MoonTheTrain

Indicators for MetaTrader 4 and 5 — 64 tools, each tested before it was sold. Backtest reports, research findings, and custom builds. 3-day trial, no card.

[MoonTheTrain @moon_the_train]

Full write-up: https://t.co/WTncL4X2iq

[MoonTheTrain @moon_the_train]

Should you trade with Jev? My verdict after testing it

After 19 Jev finance projects, a code review of QuantDinger and my own Bitcoin test: don't let Jev make or veto trades yet. Where it does help, and a checklist. #jev #backtest

[MoonTheTrain @moon_the_train]

Full write-up: https://t.co/38l1dGeexQ

[MoonTheTrain @moon_the_train]

Can Jev predict Bitcoin? It read the indicators backwards

I gave Jev 12 indicators on 1,500 unseen BTC 5-min candles. It read them right 99.8% of the time, then scored AUC 0.49. A plain model got 0.52. #jev #bitcoin

[MoonTheTrain @moon_the_train]

Every platform build, and what it can't do: https://t.co/nCHFb7MxdF

[MoonTheTrain @moon_the_train]

Your previous-day-high line is still pointing at a level price already ran through. On GBPUSD H1 since 2020, half of all daily highs and lows were gone within about 14 hours.

Liquidity Hunter drops a level on the bar that takes it. #MetaTrader #FVG #XAUUSD

[MoonTheTrain @moon_the_train]

Every platform build, and what it can't do: https://t.co/BywA8g9oQs

[MoonTheTrain @moon_the_train]

Most trailing stops bend the moment price breathes, and take you out of a position that was fine.

AlphaTrend can only move one way while the trend holds. Sometimes it sits flat for twenty bars. That's the feature. #MetaTrader #supertrend #XAUUSD

[Mark Minervini @markminervini]

Let me be clear: I'm a big fan of AI. At Minervini Markets, we are using AI models to find stocks and analyze data. But we are using it as a tool. In that way, AI is amazingly helpful and a powerful addition to our platform.

[MoonTheTrain @moon_the_train]

I'm not saying Jev is useless. But "AI understands indicators" is not the same as "AI can predict price."

Next time a Jev bot shows off profits, ask: was it tested against a simple model on data it had never seen?

[MoonTheTrain @moon_the_train]

Everyone says you can make money trading with Jev.

So I tested it: same 12 indicators, 1,500 unseen BTC 5-min candles, one question: will the next candle close up or down?

Jev read every number correctly. Then it read them backwards. 🧵
#typesaveai #jev

[MoonTheTrain @moon_the_train]

Real BTC 5-min data leans the other way: after a stretched move, the next candle tends to pull back slightly.

On the 300 candles where boosting was most confident, Jev disagreed with 294.

[MoonTheTrain @moon_the_train]

To be fair: boosting's 0.52 isn't enough to beat exchange fees either. Price-only 5-min prediction is hard.

Scope: jev-1.13.0, BTCUSDT 5m, price data only. The whole test cost $0.19 in API calls.

[MoonTheTrain @moon_the_train]

Adding Jev on top of boosting didn't help either, as a filter or as a blended probability.

The reason: Jev reads indicators like a textbook. Price above the EMA, high RSI, a green streak → it expects more upside.

[Mark Minervini @markminervini]

If AI could truly figure out how to trade, how would one AI trader beat another?
IT COULDN’T.
You need inefficiencies to generate alpha. If every AI sees the same data, finds the same patterns, and makes the same “optimal” trades, the edge gets arbitraged away.
Even Jim Simmons understands this. That's why he as admitted to manually monitoring and changing his models routinely. And that humans can override his models at the risk level.
If humans are constantly testing and changing the AI models - based on their findings - then that AI is a tool, not a solution.
In his TED interview, Simons said:
“These things fade after a while; anomalies can get washed out.
In another interview, Simmons confessed: “The computer is just a tool that we use.”
I rest my case.

[Mark Minervini @markminervini]

If AI can trade stocks so well, why are all these AI-managed portfolios at Fidelity, Schwab and elsewhere struggling to even beat the S&P 500?
And where are the AI portfolios winning the U.S. Investing Championship?
Backtests and predictions are easy. Show me the audited, real-money results. That’s where the bullshit stops.
The answer is because it can't. There is zero evidence of any enduring or consistent results from purely AI or computer manager portfolios that can even outperform the S&P 500. Fact.

[MoonTheTrain @moon_the_train]

Every platform build, and what it can't do: https://t.co/aO9lcv2NjC

[MoonTheTrain @moon_the_train]

Your ATR trend band decides the flip using the same bar it's drawn on. So the level you planned around at 09:00 isn't the one that fires at 09:59.

Money Moves tests the close against the previous bar's stop. Settled before the bar opens. #MetaTrader #supertrend #XAUUSD