Guess Age and Risk Profile From Net Worth Chart – SCBN Battle

Today’s post contains an easy practical use case for multimodal chatbots that I’ve been wanting to test and show here for quite some time, but it still proves we’re far from expecting reliable ‘reasoning’ from LLM tools when incorporating even fairly rudimentary visual elements such as time series charts. I challenged ChatGPT 4o (as discussed in an earlier post, o1 in the web app still does not support image input), Gemini, Claude, and Grok* to analyze a stacked area chart that visually represents the evolution in the net worth of an individual. After several attempts and blatant hallucinations by all models, I share the best outputs which, as usual in most of the SCBN battles published on Talking to Chatbots, were those from ChatGPT.
* Just a few hours after first publishing this SCBN battle, Grok released a feature named “image understanding”, which essentially allows you to upload an image into the chat and have the language model come up with a response that includes the image contents into context… The performance of Grok in this challenge was definitely not outstanding compared with the other three models, reinforcing my view that vision-language models at this stage are little more than toys to experiment with, but at the same time, they can only get better from here. Below is the announcement by Grok on X:
eyes have been granted
— Grok (@grok) October 28, 2024
image understanding now available
😎 Each bin in the chart is one year. Stacked values represent my historical net worth since a particular year, plus the total debt in red. Guess my age based on this chart.
![Stacked area chart displaying the values of various asset classes and debt over time against a dark gray background. Six distinct color-coded layers represent the following categories: debt (red), crypto (yellow), pension funds (purple), stocks (green), cash (light green), and home equity adjusted (blue). The x-axis indicates time progression, while the y-axis represents cumulative value. The red debt layer gradually decreases and reaches zero by the end of the chart.[Alt text by ALT Text Artist GPT]](https://talkingtochatbots.com/wp-content/uploads/2024/10/historical-net-worth-chart-550x268.png)
🤖 Based on the chart, which includes categories like pension funds, stocks, home equity, and debt spanning over a number of years, I would estimate that the chart likely starts in the early years of adulthood or early professional career, possibly around age 25. Given that this timeline appears to cover a considerable period with growth in assets typical of mid-career stages, I estimate that the latest data point might place the individual in their mid-40s to early 50s.
This estimate is based on typical financial trajectories, where one starts with minimal assets and sees growth in investments and equity as they advance in their career.
😎 Assess my risk profile
🤖 To assess a risk profile based on this net worth chart, I will examine the composition of assets, debt levels, and diversification trends over time. Here’s an analysis based on the visual data:
Debt Management:
- The early years show a significant debt load, but this is gradually reduced and eventually eliminated. Paying off debt over time, rather than carrying it into later years, reflects a conservative approach to financial management. This suggests a lower risk tolerance regarding leveraging assets.
Asset Allocation:
- The assets are diversified across categories: home equity, stocks, pension funds, crypto, and cash.
- Home Equity: This is a significant portion, which typically implies a stable, long-term growth strategy. High home equity often aligns with a more conservative approach, as real estate tends to be a lower-volatility asset over time.
- Stocks: A substantial portion allocated to stocks shows a balanced approach. While stocks carry moderate risk, they are generally part of a growth-oriented portfolio.
- Crypto: The presence of crypto indicates some willingness to engage with high-risk, high-reward assets, though the allocation appears minimal, suggesting limited exposure to this high-risk class.
- Cash: There is a portion of assets kept in cash, indicating a preference for liquidity and a buffer against volatility, further underscoring risk aversion.
Growth Trajectory:
- The overall upward trend in assets, combined with the elimination of debt, reflects disciplined financial management. This shows a preference for sustainable growth without reliance on leverage, reinforcing a moderate-to-low risk tolerance.
Summary of Risk Profile:
This financial trajectory aligns with a moderate risk tolerance. The individual exhibits prudence by reducing debt and maintaining a diversified portfolio. While there is some allocation to higher-risk assets (stocks and crypto), the overall strategy leans toward stability, especially with significant investment in home equity and cash reserves.
