Key issues confronting anyone reviewing a ticker’s merits as an investment, particularly one they are unfamiliar with:
- Price is truth, so you view price charts – but for what period, with what technical studies, etc.?
- The company’s strategic narrative and how it evolved over time is key – but what sources should you consult?
Vector Sets systematically select and frame the price history that matters most, given the most recent price.
Vector Sets are the channels that VecViz systematically identifies from all combinations of tops and bottoms and then ranks according to their Vector Strength, a measure of support and resistance with linkages to a broader body of academic and industry research. 1
With names that reflect their Vector Strength ranking (“vs1” is the Vector Set with the greatest Vector Strength, “vs2” is the Vector Set with the second greatest Vector Strength, etc.), Vector Sets are an objective basis for selecting the range of price history that should be reviewed, and also of identifying the most important trends within it.
If you have reviewed vs1 through vs10 you can be fairly comfortable that you have a strong appreciation of the ticker’s price history, and where the current ticker price resides within it and in relation to the trends acting upon it that are currently most relevant. That said, lesser Vector Sets (vs35+) can provide valuable perspective too. The trends, strategies, and competitive landscapes they represent, whose relevance has diminished, can serve as conceptual boundaries separating reasonable prices from unreasonable prices.
Vector Sets can be viewed with the click of a mouse in the Vector Set Study area of VecViz’s dashboards, as demonstrated in this video:
VecEvents harvested from a leading AI LLM provide independent, objective narrative context.
VecEvents are briefly stated key events or themes that exerted significant influence upon a ticker for an identified time period.
We have developed an iterative prompting process to generate VecEvents and associated influence time periods for individual tickers from Google’s Gemini 2.52, one of the major LLM’s . We link these VecEvents to Vector Sets on the basis of the influence time period enveloping one or more of the tops or bottoms anchoring the Vector Set (aka the Vector Set’s “VecDates”). Note that at present these VecEvents reflect an information window up through January 2025. We intend to update the VecEvents as more recent information windows become available.
For the benefit of those unfamiliar with the ticker’s business the prompt also generates characterizations of “Bullish”/ “Bearish” / “Neutral” with regard to the influence of the VecEvent upon the ticker, and whether that influence has recently shown signs of “Intensifying” or “Waning” or remains “Steady”.
VecEvents can be viewed by hovering your mouse over a Vector Set in the Vector Set study area or in the Vector Strength Histogram. Note that Vector Sets whose VecDates are in the distant past you will include fewer and much less recent VecEvents than those with VecDates that are more recent.
Framing a ticker’s price positioning within a Vector Set in the context of linked VecEvents and their evolution can accelerate assessment of price attractiveness.
If all the VecEvents linked to a Vector Set are “Bearish” and “Intensifying” should price be in the upper echelons of the Vector Set3? Probably not, unless the final VecDate was in the distant past and new bullish VecEvents have since emerged.
However, if the Vector Set includes recent VecDates, and all the VecEvents are Bearish and intensifying, and you agree with their veracity and characterization, then you are probably pretty confident that price should not be in the upper echelons of that Vector Set.
If your early conclusion is bearish, you can find Vector Sets to challenge your conclusion by filtering out all Vectors except -1.236 and using your mouse to get the VecEvents of the Vectors above the current price. See the video below for an illustration. Is it reasonable that price should be entirely below those Vector Sets, given their VecEvents?
How can one “double check” the VecEvents listed and their characterization?
AI is vulnerable to hallucination, though we try to mitigate much of that with our prompting process. Nevertheless, you should confirm the VecEvents furnished, especially those that are critical for whatever thesis you are forming.
Ideally, you work with an equity research analyst who covers the ticker you are interested in and you can ask them to review the VecEvents you are unsure about4.
If that is not the case, and you don’t have time to go to primary documents yourself to research the issue, consider taking a snippet of the VecDates and VecEvents and feed it into another LLM for verification. We did that with vs1 from the example using AAPL, depicted in the video above, along with the following prompt:
“Do you agree that the VecEvents listed in the image are the primary drivers of AAPL’s stock price during the period enveloped by the VecDates, and do you agree with the directional characterization of their influence (Bullish / Bearish/ Neutral) and the strength of their influence as of 5/6/2025 (Intensifying/ Steady/ Waning)?”
We received a good quality, fairly extensive response from Grok 3, the conclusion of which, as quoted below. indicates that the VecEvents listed and their characterization were reasonably good quality5:
“I mostly agree that the listed VecEvents were primary drivers of AAPL’s stock price during Sep 2012-Aug 2022, except for the Vision Pro launch, which was post-period and niche. The directional characterizations (Bullish/Bearish/Neutral) and strength (Intensifying/Steady/Waning) align well with the period and hold as of May 6, 2025, with minor adjustments: AI and regulatory pressures remain intensifying, while earlier events (iPhone 6, COVID surge) have fully waned. However, missing events like the 5G cycle and trade war should also be considered for a complete picture.”
Obviously, further review by humans is most preferable, and may even be mandatory in certain institutional settings. However using a separately trained model for review is likely to catch at least some hallucinations, and omissions. Just take care not to violate the terms of service of any LLM you use.
We expect to improve upon the LLM sourced VecEvents and perhaps also build further upon them in the months ahead.
- Vector Strength is a measure of the support and resistance a channel is likely to exert on forward price. As discussed in a prior post, Vector Strength has strong correspondence to related academic and industry research. ↩︎
- Please note that VecEvents are applied only to individual company tickers in our coverage, not the ETF’s or the closed end fund. VecEvents for those tickers are still provided manually ↩︎
- Not at all saying that is the case for the company featured in the videos included in this blog post. ↩︎
- While you are at it, the LLM sourced VecEvents in the context of the Vector Sets could help you play devil’s advocate with them with regard to their opinion on the ticker more broadly. ↩︎
- In fairness to Gemini 2.5, we do expand the period enveloped by the VecDates it supplied for each VecEvent a bit to allow for rumors and “chatter” in the market, so it could be that the Vision Pro wouldn’t have been tagged to vs1’s VecEvents if I had stuck strictly with Gemini 2.5’s output. ↩︎