{"id":242242,"date":"2026-10-21T15:04:49","date_gmt":"2026-10-21T15:04:49","guid":{"rendered":"https:\/\/mycryptomania.com\/?p=242242"},"modified":"2026-10-21T15:04:49","modified_gmt":"2026-10-21T15:04:49","slug":"a-stock-fell-20-i-used-claude-to-investigate-why","status":"publish","type":"post","link":"https:\/\/mycryptomania.com\/?p=242242","title":{"rendered":"A Stock Fell 20%. I Used Claude to Investigate Why"},"content":{"rendered":"

A real case of Claude stock analysis: Rambus fell 21% after earnings. I pulled prices, news and EPS data from EODHD to separate facts from explanations.<\/em><\/p>\n

When a stock drops 20% in one day, the internet gives you an answer within the\u00a0hour.<\/p>\n

The problem is that it gives you five of them. And they don\u2019t\u00a0agree.<\/p>\n

On April 28, 2026, Rambus (RMBS) closed at $111.27. One session earlier it was at $141.31. Two sessions earlier, it was sitting at an all-time high of\u00a0$158.40.<\/p>\n

By the afternoon, the explanations were everywhere: earnings miss, analyst downgrade, AI bubble, OpenAI, memory shortage, \u201csell the\u00a0news.\u201d<\/p>\n

So I ran an experiment. Instead of picking the story I liked best, I used Claude stock analysis on top of raw market data to answer one question:<\/p>\n

What do we actually know, and what are we just guessing?<\/strong><\/p>\n

If you\u2019re:<\/p>\n

building AI agents for investment research,writing a newsletter that explains market\u00a0moves,or just holding a stock that fell off a\u00a0cliff,<\/p>\n

this is the workflow worth stealing.<\/p>\n

Why \u201cWhy Is This Stock Down Today?\u201d Is the Wrong First\u00a0Question<\/h3>\n

Most people start with the news. That\u2019s backwards.<\/p>\n

News is written after<\/em> the price moves. Journalists see a red candle, call a few sources, look for a catalyst and publish. The headline is a hypothesis dressed up as a\u00a0fact.<\/p>\n

Here\u2019s what happened when I pulled the RMBS news feed from EODHD for April 20 to May\u00a05:<\/p>\n

Look at the\u00a0shape.<\/p>\n

One article a day before the drop. Seven on earnings day. Eighteen on the crash day. Six the day\u00a0after.<\/p>\n

The news didn\u2019t cause the move. The move caused the\u00a0news.<\/p>\n

And notice the numbers above the bars. EODHD\u2019s normalized sentiment score for RMBS stayed positive the entire time, including +0.81 on the day the stock lost a fifth of its value<\/strong>. Most of the articles described a company with solid results, growing product revenue and strong cash generation.<\/p>\n

Sentiment said \u201cfine.\u201d Price said \u201cget\u00a0out.\u201d<\/p>\n

That gap is where the real investigation starts.<\/p>\n

The Reframe: Separate Facts From Explanations<\/h3>\n

An investigation into a price drop needs three\u00a0buckets:<\/p>\n

Facts.<\/strong> Things the data proves. Prices, dates, reported numbers, documented events.Plausible explanations.<\/strong> Hypotheses consistent with the facts. Each one should come with what would confirm or refute\u00a0it.Noise.<\/strong> Claims that don\u2019t explain the move, or that contradict the\u00a0data.<\/p>\n

Humans are bad at keeping these apart. We read a convincing headline and promote it straight to bucket\u00a0one.<\/p>\n

Claude is good at this, if you give it the raw data and force the structure<\/em>. Without data, an LLM will happily write you a confident story. With data, it becomes a very fast, very patient analyst that checks every claim against the\u00a0numbers.<\/p>\n

The model is not the\u00a0edge.<\/p>\n

The data layer\u00a0is.<\/p>\n

AI Stock Analysis Needs a Data Layer\u00a0First<\/h3>\n

For this investigation I connected Claude to EODHD through its MCP server, so it could pull data directly during the conversation. Four datasets did all the\u00a0work:<\/p>\n

