Let's review what a traditional backtest isn't showing:

๐Ÿ”ธTimeframe

A lot of models work on high timeframes on the charts.

They work visually at least and that's why many are only sharing with high timeframes charts greater than 4H.
We all built a backtest based on a moving average cross and got a shooting 40%+ win-rate when running the engine for the very first time.

We all thought we were geniuses right :) too easy we all thought

There is a caveat though...

For margin trading (derivatives, options, futures, ..) backtests aren't account for real trading fees.
Let me explain... We all heard/saw those rollover fees that we need to pay overnight.
This is basically how the brokers are forcing the SWING traders to pay more fees.
While the explanation longs to pay for the shorts, and shorts to pay for the longs is poetic - those fees could eat out your position capital way before the price action has even moved.
Imagine a range during days/weeks.
You'll end up paying arbitrary fees and might have a very negative position size way before anything interesting (from a trading perspective) ever happened.

I saw many trades being -5/-10% PnL only because of fees. Then, imagine trading contracts with an expiration date - this adds another challenge - and most of the backtest don't even account for that.

๐Ÿ”ธ Leverage

Leverage increases disproportionally the risk compared to the opportunity.
Leverage 2 does increase the risk by 2 but the opportunity won't be multiplied by 2.
It would be in case the analysis is good in the first place. (assuming the risk/entry/exit plan is properly planned).
Assuming those analyses are made by experienced traders, then using leverage makes more sense.

๐Ÿ”ธ Past performance doesn't guarantee future performance

Probably the quote we hear the most in trading guys...
Generally in trading, what worked before has a probability to not work anymore the more time has been elapsed.
I don't mean it won't work anymore. This only means we should be cautious when trading setups valid from a while ago on a specific market

A way to not get that burden on our psychology is to indeed reduce the position sizing.
Until we are comfortable and not stressed anymore. That's the sweet spot you guys have to find.

For some, it might be a few hundreds per trade, for others a few thousands. There is not a well-formulated generic universal valid answer for what's the best position sizing

But for sure, once we get comfortable with one. we should go to the upper level direct above. Direct above means, if we trade 100 USD position sizes, the next one could be in the 110-150 USD range. (and not 1K right off the bat...)

We wouldn't lift 100 kg after getting used to only 10 kg. Trading isn't different than any skill requiring training and dedication
The difficulty is to not change our goals midway after a few wins or a few losses.
And to stick with them for a few months at least.

Literally takes me weeks of training to add a few kg to my chest press barbell or biceps curls. That's how much it took me also to increase the average position sizing by 20% or so.
Thus the more I increase it, the more time I need to get comfortable with it.

Applies in a lot of areas in life, sport, career also. It takes time! And we take pride in trying to reducing the research time for you guys by offering the algo frameworks, that we only use for our trading.
We think that reducing the number of choices in trading is the way to go. Less choices ends up with less doubts and mistakes.

The worst thing for new traders is getting early very lucky and rewarding trades.
That's what happened with many crypto traders in 2017.
They got too cocky and made that money too quickly and too easily.
Then, when the market turned bearish... they gave all it back because their experience/backtest/psychology/beliefs weren't ready for a market shift.

We're at a time where markets change constantly.
And perhaps that's why the patterns used by our predecessors 20 years ago aren't relevant anymore.

๐Ÿ”ธ Sometimes orders aren't filled

In paper trading or with a backtest, when a Take Profit is hit... well we make profit and that's cool.
But those LIVE trading know that sometimes... the limit orders aren't filled and no one can give us a logical explanation why the F they weren't filled. Even though the charts clearly show we should have been entitled to those orders...

Your broker will say slippage.
Your guru will say I got a nice 1000% profit. Hope you all exited when I told you so.
You will say But my backtest claimed that an order should always be filled...

I'm saying blaming the casino isn't useful.
And that's why trades need to be managed because we're playing in a house (exchange), competition (i.e. SMART-MONEY - understand bankers/funds with real advanced financial education) that want us liquidated.

โœ‹ Winning in trading is possible.

We must learn the rules of the game we're playing - but, not the rules taught to us by those reaping us off.
It's really doable but trading is the sum of a lot of small very important things to combine all together.

1 Backtest isn't enough.
1 indicator is not enough.
1 universal risk management rule isn't enough

We'll write another more in-depth thread on that because... well... we trade to improve our lifestyle and not only for playing.

Overfitting Charlatanism

A lot of Backtests from vendors are pure charlatanism and 99% of the cases not relevant at all
Please find below a study explaining more in-depth why most backtests are overfitting and aren't in any way predicting the future performance

The Effects of
Backtest Overfitting on
Out-of-Sample Performance

๐Ÿ”ธ Are backtests still relevant?

โœ‹ They are if measuring the correct metrics.
Sticking to a 99% win-rate doesn't mean anything. It doesn't quantify everything said above and the risk needed to achieve those.
Plus, 1 overleveraged/bad managed trade/without stop-loss can cancel everything out.

The Best Backtester engine measures the RISK

โœ‹ Almost all backtests available to retail traders are utter crap and made for losing.
Only because they don't account for real market conditions.

We made the first backtest for indicators on TradingView.
We're showing what a trader should really measure for validating a model and we take into account the brokerage fees in the calculations.
This engine allows to simulate your trading strategy with conditions as close as possible than the real live market conditions.

โœ‹Our message is the risk has to be quantified first. Once identified and reduced, then we're looking to increase the captured opportunity while keeping the risk low.
I want to throw away a misconception the majority has about Risk.

Risk is subjective

At a given time, 2 traders can enter at the same second but identify a different risk value.
Identifying the risk is a skill and a skill that needs education and practice.

This is the foundation of trading... not knowing your risk for each trade is pretty dangerous.

Would you go to a country in a war for your holidays?
Would you eat some food looking as bad as you could possibly imagine?
Would you get an unprotected physical one-night relationship with someone you met 5 min ago? (... I know it's a tough question)

If the answer is obvious is because we all thought through the advantages (opportunity) and drawbacks (risks). And when the risks are way greater than the advantages, we politely refuse and keep going on with our lives.

You won't have to calculate it using our frameworks because we made all those calculations.

โœ‹โœ‹ We wanted to simplify most of your trading by reducing the number of decisions/calculations/estimates to make for each signal.

โœ‹โœ‹ It may or not may the right way, our subscribers see quickly that the overall model makes sense and getting stopped out is pretty rare.

โœ‹ The win-rate/net profit are nice metrics but only an incomplete picture of the risk required to get that performance.
A trader should identify the risks of getting liquidated by uncertain probable futures events - which traditional backtests totally disregard. (as if they want us to lose...)

Technical guide for our Backtest engine

There is also a quick video showing what the engine does and how to use it
Video tutorial