LY Traders AI Lab

Research. Build. Deploy.
Building automation tools for MetaTrader5
Auckland, New Zealand.
[email protected]

[Sanjeev Kumar @SKSNZ]

When a chip reaches you, you don't verify its transistors — you trust the verification that ran before it. Most of us are drivers, not mechanics. But someone has to know the stack: the builders and funders. Trust is an outcome, not an education.

https://t.co/NN7Am81ZtU

[LY Traders AI Lab @lytraders]

Trust the outcome. Understand the stack.

When a chip reaches you, you don't verify its transistors. You trust the fabrication process and the verification that ran before it. The abstraction layer is what makes backend literacy optional — and that is not a shortcoming. It is the layer doing its job.

Most software users, most chip users, most traders are drivers. They are not served by being made into mechanics.

But someone has to know. The more understanding is outsourced to trust, the more a small minority must know the full stack — the people who build, fund, and steer it. That is the real division: builder and funder versus user. One needs the map. The other needs a verified outcome that behaves as promised.

This is why LY Bots ships the way it does: full source included, parameters yours, no black box. Transparency is the option to understand. Trust is the outcome of verified, consistent behaviour.

The outcome you can verify beats the internals you could memorise.

Read the full note:

https://t.co/eZOwP4Tpee

[Sanjeev Kumar @SK24DAO]

Every weekend, same routine:

Pull the week's data.
Re-run the behaviour scores.
Check each market's state.
Write down what changed — and what didn't.

Nearly 20 months of this. Same steps, every week.

Consistency in process before anything else.

[LY Traders AI Lab @lytraders]

@TradeTheCrowd This is the right rule — and it works because you pre-committed to it when you were calm, so you don't have to decide when you're down. That's the most reliable form of discipline: not willpower in the moment, but a choice made once, enforced every time. Well done you!

[LY Traders AI Lab @lytraders]

Loss → frustration → desperation → revenge trade → bigger loss.

That's the sequence that wipes out accounts. Not a market crash, not a black swan — one bad trade, then another, each trying to undo the damage of the last.

The money isn't lost to bad analysis. It's lost to a psychological loop: the second trade triggers anger, the third stops reading the chart entirely. The market doesn't care how anyone feels. It doesn't owe anyone a win.

The three mistakes, in order:

No risk limits — no circuit breaker that says "stop and walk away"
Letting emotion drive decisions — fear and frustration aren't character flaws, acting on them is
No automated guardrails — nothing between the feelings and the account except willpower
And willpower is not a strategy.

The defence isn't "try harder to stay calm." It's removing the hands from the controls. The bot waits when conditions aren't there. It doesn't know anyone is down. It only knows the parameters it was configured with.

Emotions can't be eliminated from trading. But they can be removed from execution.

https://t.co/qFa94NaJQq

[LY Traders AI Lab @lytraders]

"Perfect Sharpe. Robust to regime changes." In trading, those words are the red flag, not the proof. Internal consistency checks filter an LLM's own output — they don't create alpha. The operational-layer map is the real insight; the "collapse of the headcount moat" is the marketing around it.

[LY Traders AI Lab @lytraders]

@venturetwins And that too with German like precision!

[LY Traders AI Lab @lytraders]

@Raj5552 @irabukht And that's even harder to measure in the medium to long-term?

[LY Traders AI Lab @lytraders]

@hussainanjar Likewise, the magic is that it's mostly config files and no server to babysit. Vercel CLI ships the static build, gh CLI handles the repo/CI half. Two tools, no build step — Java deployments never had it this easy.

[LY Traders AI Lab @lytraders]

@AravSrinivas Surely Finance is where Perplexity has a real edge, nothing can beat that at this stage!

[LY Traders AI Lab @lytraders]

@MilksandMatcha Plus 💅 $

[Sanjeev Kumar @SK24DAO]

We don't slice the study into 11 equal pieces.

Each market's weight follows its behaviour score. Orderly markets earn more room. Erratic markets get less.

Same 11 markets. Different sizes. Decided by structure, not by hunch.

Weekly research update. Not live trading. #LYBots

[LY Traders AI Lab @lytraders]

@steipete FOMO explains the price going up. Confusion explains it staying up.

[LY Traders AI Lab @lytraders]

Three markets. One licence.

LY Bots covers three uncorrelated markets on a single MetaTrader 5 install:
• LY-HK50 — the Hang Seng Index — $750
• LY-JP225 — the Nikkei 225 — $950
• LY-Gold — XAUUSD — $1,250

Sold separately that's $2,950. The full bundle — all three, one licence — is $2,100. That's $850 less than building them up one at a time.

One-time. Lifetime updates. Full source included. Parameters stay yours; execution becomes systematic.

The market changes. The rules don't.

https://t.co/qFa94NaJQq

[LY Traders AI Lab @lytraders]

@RoundtableSpace Something I couldn't do even when wide awake!

[LY Traders AI Lab @lytraders]

Most people don't need to know how the chip is made. They need to trust the outcome — that inputs produce correct, reliable results. Trust is earned by verified behaviour, not by backend literacy. That's an abstraction layer working as designed.

But here's the part worth adding: the more we outsource understanding to trust, the more a small minority actually has to know the full stack — the people building, funding and steering it. That minority is your audience. So yes, make the end-to-end content. Users trust the chip; the people who decide its future need the map.

[LY Traders AI Lab @lytraders]

@AravSrinivas Congratulations Aravind on new launch, previous one didn't work and I had to cancel the subscription but I do like the Power of Perplexity in compliance!