anawan

[anawan @anawan1]

@MrMilkTrading Got me!

[anawan @anawan1]

@MoonTrades7 Not trying to be a hater. I made a bot to capture each challenge trade and track from the time the announcement was made. This example uses $500 per trade. 1 winner so far out of 13. Am I missing something?

[anawan @anawan1]

@Prophitcy SPX

[anawan @anawan1]

@apoorvsinghasr Ran a 5 yr backtest on 5 random tickers and SPY. Tried to follow your rules exactly. Pretty awful results. Do you have data to support your claims? The small number of trades says curve-fitting

[anawan @anawan1]

@RyanHoffec57252 @Sandeman52 This.

[anawan @anawan1]

@MoonTrades7 $QQQ

[anawan @anawan1]

@MrMilkTrading @therealmorph835 @McClellanOsc @LindaRaschke Good one! I almost choked on my G&T!

[anawan @anawan1]

@1harryrand @aw_trades_ FWIW, I used Sol set on Ultra to reproduce the pdf as a backtest. I got zero trades over a full year. Ran a sensitivity test to find what would have to change to get trades and "extreme" relaxation of requirements got me to 40 trades.

[anawan @anawan1]

@aw_trades_ Thanks for this! Do you typically get 1-2 trades per day or significantly less?

[anawan @anawan1]

@MarvintheMmd @UCompounders Free Cash Flow

[anawan @anawan1]

@MarginCall4 If it looks too good to be true....

[anawan @anawan1]

@Zabaroptik @MrMilkTrading I'm in!

[anawan @anawan1]

@MarginCall4 Can you share the yearly numbers? I found it crapped out pre-2020, but I didn't go back as far as you.

[anawan @anawan1]

@L1vsun Go away. The post is engagement-bait wrapped around a real concept, and its central claim — that the transition matrix is the edge — is demonstrably false. Using 10 years of SPY daily data, the "edge" fails a placebo test. Pure BS.

[anawan @anawan1]

@algojan The pattern is clear - optimize via lookahead in the basic tradingview window, post that beautiful curve and hope for suckers. I have quite a bit of 1m 10 year data (equities/options/some futures). Happy to backtest for folks.

[anawan @anawan1]

@Tejanops @PropQuantX I did a 10 year and it failed miserably other than the tradingview window. Classic case of curve-fitting. Still fun to get access and check it out!

[anawan @anawan1]

@PatternProfits Definitely interested!

[anawan @anawan1]

@EdwardAlerts $SPX

[Marceline Yara @AI_creator_Yara]

A former Goldman Sachs quant trader shared a simple but powerful principle with me in a single conversation:

���We don’t forecast. We only take positions when pricing deviates from estimated probability by more than ~6%.”

[ 𝐍𝐨𝐭𝐞: 𝐅𝐨𝐥𝐥𝐨𝐰 𝐌𝐞: @AI_creator_Yara (𝐅𝐨𝐫 𝐢𝐧𝐬𝐭𝐚𝐧𝐭𝐥𝐲 𝐚𝐮𝐭𝐨 𝐃𝐌)]

That was it. No complexity—just execution logic used on institutional desks.

The result was a system that evaluates 400+ markets per hour.

It focuses on contracts in the ~7–19 cent range where implied pricing diverges significantly from modeled probability (roughly 60–90% confidence zones in backtested scenarios).

Over a three-month period (paper + small capital testing phase):

* Starting point: $2,000
* Peak value: $8,191
* 99 executed trades
* Sharpe ratio: ~2.3

A few example signals the system flagged:

* ETH Merge-related event: 72¢ market price vs ~88% modeled probability → +19¢ move
* SOL above $200 breakout scenario: 44¢ vs ~81% model probability → +15¢
* Florida hurricane (Cat 3+ threshold): 81¢ vs ~92% → +7¢
* Wheat above $800: 53¢ vs ~68% → +20¢

Each of these was identified automatically by the scanner based on deviation thresholds and liquidity filters.

When I showed the setup to him recently, he said:

“This is essentially what we run with an 800M AUM team and dozens of engineers.”

The difference is cost.

My current stack:

* Claude: ~$20/month
* VPS: ~$5/month
* Open-source repos: free
* APIs: minimal / free tiers

On top of that, I’ve deployed multiple lightweight agents running continuously to monitor different market clusters.

Performance tracking labels (internal testing):

* velvet_void: +$697
* nano_alpha: +$541
* ratking_eth: +$407
* darkpool_7: +$356

For context, his fund reportedly returned ~19% last year, while this experimental setup has shown significantly higher short-term volatility-adjusted returns in a limited testing window.

Nothing here is magic or prediction-based—just probabilistic pricing, systematic filtering, and execution discipline.

If you want access to the breakdown and setup details:

1. Like & repost
2. Comment " 𝐂𝐥𝐚𝐮𝐝𝐞 "
3. Follow @AI_creator_Yara (so I can DM you)

[anawan @anawan1]

@MrMilkTrading Another angle to try might be to use the pre-2020 data as your in-sample and post 2020 out of sample. Makes you wonder if the reverse happens...