Yoav Shkedy

Playing with software and trading

[Yoav Shkedy @Yoav_Shkedy]

@cryptorover 🤔

https://t.co/snC6031LM0

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

@KobeissiLetter Who is accumulating?

https://t.co/snC6031LM0

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

@cryptorover 👀

https://t.co/snC6031LM0

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

@TrendSpider 👀

https://t.co/9JnRg1dxns

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

@Sylvia_xh6 Are we in accumulation phase?

https://t.co/9JnRg1dxns

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

@BitcoinArchive We at least want the next candle to close above it, but might be

https://t.co/9JnRg1dxns

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

@Mr_Derivatives Accumulation phase starting?

https://t.co/9JnRg1dxns

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

@brettmacro Accumulating? 👀

https://t.co/9JnRg1dxns

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

@amitisinvesting https://t.co/9JnRg1dxns

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

Have we entered Bitcoin’s accumulation portion of the cycle?

bitcoin:native

[Yoav Shkedy @Yoav_Shkedy]

This is where Jev could get genuinely interesting for trading.

A lot of strategies want news-aware rules, but they’re painful to backtest and execute.

Example: your strategy trades NQ daily, but you want to stay out on days where Trump says something bad about Iran.

Historically, that means feeding huge amounts of news into an LLM or a more traditional classifier and asking:
“Should this strategy be active today?”

That’s slow and expensive to backtest at scale - and cumbersome for real-time execution too.

Jev could make this much cleaner and cost-effective:

news → structured probability/decision → strategy on/off

Run that across years of historical data, then use the same logic live.

That opens up a much bigger design space for strategies.

[Yoav Shkedy @Yoav_Shkedy]

@obrainarc @arc This is a better use of Jev in trading:

https://t.co/frTD4A1rrS

[Yoav Shkedy @Yoav_Shkedy]

This is where Jev could get genuinely interesting for trading.

A lot of strategies want news-aware rules, but they’re painful to backtest and execute.

Example: your strategy trades NQ daily, but you want to stay out on days where Trump says something bad about Iran.

Historically, that means feeding huge amounts of news into an LLM or a more traditional classifier and asking:
“Should this strategy be active today?”

That’s slow and expensive to backtest at scale - and cumbersome for real-time execution too.

Jev could make this much cleaner and cost-effective:

news → structured probability/decision → strategy on/off

Run that across years of historical data, then use the same logic live.

That opens up a much bigger design space for strategies.

[Yoav Shkedy @Yoav_Shkedy]

@eternitybro Try this instead

https://t.co/frTD4A1rrS

[Yoav Shkedy @Yoav_Shkedy]

This is where Jev could get genuinely interesting for trading.

A lot of strategies want news-aware rules, but they’re painful to backtest and execute.

Example: your strategy trades NQ daily, but you want to stay out on days where Trump says something bad about Iran.

Historically, that means feeding huge amounts of news into an LLM or a more traditional classifier and asking:
“Should this strategy be active today?”

That’s slow and expensive to backtest at scale - and cumbersome for real-time execution too.

Jev could make this much cleaner and cost-effective:

news → structured probability/decision → strategy on/off

Run that across years of historical data, then use the same logic live.

That opens up a much bigger design space for strategies.

[Yoav Shkedy @Yoav_Shkedy]

@mynameisroti Yes, as opposed to feeding it market state and ask for BUY/SELL probabilities

[Yoav Shkedy @Yoav_Shkedy]

This is where Jev could get genuinely interesting for trading.

A lot of strategies want news-aware rules, but they’re painful to backtest and execute.

Example: your strategy trades NQ daily, but you want to stay out on days where Trump says something bad about Iran.

Historically, that means feeding huge amounts of news into an LLM or a more traditional classifier and asking:
“Should this strategy be active today?”

That’s slow and expensive to backtest at scale - and cumbersome for real-time execution too.

Jev could make this much cleaner and cost-effective:

news → structured probability/decision → strategy on/off

Run that across years of historical data, then use the same logic live.

That opens up a much bigger design space for strategies.

[Yoav Shkedy @Yoav_Shkedy]

Faster execution of a zero-edge signal just loses money faster.

Feed it price, volume, order book, news, positions - and you’ve built a very fast decision loop. You have not built alpha.

