Strategyquant Success

HI Everyone,

Someone emailed me today asking me about how I have found StrategyQuant after seeing that I’ve been a user for over 2 years and this prompted me to write about my journey using it and where it’s led me to this date. I’ve very much enjoyed using the product and initially after clicking a few buttons and getting some decent looking backtests, I thought.. “this is it! I’ve now got algos which will just make me an immense amount of money!” However, if only it was this easy.. indeed like a lot of beginners, I put these into live operation to find that the results were nothing like backtests… even getting the backtests to look as good in Metatrader with tick data didn’t seem to work.

However, I stuck at it and you have to remember that StrategyQuant is just a tool to help you pretty much like how a calculator can help you do complex sums quickly and accurate BUT… it depends on you pressing the right buttons and feeding in the correct information. StrategyQuant is exactly the same.. you need to make as the operator you know what you’re doing.. take time to learn about the software and also take the time to learn about how algo trading works in the real world. Most systems fail because of curve fitting…  then every broker differentiates in terms of their prices, spreads, timezones.. even their data feeds. What may seem like small differences in their data feeds can make a world of difference and depending on the entry / exit criteria and the indicators you’re using can all yield very different results.

To cut a very long story short, StrategyQuant is an amazing product but take time and dedication to learn.. and experiment. I’ve put in at least an hour a day over the past two years experimenting, testing and refining. There is no perfect solution and no perfect algo… markets will change but the key is to build a portfolio of algorithms and to constantly refine the portfolio.. monitoring them and removing when they start to fail and adding ones which seem to be working into your live working portfolio.

Having a workflow with StrategyQuant in the middle of it has helpded me make it work. This year, I replaced a portfolio I was running with all StrategyQuant algos I took through  this workflow. The result is a portfolio that has won multiple rounds of funding with the well respected Metatrader broker Darwinex. And as I write this, the portfolio is currently 5th out of over 3500 other systems for December 2020.. if it can remain here, then it will be the 4th allocation of funding in just 6 months (funding is allocated every month)

So this is all thanks to StrategyQuant.. I really appreciate the product and the team for continually updating and improving it! Thank soooo much!

My trading system (or portfolio) is called YFC. Not sure if I can post the link to it so you can see it’s results live:

Darwinex YFC system

how many net pips does a good strategy like that make?

ivan wrote:

how many net pips does a good strategy like that make?

Looking to average 350 pips per week total across 6 pairs from 80-90 trades each week. Historically that is how it has performed on average the last 3 years.

by 6 pairs you mean 6 individual EAs or strategies ?

the number you gave, would translate into aprox 166 pips/month per strategy which is very good given the lowest timeframes which are notorious difficult. My record was a strategy on XAUUSD on H1 which had 480 pips/month on many relevant months. Anything below 30 pips/month i throw away

i never tried lower timeframes than M15 and even M15, only on a few pairs but i will generate in the near future

i am using mostly M15-H4 TFs on those basic pairs – EU, EJ, UJ, GJ,GU and GOLD

every market behaves differently, some markets could be in a loss for whole year

ivan wrote:

by 6 pairs you mean 6 individual EAs or strategies ? the number you gave, would translate into aprox 166 pips/month per strategy which is very good given the lowest timeframes which are notorious difficult. My record was a strategy on XAUUSD on H1 which had 480 pips/month on many relevant months. Anything below 30 pips/month i throw away i never tried lower timeframes than M15 and even M15, only on a few pairs but i will generate in the near future

Yes 166 pips/month per pair on 1 EA is correct. The exact same EA gives great results on all 6 pairs.

My just developed EA was profitable in M1, M5 and M15 in most of those pairs, but M1 was the standout, so that is all that I am trading. On all 6 pairs the equity curves I have been getting in backtesting have been fantastic.

I just started using SQ 3 weeks ago, will add other strategies as I find them. Ideally I want EAs that are each individually profitable and will give profits in any market condition. When a range trading EA is performing badly, have a trending EA that performs great at the same time, to avoid drawdown. Currently I am only opening trades 6 hours a day, so would be great to have systems that work on the other hours as well.

