisition at x. Success OK”. It means the low of the day was found later in the dataset.
As for your question which pair it works on v.s which pair it don’t work on. Well the strategy works on any pair but there are two important things to watch out for. First off, the spread. On some ‘exotic’ pairs like MXNUSD, PLNEUR, SEKEUR the spread is higher than EURUSD, GBPUSD. Not always, but especially early in the morning exchange hours. Second, at one time my algorthim bought a pair so high in price that it was not able to find the lowest price later. I was down $1000 at end of the day, next day down $2000, then down $3000 following days. I am going to try this algorthim on 1H timeframe instead of 1D timeframe because of this particular problem.
I will keep you guys updated after a few more backtests
Between 2024-08-01 to 2024-10-08:
Total trades 258
Winning trades 246
Losing trades 12
Total profit 14.07%
Lot size: 1
Total profit in USD: $14.070
Winning percentage: 95%
Best timeframe is 1H. Everything less than 1H the spread and commisions are eating up the profits. This is only for short positions btw. Not yet tried for long positions (probably total profit increase by another 15%).
Now for the challenging part. Where to take profit we already figured out now. The 12 losing trades we must minimalize the damage without effecting the profit percentage too much. The lossing trades are already included in the profit percentage.
The id after each Success OK tells where the position exited at later date. So far 26% gains in 2 months. With 2 lot size and $6 commisions that would amount to 52 000 USD in gains and -3378 USD in commisions. We could probably increase the lot size a lot more. To reach $100.000 in gains it would take a minimum of 4 lot size on each trade.
If close > higher than previous high put a sell position on next bar. If close < lower than previous low put a buy position. Take profit is whenever the high/low reaches of the current bar at a later date. As you can see some positions close a couple of days later.
Winning percentage:
Of 76 trades all are 100% winning. Tested on botth EURGBP and EURUSD both 100% win rate.
Logic:
Probably mean reversion? Price will return back to it’s normal level after a high/low.
Yes, MA could be useful but not using for this setup. I have another setup using using MA for entry and RSI for exit on weekly charts.
When people realize that trading is 50/50 gambling. And you are using different strategies like looking at stock opens, high/low, following the trend with MA, oversold/undersold conditons with RSI to increase the edge.
Other than that you could as well just flip a coin.
I think for most people the idea to make incredible amount of money with small initial investment on margin is what is most attractive about it instead of going to the casino.
I do keep to heart what Charlie Munger said about Jim Simons that he is one of the greatest non-degenerate gamblers.
I spent maybe 1000 hours looking at charts and 50 hours listen to other people. Now I come to this point looking at charts is not even fun any longer and I prefer just to do everything with OHLC values instead.
Very interesting approach and impressive results. 
I was also going to ask about stop losses, but I see you’ve already answered them on your latest responses. I just wonder, have you noticed any specific pair where this works particularly well? Any specific pair that it doesn’t work on?
