JTH's Account Talk

[rangerray]

Your move to C Fund was timed just about right!!!


Sent from my iPhone using Tapatalk

 

[JTH]

Your move to C Fund was timed just about right!!!


Sent from my iPhone using Tapatalk

Yea, that was an easy one, I had nowhere to go but up

 

[JTH]

Did you want to start an Autotracker account for this system?

You know I hadn't thought about that, but I think it's more than possible to build a Diagnostic for Intra-Fund Transfers.

Sorta ran into a snag, I can't tell you how many hours I got in, but I'm starting over from scratch.
I was about 80% done, but developed some new methodologies along the way and decided it's just easier to start over and apply the lessons accumulated during the first builds into the new build. This will take longer, but the end result it will be cleaner, with less indicator redundancy, it should process faster, and potentially it should transfer over to other AI engines, so eccentrically I'll own it and not be co-dependent on a particular AI ecosystem.

 

[FAAM]

Gosh - makes me dizzy just looking at your sample on weekly output.

 

[JTH]

You know I hadn't thought about that, but I think it's more than possible to build a Diagnostic for Intra-Fund Transfers.

Sorta ran into a snag, I can't tell you how many hours I got in, but I'm starting over from scratch.
I was about 80% done, but developed some new methodologies along the way and decided it's just easier to start over and apply the lessons accumulated during the first builds into the new build. This will take longer, but the end result it will be cleaner, with less indicator redundancy, it should process faster, and potentially it should transfer over to other AI engines, so eccentrically I'll own it and not be co-dependent on a particular AI ecosystem.

So I did a hard-review on this and have added it to my roadmap. Most of it will be automated, I also think I can get an Intra-updated engine that gets fed data at a 1030 cuttoff.

I've completed 8 of 53 roadmap task, about 15% finished.
Estimated time to completion Light Mode 2-3 weeks, Full Mode 6-8 weeks
___Of course that's just an estimate and every new idea adds a new task....

 

[JTH]

Gosh - makes me dizzy just looking at your sample on weekly output.

Good point, I'm trying to balance the line between too wordy and too much brevity.

Basically I need the output to be within reach of a 3-5-Minute read, but with the Depth of a full comprehensive outlook.
I'm also going to work on consistency, so when an updates arrive, you know where to quickly scan to find information you deem relevant to you.

 

[JTH]

That is one hell of an opening gap. Red-Dashed line is 50% retracement of Feb-High to Apr-Low.

Expectations:

• People in bad positions are going to be looking to get out (I did that already on some leveraged plays)
• People who think they'll be left behind might be looking to get in.
• Smart money already left in Feb, then perhaps got in on the lows, so they can ride it out.
• Personally it's not my setup, I'd like to see how this session ends and if we can get both a close and confirming close above the 50% retracement.

 

[TommyIV]

That is one hell of an opening gap. Red-Dashed line is 50% retracement of Feb-High to Apr-Low.

Expectations:

• People in bad positions are going to be looking to get out (I did that already on some leveraged plays)
• People who think they'll be left behind might be looking to get in.
• Smart money already left in Feb, then perhaps got in on the lows, so they can ride it out.
• Personally it's not my setup, I'd like to see how this session ends and if we can get both a close and confirming close above the 50% retracement.

The S&P 500 is losing steam heading into the trade deadline. The current price (5,399.96) is around 1.3% lower than its morning highs (5,469.69). How much more will it shed?

 

[tsptalk]

Breathe in, breathe out.

 

[JTH]

I've been working a Quant AI system, it will never be finished, but I plan to start a thread and start displaying the results.

So I did a hard-review on this and have added it to my roadmap. Most of it will be automated, I also think I can get an Intra-updated engine that gets fed data at a 1030 cuttoff.

So Wednesday's progress was much slower than anticipated.

Within a Project, when an AI session gets long it bogs down the browser cache. If you create a sub-folder within the project it speeds thing up but it can't read the previous history of other folders and you're AI becomes a complete stranger losing the lessons learned. So much of last night was spent creating a turnover file like you do for a shift-change with a co-worker.

But then I encountered another issue that requires another pre-stage for the project. When you spend a lot of time within a project folder, the AI learns and adapts to your line of thinking and you develop a Human-AI Symbiosis relationship. But as the folder conversation exceeds the cache limits of the browser, the AI gets progressively dummer, making very simple avoidable mistakes. When you switch over to a new folder you can port over a turnover list, but you lose your Human-AI Symbiosis realtionship.

So I need to create a custom AI assistant customized for my work flow and tailored to my skill-set and thinking profile. This combined with a turnover list will help keep the project manageable as it expands.

 

[tsptalk]

Sounds easy.

This is going to be fun to follow.

 

[JTH]

So Wednesday's progress was much slower than anticipated.

Within a Project, when an AI session gets long it bogs down the browser cache. If you create a sub-folder within the project it speeds thing up but it can't read the previous history of other folders and you're AI becomes a complete stranger losing the lessons learned. So much of last night was spent creating a turnover file like you do for a shift-change with a co-worker.

But then I encountered another issue that requires another pre-stage for the project. When you spend a lot of time within a project folder, the AI learns and adapts to your line of thinking and you develop a Human-AI Symbiosis relationship. But as the folder conversation exceeds the cache limits of the browser, the AI gets progressively dummer, making very simple avoidable mistakes. When you switch over to a new folder you can port over a turnover list, but you lose your Human-AI Symbiosis realtionship.

So I need to create a custom AI assistant customized for my work flow and tailored to my skill-set and thinking profile. This combined with a turnover list will help keep the project manageable as it expands.

I'm almost finished with Phase 1. Good progress, the Turnover Files are working, the boot-up sequence gets the database mapped, configured and ready to deploy from scratch. I've configured my AI-Team for Project HAL-63, the team lead is Flo, she manages the commoners for me. There's also a reservist unit for special needs, but those are sub-contractors so I don't have to put them on my health insurance.

I'm, almost done with the core indicators, but have to build a candlestick database, then configure a PDS (Pattern Detection System for all indicators). From there we back-test all patterns then archive patterns prior to 2025, to free up resources for the diagnostic engines. Then I'll work-in position sizing strategies and correlation analysis for matching up ETFs which pair well with the S&P 500's direction.

There is likely to be a TSP Diagnostic Engine, that's Phase IV so it will take some time to get to that point.

Here's the team I've assembled, they are still on probation, they need to prove their worth

Name	Title
Flo	Lead Analyst
Art Cashin	Vice Strategist
Thomas Bulkowski	Pattern Detection Specialist
Michael Howell	Liquidity Modeler
Lyn Alden	Narrative Synthesizer
Jason Goepfert	Sentiment Strategist
Ralph Vince	Risk Manager / Capital Allocator
Russell Napier	Market Historian
Martin Zweig	Breadth Strategist

 

[JTH]

Finally when the next European ask me what I think about DJT, I can reply with "What you think about my Pope?"

That's called "Winning"

 

[JTH]

IFT From 100C to 100G EoB Today