Congratulations Alan on coming 4th in the Darwinex table this month and a nice capital allocation through that.

you just caught corona virus moves and trends

ivan wrote:

my experience or story is far worse than allan lee’s and its possible that i will scare off beginners, that’s why i hesitated to write after some 3 years of generating and testing, i failed to make a single euro or eurocent, not even recover some of the investments, not to mention real profit my best strategies were on timeframes ranging from M15 to H4. The daily timeframe is the only one i haven’t generated because simply forward testing on such huge timeframe, takes years the pairs are in descending order of profitability: XAUUSD, GBPJPY and GBPUSD. Some very isolated cases, insignificant on EURUSD, USDCHF. The best strategies tested for several months until some 20 closed trades, had 65 – 100 pips/trade and an average of 150 – 450 pips/month. Other less profitable which i kept, had some 30 – 50 pips/trade and 100 – 200 pips/month. Below this line, any strategy is garbage. The number of final working strategies were 10 – 12, at the end of a complete generation process on most of forex pairs and timeframes. If we assume a ratio of 1 EUR/100 pips, that in theory would be translated into 30 EUR/month (3 EUR/strategy x 10 strategies). This ratio of 1 EUR/100 pips would be reasonable for most small beginner accounts with volumes of 0.01 and medium leverage. In theory, such a trading portfolio would double the account every 8 – 10 months. This is on paper, because in real life, there were always unexpected or unplanned hiccups unrelated to SQ software, like a broker deleting the history or unexplained differences in behavior. I use this unit of measure, pips/trade or pips/year because its the easiest for beginners to grasp and to convert of all. Once you input the leverage and volume, you can easily convert this in exact sum for every trade. This post and what i wrote is by no means a negative review of the software itself but a warning for not all beginners but for those who underestimate the complexity of trading itself and those who seek rapid and easy profit. I can assure them that this is not one of those solutions.

I’m sorry to hear this but it’s definitely true that it’s not as easy as clicking a few buttons and getting a working strategy. You have to be really harsh in your selection process with StrategyQuant.. only after selecting the best ones, there’s still no guarantee.. you have to test it in your broker to see if it executes and similar to the backtest results you are seeing.

In StrategyQuant, I test with multiple data sets.. import historical data from your broker, use the Dukascopy and another reference. If the backtests graphs don’t look similar, immediately discard. All my strategies are relatively simple.. I don’t select too many options for the criteria.. just 1 or 2 indicators max and I’m using orders to get in and out. Also keep away from the smaller timeframes…M1 and M5.. dont’ usually work. M15 I’ve had success with and H1. Also about 1/3 of my testing is out of sample. Also don’t expect your algos to last forever.. at some point they will fail, your job is to identify when this is and replace with another that you’re testing in the background. You know when they are failing as they are not performing as per your backtest results… ie no. of wins to losses, trading frequency etc.

Gin wrote:

you just caught corona virus moves and trends

TBH that’s the whole point.. you generate algos that work across a variety of conditions.. my portfolio is a mix of algos which work in trend and non-trend situations.

Right now, the ones which are doing most of the trading are the trending and breakout ones… the others are in the account not doing much. Once the trends stop being trends, I’m expecting the reversion strategies will kick in and the trending ones will sit it out!

mattedmonds wrote:

Congratulations Alan on coming 4th in the Darwinex table this month and a nice capital allocation through that.

Thank you Matt.. this is actually the 4th allocation of funding from Darwinex now bringing me to 197k euro allocation!

alanhere wrote:

Gin wrote:

you just caught corona virus moves and trends

TBH that’s the whole point.. you generate algos that work across a variety of conditions.. my portfolio is a mix of algos which work in trend and non-trend situations. Right now, the ones which are doing most of the trading are the trending and breakout ones… the others are in the account not doing much. Once the trends stop being trends, I’m expecting the reversion strategies will kick in and the trending ones will sit it out!

Thats what I am working towards. I have some great non-trending EAs that scalp ranges. Now need to work out some great trending EAs, so they balance each other out.

Hi Alan,

Just for curiosity, what tests (e.g. Monte Carlo, Slippage, WFM,…) are involved to generate your strategies?

Hi,

great stuff, thanks for sharing and also for mentioning the hard work that is required