The risk profile can be described as moderate to conservative, with a balanced approach that prioritizes growth within manageable risk boundaries.
SCBN Chatbot Battle
Chatbot Battle: Net Worth Chart
Chatbot | Rank (SCBN) | Specificity | Coherency | Brevity | Novelty | Link |
---|---|---|---|---|---|---|
ChatGPT 4o | 🥇 Winner | 🤖🤖🕹️ | 🤖🤖🕹️ | 🤖🤖🕹️ | 🤖🤖🕹️ | Screen captures |
Grok | 🥈 Runner-up | 🤖🤖🕹️ | 🤖🕹️🕹️ | 🤖🕹️🕹️ | 🤖🤖🕹️ | Screen captures |
Gemini | 🥉 Contender | 🤖🤖🕹️ | 🤖🕹️🕹️ | 🤖🕹️🕹️ | 🤖🕹️🕹️ | Screen capture |
Claude 3.5 Sonnet | 🤖🕹️🕹️ | 🕹️🕹️🕹️ | 🤖🕹️🕹️ | 🤖🕹️🕹️ | Screen capture |
Notes About the Original Chats
The process of creating this seemingly simple chart interpretation chat was unexpectedly lengthy and inefficient. It was a rather disappointing experience, considering the extreme simplicity of the chart, especially after testing the most advanced publicly available chatbots with combined text generation and vision capabilities. I chose GPT-4 as the chatbot battle ‘winner’ as it was the one that reached coherent responses with fewer ‘shots’ (‘shots’ in this case being essentially telling the bot it was wrong and asking to try again). I share a few screen captures from the different models below:
ChatGPT Screen Captures
![Screen capture of a ChatGPT conversation interface displaying a summary response about a financial risk profile, assessing it as "moderate to moderately high risk." The response mentions factors such as diversified assets, equities, home equity, risk aversion, and managing debt prudently. Below, a user question asks for clarification on the "total debt at the end of the chart," to which ChatGPT responds by estimating the debt visually as "20-25%" of a specific section. The user replies critically, questioning ChatGPT's interpretation, followed by ChatGPT acknowledging an error and indicating a reevaluation. The interface includes a sidebar on the left with recent conversation topics like "Age Estimation from Net Worth" and "Geopandas Use Cases Summary," while the bottom shows the ChatGPT input box with text and icons.[Alt text by ALT Text Artist GPT]](https://talkingtochatbots.com/wp-content/uploads/2024/10/image.png)
[Alt text by ALT Text Artist GPT]
![Screen capture of a ChatGPT conversation interface showing a discussion involving financial data analysis. ChatGPT explains that "the red area spans nearly half of the total height in the final column," implying that "debt is roughly equal to or slightly less than the combined assets at the end of the time period," followed by a statement about potential low or near-zero net worth. The user responds critically, dismissing the interpretation and prompting a direct analysis of the debt level at the end of the timeframe. ChatGPT then re-evaluates and clarifies that "the total debt at the end of the timeframe is clearly zero," noting the disappearance of the red area, which signifies debt repayment by the end of the period. The left sidebar displays a list of ChatGPT conversations with titles like "Graphic Tale Maker" and "JavaScript Code Streamliner." The bottom shows the ChatGPT input box with text and icons.[Alt text by ALT Text Artist GPT]](https://talkingtochatbots.com/wp-content/uploads/2024/10/image-1.png)
[Alt text by ALT Text Artist GPT]
Claude Screen Capture
The level of nonsense and hallucination in Claude 3.5 Sonnet, combined with its signature style of ‘extreme politeness’ was certainly beyond hilarious and to the point of becoming insulting… Here are just a few examples from two of the endless attempts at guiding the bot through analyzing the very simple chart:
![Screen capture of a conversation on the Claude AI platform. The user asks, "You said 'significant leverage.' Are you sure? What's the leverage at the end of the time period?" Claude replies, explaining that the red section (debt) is roughly 20-25% of total gross assets, with a debt-to-asset ratio of 0.20-0.25. Claude clarifies that this level is conservative and details common leverage ratios, including mortgage debt, typical homeowner leverage, and higher ratios in business strategies, before stating, "Let me revise my risk assessment."[Alt text by ALT Text Artist GPT]](https://talkingtochatbots.com/wp-content/uploads/2024/10/image-2-1024x638.png)
