Tyche

building algorithmic trading systems

[Tyche @xtyche1]

I am a 21 y/o Economics student from Germany.
Got pulled into systematic trading and quant finance during my degree.

Before this I traded discretionary for 2 years. Mostly ICT.
I knew the theory — risk management, HTF bias, psychology… Stayed around breakeven the whole way, couldn't really make it work.

I still believe discretionary trading works. But honestly?

Most "traders" online are larping. One winning streak, big flex, no proof of an actual long-term track record or edge.

What can they show? A few payouts. Over what horizon? Markets carry enough noise that anyone can win for a while. Very few sustain it for years — way less than you see online.

Why I fell in love with systematic trading:
→ I can backtest decades of data
→ I can define exactly what and where my edge is
→ I know with higher probability which regimes will hurt me and which will feed me
→ I don't need to be 100% psychologically resilient every single day to follow the rules
→ No cap on how many systems I can run in parallel

Over time I built — with my own research, experience + Claude — 3 ready-to-deploy MT5 algo EAs, each running on its own CFD prop firm account:

Portfolio A → 6 strategies
Portfolio B → 5 strategies
Portfolio C → 6 strategies
11 unique strategies, 8 assets (some overlap across portfolios).

Mainly momentum / trend-following. A few mean-reversion plays sprinkled in.
Everything purely technical, fully rule-based, no discretion.

Backtested 2018 → today. Each portfolio sized for its own 100k prop firm account → 300k combined reference. Metrics across the full portfolio:

• Sharpe ~3.0
• Profit Factor ~1.4
• Max DD <10% on every single account
• Annual vol ~5%
• Zero blowups

Yes, 5% annual vol looks low. That's the point.
Each EA is sized for its own 100k account. Drop all 3 onto a 300k view → vol drops further.

Boring on purpose. Prop firms reward consistency, not heroics. And of course I can just increase the global risk 2x if I have a 200k account...

Past results never guarantee future ones. I know that. But the edge is defined, repeatable, falsifiable. That's the difference.

Your favorite guru can laugh at 2%/month. I'd rather scale across multiple prop firms with consistent gains than chase 10%+ / month and blow up once a year.

Everything is still theory.
My goal: scale aggressively across prop firms now, while I keep refining systems, learning more technical skills, and finishing my degree.

I'm gonna share more insights into how I build, my workflow, research, strategy evaluation.

You can follow me on this journey for free.
We build in public from here!

[Tyche @xtyche1]

I chose algo because I want to scale beyond my own attention span.

Not because discretionary is fake.

If you trade discretionary and you've got multiple years of consistent live results — respect.

If you trade algo and you laugh at every discretionary trader by default — you're flexing your tools, not your understanding.

Build edge. Measure it as honestly as your medium allows. That's the whole game.

[Tyche @xtyche1]

The quant corner of Twitter loves to dunk on discretionary traders.

"No backtest. No edge. Just vibes."

I get it. I went algo for a reason. But the take is lazy.

Discretionary traders absolutely can make money. The reason most quants don't see it has nothing to do with whether the edge exists. 🧵

[Tyche @xtyche1]

The actual goal — for both sides — is the same:

A defined system that produces positive expected value with the lowest vol you can manage.

The discretionary trader's "system" is just stored in their head + reps instead of in code. If they can't articulate it, they don't have one. Same standard as a quant.

Edge is edge. Process is process. The medium doesn't matter.

[Tyche @xtyche1]

Rough thresholds I use:
• Return Spearman < 0.3 → useful diversifier
• DD Pearson < 0.4 → ok on same account
• λ_L < 0.2 → low tail co-crash risk

Most unexperienced algo guys backtest 8 EAs, dump them on one account, call it a portfolio.

They built one strategy in 8 files.

A portfolio is a correlation structure. Not a list.

[Tyche @xtyche1]

"Diversification is the only free lunch in investing."
— Harry Markowitz, Nobel Prize 1990

Sounds like textbook filler. It's the most concrete edge I have.

Stack 3 mediocre strategies and you get a portfolio that beats every one of them.

Same trades. Same risk. Higher Sharpe. For free.

The catch — only if you measure correlation the right way. And most don't. 🧵

[Tyche @xtyche1]

One correlation number isn't enough. I look at three:

→ Returns — Spearman on daily P&L. Rank-based, robust to outliers, catches non-linear co-movement.

→ Drawdowns — Pearson on daily DD series. Here magnitude matters: I want to know if A is down 10% when B is down 12%. Ranks would blur that.

→ Tails — Lower Tail Dependence (λ_L). The probability that B has a worst-5% day WHEN A has a worst-5% day. This is what actually blows up accounts

[Tyche @xtyche1]

@x256xx Nice ragebait

[Tyche @xtyche1]

@DylanElchami Pleaaase offer 200-500k accounts on this

[Tyche @xtyche1]

@ZenomTrader TJR himself said hes renting the cars for marketing purposes, not because he likes them so much 🤣

[Tyche @xtyche1]

@tassadaro22393 Hyperliquidated

[Tyche @xtyche1]

@merkelfanboi @Tedmarsgoi2 Ne, aber OF...

[Tyche @xtyche1]

@TheGreekTrader @Polymarket Is holding time relevant here???

[Tyche @xtyche1]

@alex_hunter20 fugazi

[Tyche @xtyche1]

@ByteSilix drainer, do not connect guys