End-of-day prices<\/strong> for RMBS and a sector benchmark (SMH, the VanEck Semiconductor ETF)Earnings calendar<\/strong> with actual EPS, consensus estimates and surprise\u00a0%Financial news API<\/strong> with full article text tagged by\u00a0tickerSentiment API<\/strong> with a daily normalized score and article\u00a0count<\/p>\n

That\u2019s it. No terminal, no scraping, no 15 browser\u00a0tabs.<\/p>\n

Want to run the same investigation on your own watchlist?<\/strong>
EODHD gives you prices, fundamentals, earnings, news and sentiment from one API (and an MCP server for Claude).
\u2192 <\/strong>Try EODHD\u00a0APIs<\/strong><\/a><\/p>\n

Step 1: Measure the Move Against the\u00a0Sector<\/h3>\n

A 20% drop means nothing on its own. If the whole sector fell 18%, Rambus didn\u2019t have a problem. The market\u00a0did.<\/p>\n

So the first thing I asked Claude to do was index RMBS against\u00a0SMH.<\/p>\n

This chart changed the whole\u00a0story.<\/p>\n

From March 27 to April 24, Rambus went up 76.5%<\/strong>. The semiconductor ETF went up 35.3%<\/strong>. From the March 30 low, RMBS had almost doubled (+98.7%) in four\u00a0weeks.<\/p>\n

The fall only makes sense next to the\u00a0rise.<\/p>\n

Then I zoomed in on daily\u00a0returns:<\/p>\n

Two things jump\u00a0out:<\/p>\n

Friday, April 24:<\/strong> RMBS +14.4% in a single day, ahead of earnings. The sector was up\u00a05.1%.Monday, April 27:<\/strong> RMBS fell 10.8% during regular trading<\/strong>. The earnings report came out after<\/em> the close. The sector was flat (-0.04%).<\/p>\n

That second point is the one most headlines skipped. A big chunk of the drop happened before anyone saw the\u00a0numbers.<\/p>\n

Peak to trough over two sessions: -29.8%<\/strong>. Against the sector, Rambus underperformed by 10.8 points on Monday and 18.3 points on\u00a0Tuesday.<\/p>\n

Step 2: Check What the Earnings Actually\u00a0Said<\/h3>\n

Here\u2019s what Rambus reported for Q1 2026 (after the close on April\u00a027):<\/p>\n

Management said results were \u201cin line with guidance.\u201d Q2 guidance implied roughly 11% sequential product revenue\u00a0growth.<\/p>\n

That doesn\u2019t read like a disaster.<\/p>\n

So I asked Claude to compare the numbers against consensus. And this is where it got interesting.<\/p>\n

The sources didn\u2019t agree on what \u201cconsensus\u201d was.<\/strong><\/p>\n

EODHD\u2019s earnings calendar: estimate $0.64<\/strong>, actual $0.63. A miss of 1 cent\u00a0(-1.6%).One widely syndicated analysis: estimate $0.62<\/strong>. A\u00a0beat.A Zacks headline the same evening: \u201cQ1 Earnings and Revenues Beat Estimates.\u201d<\/p>\n

Same quarter. Same company. Miss or beat depending on who you\u00a0read.<\/p>\n

Claude flagged a second conflict on valuation: one article said RMBS was trading at roughly 48x<\/strong> this year\u2019s earnings estimates, another said 63x forward earnings<\/strong>. Different estimate sets, different denominators.<\/p>\n

And a third one, in a pre-earnings preview: it claimed Rambus\u2019s product revenue guidance of $84M to $90M implied a \u201c46 to 50 percent decline\u201d from $166.7M a year earlier. But $166.7M was total<\/strong> revenue, not product revenue. Product revenue actually grew 15%. Claude caught the apples-to-oranges comparison in seconds because it had the reported figures right next to the\u00a0claim.<\/p>\n

This is the most useful thing an AI does in stock analysis. Not predicting. Cross-checking.<\/strong><\/p>\n