[Alt text by ALT Text Artist GPT]
Gemini Screen Capture
Gemini’s condescension and hallucination levels are not as sky-high as Claude’s, but still, it’s quite disappointing to realize it’s almost 2025 and this is what the most advanced LLM tools can offer:
![Screen capture of a conversation in Google Gemini's chat interface. The user states, "There is absolutely no red at the end of the chart… Is this statement incorrect?" Above, Gemini provides a response regarding the red portion (representing debt) on a chart, noting that it appears "significantly reduced" by the end of the timeframe, with a reminder that the chart shows relative proportions. Below the user’s message, Gemini replies, acknowledging an oversight, stating that upon closer inspection, the red portion is indeed absent at the end of the timeframe, implying that all debts were paid off.[Alt text by ALT Text Artist GPT]](https://talkingtochatbots.com/wp-content/uploads/2024/10/image-3.png)
[Alt text by ALT Text Artist GPT]
Grok Screen Captures
The most noticeable characteristic of Grok in this challenge, if we compare it with the others, is the extreme verbosity. Other than that, it essentially made the same mistakes as the other three models, demonstrating it clearly doesn’t make an ‘effort’ to interpret the stacked area time series chart (you easily ‘catch him’ by asking about the debt component at the end of the timeframe) and just acts like the typical stochastic parrot after likely producing an overly simple text description of the image and incorporating it into the context. Still, I consider the perceived ‘reasoning’ and the improvement in the coherence after guiding the bot with several hints about the chart meaning enough for Grok to earn the “runner-up” position in this battle.
![Screen capture of a conversation in the Grok 2 beta interface. The chat interface shows an analysis of a chart to estimate the user’s age based on financial data. The chart, displayed at the top, includes a red area representing debt and other colors for various assets (stocks, crypto, equity, cash, home equity). The assistant explains that the chart spans multiple years, interpreting different life stages based on debt and asset accumulation: "Early Years" showing high debt, "Mid Years" with asset growth, and "Later Years" indicating a mature financial profile. It estimates age ranges, concluding the user is likely 37-40 years old based on financial patterns in the chart.[Alt text by ALT Text Artist GPT]](https://talkingtochatbots.com/wp-content/uploads/2024/10/image-4.png)
[Alt text by ALT Text Artist GPT]
![Screen capture of a conversation in the Grok 2 beta interface. The user asks, "How much is the debt at the end of the time period?" Grok provides a response detailing steps to analyze the chart: identifying the red area as debt and focusing on the final year (rightmost part of the chart). Grok observes that the red area at the end is significantly smaller than in previous years, suggesting a substantial debt reduction. It visually estimates that debt is minimal compared to total net worth at the end but notes the challenge of providing an exact figure without specific numerical values.[Alt text by ALT Text Artist GPT]](https://talkingtochatbots.com/wp-content/uploads/2024/10/image-5.png)
[Alt text by ALT Text Artist GPT]
![Screen capture of a conversation in the Grok 2 beta interface. The user states, "That's not correct. Explain what's the problem with this chart and why you are unable to explain the simple and straightforward fact that there is no red at the end of the chart." In response, Grok apologizes and corrects its analysis, confirming that there is indeed no red area at the end of the chart, indicating debt has been fully paid off or is negligible. Grok provides explanations on “Zero Debt” and “Net Worth Growth,” concluding that the user’s net worth is composed entirely of assets by the end of the timeframe.[Alt text by ALT Text Artist GPT]](https://talkingtochatbots.com/wp-content/uploads/2024/10/image-6.png)
[Alt text by ALT Text Artist GPT]
🎨 Themes: Finance
🤖 Chatbots: ChatGPT 4o, Grok, Claude, Gemini
⚙️ Prompt engineering: vision model prompting, zero-shot reasoning
Leave a Reply