Jev could still be extremely useful as part of a trading system:

Think of a meta-model instead of an alpha model. Imagine five specialized strategies/models producing signals. Jev sees regime, volatility, their predictions, confidence, exposure, recent degradation, and market context, and decides which one to activate or how much risk to allocate.

The edge doesn’t come from Jev alone.

It comes from the data, feature construction, target definition, validation, execution, and feedback loop around it.

[Yoav Shkedy @Yoav_Shkedy]

@punk6995 Not in the way everyone thinks

https://t.co/frTD4A1rrS

[Yoav Shkedy @Yoav_Shkedy]

This is where Jev could get genuinely interesting for trading.

A lot of strategies want news-aware rules, but they’re painful to backtest and execute.

Example: your strategy trades NQ daily, but you want to stay out on days where Trump says something bad about Iran.

Historically, that means feeding huge amounts of news into an LLM or a more traditional classifier and asking:
“Should this strategy be active today?”

That’s slow and expensive to backtest at scale - and cumbersome for real-time execution too.

Jev could make this much cleaner and cost-effective:

news → structured probability/decision → strategy on/off

Run that across years of historical data, then use the same logic live.

That opens up a much bigger design space for strategies.

[Yoav Shkedy @Yoav_Shkedy]

@Joshuwa Right. Try this instead:

https://t.co/frTD4A1rrS

[Yoav Shkedy @Yoav_Shkedy]

This is where Jev could get genuinely interesting for trading.

A lot of strategies want news-aware rules, but they’re painful to backtest and execute.

Example: your strategy trades NQ daily, but you want to stay out on days where Trump says something bad about Iran.

Historically, that means feeding huge amounts of news into an LLM or a more traditional classifier and asking:
“Should this strategy be active today?”

That’s slow and expensive to backtest at scale - and cumbersome for real-time execution too.

Jev could make this much cleaner and cost-effective:

news → structured probability/decision → strategy on/off

Run that across years of historical data, then use the same logic live.

That opens up a much bigger design space for strategies.

[Yoav Shkedy @Yoav_Shkedy]

@prpatel05 A better idea instead:

https://t.co/frTD4A1rrS

[Yoav Shkedy @Yoav_Shkedy]

This is where Jev could get genuinely interesting for trading.

A lot of strategies want news-aware rules, but they’re painful to backtest and execute.

Example: your strategy trades NQ daily, but you want to stay out on days where Trump says something bad about Iran.

Historically, that means feeding huge amounts of news into an LLM or a more traditional classifier and asking:
“Should this strategy be active today?”

That’s slow and expensive to backtest at scale - and cumbersome for real-time execution too.

Jev could make this much cleaner and cost-effective:

news → structured probability/decision → strategy on/off

Run that across years of historical data, then use the same logic live.

That opens up a much bigger design space for strategies.

[Yoav Shkedy @Yoav_Shkedy]

@RoundtableSpace Try this instead

https://t.co/frTD4A1rrS

[Yoav Shkedy @Yoav_Shkedy]

This is where Jev could get genuinely interesting for trading.

A lot of strategies want news-aware rules, but they’re painful to backtest and execute.

Example: your strategy trades NQ daily, but you want to stay out on days where Trump says something bad about Iran.

Historically, that means feeding huge amounts of news into an LLM or a more traditional classifier and asking:
“Should this strategy be active today?”

That’s slow and expensive to backtest at scale - and cumbersome for real-time execution too.

Jev could make this much cleaner and cost-effective:

news → structured probability/decision → strategy on/off

Run that across years of historical data, then use the same logic live.

That opens up a much bigger design space for strategies.

[Yoav Shkedy @Yoav_Shkedy]

@ibocodes Try this

https://t.co/frTD4A1rrS

[Yoav Shkedy @Yoav_Shkedy]

This is where Jev could get genuinely interesting for trading.

A lot of strategies want news-aware rules, but they’re painful to backtest and execute.

Example: your strategy trades NQ daily, but you want to stay out on days where Trump says something bad about Iran.

Historically, that means feeding huge amounts of news into an LLM or a more traditional classifier and asking:
“Should this strategy be active today?”

That’s slow and expensive to backtest at scale - and cumbersome for real-time execution too.

Jev could make this much cleaner and cost-effective:

news → structured probability/decision → strategy on/off

Run that across years of historical data, then use the same logic live.

That opens up a much bigger design space for strategies.