The 1-cent difference can\u2019t explain a 21% drop. Whatever happened, it wasn\u2019t the EPS\u00a0number.<\/p>\n

Step 3: Let Claude Sort the\u00a0Evidence<\/h3>\n

With prices, earnings, news and sentiment loaded, I gave Claude one instruction: classify everything into facts, plausible explanations and noise. Here\u2019s a condensed version of the\u00a0output.<\/p>\n

Facts (the data proves\u00a0it)<\/h3>\n

RMBS rose 76.5% in four weeks into earnings, more than twice the semiconductor ETF.It fell 10.8% on April 27 before<\/strong> the report, with the sector\u00a0flat.It fell another 21.3% on April 28, while SMH fell 3.0% and the Nasdaq 100 fell\u00a01.0%.Q1 EPS was within one cent of every consensus figure published.Baird downgraded RMBS from Outperform to Neutral on April 28, with a $120 price target. Their reasoning: growing DRAM shortages could slow RDIMM unit growth into 2027, and Rambus is \u201ca unit-driven top-line\u201d without the benefit of higher\u00a0pricing.Late on April 27, a Wall Street Journal report said OpenAI had missed internal user and revenue goals. On April 28 the PHLX Semiconductor Index fell more than 3% intraday, after an 18-day winning\u00a0streak.<\/p>\n

Plausible explanations (ranked)<\/h3>\n

1. Positioning: buy the rumor, sell the news.<\/strong>
The stock priced in perfection. A +14% day right before earnings means buyers were paying for a blowout, not a solid quarter. \u201cIn line\u201d was a disappointment relative to price.
Would confirm:<\/em> other high-momentum chip names falling hard on the same day without bad results.
Would refute:<\/em> RMBS falling while its peers\u00a0rallied.<\/p>\n

2. A forward-looking supply risk, made visible by the Baird note.<\/strong>
Management itself said back-end supply tightness would persist into 2027 and that the MRDIMM ramp depends on Intel and AMD platform timing. The downgrade gave the market a clean narrative for something already in the call.
Would confirm:<\/em> weakness in later quarters tied to units, not demand.
Would refute:<\/em> product revenue accelerating in Q2 and\u00a0Q3.<\/p>\n

3. Sector de-risking after the OpenAI report.<\/strong>
Yahoo Finance\u2019s markets desk noted the selloff hit \u201cthe names that had already run the hardest\u201d: Rambus, Arm, FormFactor, Navitas, Wolfspeed.
Would confirm:<\/em> the same high-beta group underperforming together.
Would refute:<\/em> RMBS being the only one\u00a0down.<\/p>\n

Noise<\/h3>\n

The EPS \u201cmiss.\u201d<\/strong> One cent, and contested by other\u00a0sources.The new CFO appointment<\/strong> on April 29. It came after the\u00a0drop.\u201cCanary in the coal mine for AI\u201d takes.<\/strong> Possibly right in the long run, but not testable with this\u00a0data.<\/p>\n

Notice what Claude didn\u2019t do: pick a single cause. A 21% drop is almost never one thing. It\u2019s a crowded trade, a reason to sell and a sector mood all arriving at the same\u00a0time.<\/p>\n

Why Do Stocks Fall After Good Earnings?<\/h3>\n

Because the price already\u00a0knew.<\/p>\n

Earnings move stocks relative to expectations<\/em>, and expectations live in the price, not in the consensus EPS. When a stock rallies 76% in four weeks, the real bar isn\u2019t $0.64. It\u2019s whatever number justifies the\u00a0rally.<\/p>\n

The best evidence for this came three months\u00a0later.<\/p>\n

On July 27, 2026, Rambus reported Q2 EPS of $0.64 against a $0.58 estimate<\/strong>. A clean 10.3%\u00a0beat<\/strong>.<\/p>\n

The stock fell 9.0%<\/strong> the next day and closed at $82.81 two sessions\u00a0later.<\/p>\n

