@RHerman One thing in backtesting that works really well is to split your backtesting data into two sets. The first timeframe for backtesting, the second timeframe is what you use to forward test your model. Also look into MAE and MFE
English
Onchain Edge
3.1K posts
Onchain Edge
@onchain_edge
I write the Onchain Edge — a weekly 5-min crypto breakdown + Bitcoin dashboard to help crypto traders cut noise, manage risk, and trade with conviction.
🔴 Onchain Edge → Katılım Nisan 2022
124 Takip Edilen4.6K Takipçiler
I wasn’t planning to build an algo model.
But I accidentally found something interesting while backtesting.
And the numbers surprised me.
So of course…
I went deeper.
The idea started as a simple observation in $NQ futures.
A very specific time-based behavior that kept repeating in the data.
At first I assumed it was just noise.
So I tested it.
Backtest range:
Jan 1, 2020 → Today
Results:
• 1410 trades
• 77.9% win rate
• Max portfolio drawdown $1,880
At that point I still assumed:
“Probably curve-fit.”
So I did the only thing that actually matters.
I built a fully mechanical strategy in Pine Script and tested it again.
Same logic.
Same rules.
No discretion.
AND THE EDGE STILL HELD.
—
The model is extremely simple.
Every day at exactly 2:00am New York time
the system gives one trade on Nasdaq.
No prediction.
No analysis.
No chart watching.
Just:
Long or Short.
—
Execution rules:
• 1 micro contract (MNQ)
• TP: $300 (150 points)
• SL: $300
If price reaches +$250
I move stop to +$200 locked profit
So the outcome becomes:
• +$300 win
• +$200 secured profit
• -$300 loss
—
Backtests are one thing.
Markets are another.
So I started forward testing the system live.
A dedicated prop firm evaluation account LUCID
running the model exactly as designed.
No overrides.
No discretion.
Just execution.
—
Day 1 of forward testing (today)
+$300
System followed exactly.
Goal of this experiment:
Can a simple mechanical model
pass a $50k prop firm evaluation
trading one trade per day?
Every day.
2:00am New York time.
The indicator decides.
—
I’ll be posting results publicly.
Wins.
Losses.
Drawdowns.
Everything.
Because most traders don’t fail because their strategy is bad…
They fail because their execution is inconsistent.
This experiment removes that excuse.
—
Let’s see if the edge survives real markets.
Would you trust a system that trades
once per day at the exact same time?
#AlgoTrading #FuturesTrading
Herman Trading@RHerman
I got interested in algo trading recently. And I think I may have found a real edge. It’s 100% mechanical on $NQ Every day at exactly 2:00am New York time, the system gives a trade automatically. No prediction. No overthinking. No emotional bias. The indicator shows: long or short entry exactly at 2am. stop loss take profit break-even trigger It also has built-in alerts. The framework is simple: TP: 150 points SL: 150 points On 1 mini contract that means: +$300 win -$300 loss If the price reaches +$250, I will stop to BE in profit at +$205. That’s the whole model. What makes it interesting is this: In backtesting, it delivered 75%+ win rate and the historical drawdown never reached a level that would have blown a $50k prop firm account. So now I’m doing what most traders avoid: forward testing it on a live market with a dedicated eval account Because backtests are not enough. But when a model is: mechanical repeatable risk-defined time-based simple to execute …it deserves to be tested properly. Too many traders lose money not because their idea is bad… but because their execution is inconsistent. This experiment removes that excuse. Now the question is simple: Does the edge survive real conditions? That’s what I’m about to find out. Would you run a system like this live? #AlgoTrading #PropFirm
English
@minhtrungct Mostly busy with real estate and my two kids. Priorities shifted. Got burnt out trying to grow on twitter
English
Bitcoin onchain metrics worked differently in 2017 than 2024.
MVRV hit 3.5 at the 2018 top. It barely crossed 2.8 in 2024.
Static thresholds miss cycle tops.
I rebuilt OnchainEdge from scratch.
English
@anicesatnight Thanks man! Great input will try and fix it soon
English
@onchain_edge Thank you so much for sharing, and congratulations on the improvements! For now, I should mention that the UI on phones doesn't adjust well to the screens.
English
I didn't rebuild this to fit the backtest. I rebuilt it to fix the flaws. The thresholds were broken. The sell logic was buggy. Flat dollar deployment didn't scale with portfolio.
v2 adapts. Percentage-based deployment scales.
English
English
@digitalartchick @Bluntz_Capital Narratives always follow price. Not the other way around. People always try to find a reason why price moves
English
@Bluntz_Capital You sure that pump was the TA and not the Robinhood listing?
English
33% in under a day for moodeng🦛
we're back baby
Bluntz@Bluntz_Capital
quite liking moodeng here, nice 5 wave rise and abc down to the 0.618 and trying to break the bull flag now. if it does i think it looks good for a strong leg up 🦛
English
@intuitio_ Excellent advice. But most will ignore it and get rekt
English
Follow @onchain_edge for deep market threads like this.
Or join 700+ investors reading Onchain Edge Newsletter
The only crypto newsletter that combines narrative, positioning, and onchain structure.
📩newsletter.onchainedge.com
English
So who is Satoshi?
Maybe Hal Finney.
Maybe someone we’ve never heard of.
Maybe it doesn’t matter.
What matters is the idea:
A system so strong, its creator had to disappear.
English
He created the most powerful financial system of the 21st century.
Then vanished without spending a single cent.
Now, $113 billion worth of Bitcoin is just... sitting.
Here’s the story of the ghost wallet that could crash the entire market:
🧵
English