Gemini Deep Stock Research

Task for Gemini Deep Research Max

Ask this to Gemini Deep Research Max called from https://aistudio.google.com

Suppose every 1st day of the month I want to predict whether by the end of the month a stock will go up or down, in the form of score ranging from -100 (full conviction that it will go down OR being suspended) to +100 (full conviction that it will go up AND NOT being suspended).

What publicly available data must be gathered by Gemini Deep Research so that it can predict the price movement with high accuracy? Think about top 100 factors. Focus on Indonesian market. List all of them sorted from the most important.

IMPORTANT: we’re not building a new AI/machine learning system, also we won’t use any API or any other tools. Only Gemini Deep Research.

The prompt above will result in a document, we will attach it in the next step

DateAgent NameURL
2026-04-25deep-research-max-preview-04-2026https://docs.google.com/document/d/1EZOg9ESOlPgGOztpCQtVdF8Cb3Dp-CbZBUhW3bqmrYE/edit?usp=sharing
2026-04-24deep-research-preview-04-2026https://docs.google.com/document/d/18_LWHXwISbPPH7T_2b2c4D0p4HkeAQWLDfg6lI2rmrU/edit?usp=sharing

Task for Gemini Deep Research

To save tokens pricing, call this from https://gemini.google.com

Check IDX:AKRA (JK.AKRA)
.
First part: predict whether by the end of the month the stock will go up or down compared to the price of the start of the month, in the form of score ranging from -100 (full conviction that it will go down OR being suspended) to +100 (full conviction that it will go up AND NOT being suspended). Output your conviction before explanation, then explain your reasoning in details. Use your original analysis, do not parrot any other analysts.
.
Second part: output the supporting data, i.e.: everything we found that support our conviction, in table format with these columns: #1-100, Factor Name, Factor Score, Explanation. The 100 factors that build our conviction are listed in this document [attached].

Leave a Comment