A tiny miss crashed it. A real beat didn\u2019t save\u00a0it.<\/p>\n

If you only read EPS surprises, this makes no sense. If you read price, positioning and guidance together, it\u00a0does.<\/p>\n

How to Reproduce This Claude Stock Analysis With\u00a0Python<\/h3>\n

I ran the investigation interactively, with Claude connected to EODHD\u2019s MCP server. Here\u2019s the reproducible version as a script, using EODHD\u2019s REST API and the Anthropic SDK.<\/p>\n

pip install requests pandas anthropicimport os, json
import requests
import pandas as pd
import anthropicEODHD_KEY = os.getenv(“EODHD_API_KEY”)
BASE = “https:\/\/eodhd.com\/api”
TICKER, BENCH = “RMBS.US”, “SMH.US”
START, END = “2026-04-17”, “2026-05-01″def get(path, **params):
params.update({“api_token”: EODHD_KEY, “fmt”: “json”})
r = requests.get(f”{BASE}\/{path}”, params=params, timeout=30)
r.raise_for_status()
return r.json()# 1. Prices: stock vs sector
stock = pd.DataFrame(get(f”eod\/{TICKER}”, **{“from”: START, “to”: END})).set_index(“date”)
bench = pd.DataFrame(get(f”eod\/{BENCH}”, **{“from”: START, “to”: END})).set_index(“date”)
rets = pd.DataFrame({
“stock”: stock[“close”].pct_change() * 100,
“sector”: bench[“close”].pct_change() * 100,
}).dropna().round(2)
rets[“excess”] = rets[“stock”] – rets[“sector”]# 2. Earnings, 3. news, 4. sentiment
earnings = get(“calendar\/earnings”, symbols=TICKER, **{“from”: “2025-01-01”, “to”: END})
news = get(“news”, s=TICKER, limit=50, **{“from”: START, “to”: END})
sentiment = get(“sentiments”, s=TICKER, **{“from”: START, “to”: END})print(rets)<\/p>\n

Output:<\/p>\n

stock sector excess
date
2026-04-24 14.37 5.10 9.27
2026-04-27 -10.79 -0.04 -10.75
2026-04-28 -21.26 -2.97 -18.29
2026-04-29 0.80 1.70 -0.90
2026-04-30 2.63 1.43 1.20
2026-05-01 -2.76 0.61 -3.37<\/p>\n

That table alone kills half the headlines. Now hand the evidence to Claude, with a prompt that forces the structure:<\/p>\n

evidence = {
“daily_returns_vs_sector”: rets.to_dict(orient=”index”),
“earnings_history”: earnings[“earnings”][-4:],
“sentiment”: sentiment.get(TICKER, []),
“news”: [{“date”: n[“date”][:16], “title”: n[“title”], “excerpt”: n[“content”][:600]}
for n in news],
}PROMPT = “””You are an equity research assistant. Explain this price move using ONLY the evidence.
Return three sections:
1. FACTS: what the data proves. Cite the field you used.
2. PLAUSIBLE EXPLANATIONS: ranked hypotheses. For each, say what would confirm or refute it.
3. NOISE: claims that don’t explain the move or contradict the data.
Flag any number that appears with different values in different sources.
Do not give investment advice.”””client = anthropic.Anthropic()
msg = client.messages.create(
model=”claude-sonnet-4-5″, # use the latest Claude model on your account
max_tokens=2500,
messages=[{“role”: “user”, “content”: PROMPT + “nnEVIDENCE:n” + json.dumps(evidence, default=str)}],
)
print(msg.content[0].text)<\/p>\n

From here you can\u00a0build:<\/p>\n

a daily \u201cbiggest movers explained\u201d report for your watchlist,an alert that fires when a stock underperforms its sector by more than 10\u00a0points,a research agent that writes the first draft of your post-earnings notes.<\/p>\n

Key Takeaways<\/h3>\n

Start with price, not news.<\/strong> Index the stock against its sector before reading a single headline.Check the timeline.<\/strong> 10.8% of Rambus\u2019s drop happened before the earnings report was\u00a0public.Use Claude as a cross-checker, not an oracle.<\/strong> Its biggest value here was catching conflicting numbers across sources, not predicting anything.<\/p>\n

Final Thoughts<\/h3>\n

Rambus didn\u2019t fall 21% because it missed by a\u00a0penny.<\/p>\n

It fell because the price had run far ahead of the story, and the story stopped getting\u00a0better.<\/p>\n

The internet will always give you an explanation within an hour. A good Claude stock analysis workflow gives you something better: the evidence to know which explanations deserve your\u00a0trust.<\/p>\n

FAQs<\/h3>\n

\u2753 Can Claude analyze stocks?<\/strong>
\u2705 Yes, but only as well as the data you give it. Connected to a market data source like EODHD (via API or MCP server), Claude can pull prices, earnings, news and sentiment, then cross-check claims against the numbers. Without data, it can only repeat general knowledge.<\/p>\n

\u2753 Why is a stock down today when the news looks positive?<\/strong>
\u2705 Price reacts to expectations, not to whether news is good in absolute terms. If a stock rallied hard before the news, \u201cgood\u201d may already be priced in. Compare the move against its sector and check what happened in the days\u00a0before.<\/p>\n

\u2753 Why do stocks fall after good earnings?<\/strong>
\u2705 Usually because of positioning and guidance. A strong run before the report raises the real bar above published consensus. Rambus beat Q2 2026 EPS estimates by 10% and still fell 9% the next\u00a0day.<\/p>\n

\u2753 What does \u201cbuy the rumor, sell the news\u201d mean?<\/strong>
\u2705 Traders buy ahead of an expected positive event and sell once it happens, locking in gains. It explains why a stock can drop on results that meet or beat expectations.<\/p>\n

\u2753 Is AI stock analysis reliable?<\/strong>
\u2705 It\u2019s reliable for organizing evidence, spotting contradictions and summarizing large volumes of news. It\u2019s not reliable for predicting prices. Treat it as a fast research assistant, and always verify the facts it\u00a0cites.<\/p>\n

Investigate the next big mover yourself<\/strong>
Prices, earnings calendar, financial news and sentiment in one API, plus an MCP server that plugs straight into Claude.
\u2192 <\/strong>Get started with\u00a0EODHD<\/strong><\/a>Need content like this for your API or fintech product?<\/strong>
I write data-driven technical articles that explain products through real use cases, not marketing fluff.
\u2192 <\/strong>Work with\u00a0me<\/strong><\/a><\/p>\n

Looking for technical content for your company? I can help\u200a\u2014\u200a<\/em>LinkedIn<\/em><\/a> \u00b7 <\/em>kevinmenesesgonzalez@gmail.com<\/em><\/a><\/p>\n

A Stock Fell 20%. I Used Claude to Investigate Why<\/a> was originally published in Coinmonks<\/a> on Medium, where people are continuing the conversation by highlighting and responding to this story.<\/p>","protected":false},"excerpt":{"rendered":"

A real case of Claude stock analysis: Rambus fell 21% after earnings. I pulled prices, news and EPS data from EODHD to separate facts from explanations. When a stock drops 20% in one day, the internet gives you an answer within the\u00a0hour. The problem is that it gives you five of them. And they don\u2019t\u00a0agree. […]<\/p>\n","protected":false},"author":0,"featured_media":242243,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[],"class_list":["post-242242","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-interesting"],"_links":{"self":[{"href":"https:\/\/mycryptomania.com\/index.php?rest_route=\/wp\/v2\/posts\/242242"}],"collection":[{"href":"https:\/\/mycryptomania.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/mycryptomania.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/mycryptomania.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=242242"}],"version-history":[{"count":0,"href":"https:\/\/mycryptomania.com\/index.php?rest_route=\/wp\/v2\/posts\/242242\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/mycryptomania.com\/index.php?rest_route=\/wp\/v2\/media\/242243"}],"wp:attachment":[{"href":"https:\/\/mycryptomania.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=242242"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/mycryptomania.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=242242"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/mycryptomania.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=242242"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}