Leaders & Laggards

Sector rotation, scanned nightly

The market is always rotating.Most people notice late.

Every trading night we rank all eleven S&P sectors against the index, then rank the biggest names inside each one against their own sector. You wake up knowing where strength moved — and which names carried it. Then go deeper: a model forecast for every name, and an AI research desk that will write up any US stock on request.

Get started$x.xx/month. Cancel anytime.

01 — What sector rotation is

Money doesn't leave the market. It moves.

When the outlook shifts, capital rotates out of one part of the market and into another — out of utilities and staples, into semiconductors and banks, and back again. The index barely moves. Underneath it, leadership changes completely.

That rotation is visible in the data weeks before it is obvious in the headlines. It shows up as one sector quietly, persistently beating the S&P while another quietly stops.

Sector20-day vs SPY60-day vs SPY
XLK+6.1%+11.4%leading
XLF+2.3%+4.0%leading
XLU−2.4%−5.8%lagging

Illustrative figures.

02 — A rotation you lived through

The AI cycle didn't begin on the news.

By the time AI was the only story on financial television, technology had already been beating the S&P for months — and inside it, the memory and semiconductor names had been beating technology itself. Nothing was hidden. It was in the returns, every night, for anyone comparing sectors to the index and stocks to their sector.

The question was never “is AI real?” It was “which part of the market is already being paid for it?”

That is the only question this product answers. Not what will happen — what is already happening, ranked, every night, before you have to decide anything.

03 — Why we only watch the heavy names

Five hundred tickers is not a shortlist.

Most screeners hand you the whole market and let you sort it. That is how you end up with forty names you do not recognise and no conviction about any of them.

We go the other way. Within each sector we look at the index-heavy names — the largest weights, the ones that actually move the sector — and show you how each is performing against it. A handful of names per sector, each one big enough to matter.

Weight

How much of the sector this name actually is.

Weight change

Which names are gaining share of their sector.

Relative return

Whether it is beating the sector it belongs to.

04 — What you get

A shortlist, and a straight answer about your positioning.

Which sectors are leading

All eleven ranked against SPY over 20, 60 and 90 days — so you see brand-new strength and established strength separately.

Which names inside them are leading

The heavy names in each sector, ranked against that sector. Beating the sector that is beating the market is a short list.

Whether you are in the right rotation

Hold something already? See at a glance whether its sector is still leading or has quietly rolled over.

05 — Predictions

A forecast from the chart alone.

A foundation model trained on billions of candlesticks reads each symbol’s recent price bars and continues the pattern, bar by bar. We run it many times per symbol: the middle path is the forecast, the spread is the range.

Every forecast is scored against what actually happened, and each name’s hit rate sits right beside its forecast — so you can see where the model has earned attention and where it hasn’t.

How a forecast is made
  1. STEP 1

    Reads the chart

    The last 252 daily bars of one symbol — open, high, low, close and volume.

    Price only. No news, earnings or fundamentals.
  2. STEP 2

    Continues the pattern

    Each candle becomes tokens, and a transformer predicts the next bar from the ones before — then feeds it back in and repeats, 12 bars ahead.

  3. STEP 3

    Many paths, one forecast

    It runs 10 times. The middle of those paths is the forecast; their spread is the range, and it widens the further out it looks.

    Median = forecast · band = range

Can fail badly around earnings. An earnings report is news the model cannot see. When a stock gaps on results, the forecast runs straight through it.

06 — Stock Analysis

A research desk of AI agents, on any US stock.

Ask for a report on any US-listed stock and a team of specialised AI agents goes to work. Four analysts gather the market, sentiment, news and fundamentals. Two researchers argue the bull and bear cases against each other. A trader drafts a plan, a risk team pulls it apart, and a portfolio manager makes the call.

You get the whole argument, written out — not just a rating. It is research to think with, not investment advice.

How a report is made
  1. STAGE 1

    Research desk

    MarketSentimentNewsFundamentals

    Four analyst agents gather the evidence, each on its own beat.

  2. STAGE 2

    Bull vs Bear

    Two researcher agents argue opposite cases, round after round. A research manager judges the debate.

  3. STAGE 3

    Plan and risk check

    Trader→AggressiveNeutralConservative

    A trader agent drafts the plan. Aggressive, neutral and conservative risk agents argue it out for the same number of rounds.

  4. STAGE 4

    The call

    BuyOverweightHoldUnderweightSell

    A portfolio manager agent weighs everything and gives one rating.

Rounds per depth: Shallow 1 · Medium 3 · Deep 5

Every agent’s work is written out in the report — you read the argument, not just the verdict.

  • Research desk: Market Analyst, Sentiment Analyst, News Analyst, Fundamentals Analyst
  • Debate: Bull Researcher, Bear Researcher, Research Manager
  • Plan and risk: Trader, Risk team
  • Decision: Portfolio Manager
DepthShallowMediumDeep
Debate1 debate round3 debate rounds5 debate rounds
Typical waitabout 11 minutesabout 14 minutesabout 20 minutes
Price per reportIncluded · 4 a month, then $x.xx$x.xx$x.xx

07 — MCP

Give your own agent the same nightly scan.

Everything on this site is also served over the Model Context Protocol, the open standard AI assistants use to call tools. Connect Claude — or any agent that speaks MCP — and it can query the ranked sectors, the heavy names, the forecasts and their track record directly: the same tables, the same numbers, fresh every trading night.

Read-only by design. Your agent can ask; it cannot trade, change anything, or see anyone else’s account. It is data for the agent to reason with, not a signal to follow.

Sector scan

All eleven sectors against SPY over every window, and which are leading.

Heavy names

The index-heavy names in a sector against that sector, with their weights and weight changes.

IndexHeavy list

The merged SPY, QQQ and DIA shortlist, ranked against each index.

Prices

A year of daily closes for any tracked ticker, aligned to the trading calendar.

Predictions

The current forecast table, each symbol’s forecast history, and the scorecard.

Watchlists

A TradingView import file for any sector or the IndexHeavy list.

How it helps while you trade

Check the rotation before acting on an idea

Ask whether a name’s sector is still leading, and whether the name is beating that sector — the two questions this product exists to answer, now answerable by your agent in the moment.

Start from a shortlist, not the whole market

An agent that begins with the names that actually move each sector spends its reasoning on a handful of candidates instead of five hundred tickers.

Weigh every forecast by its own record

Each forecast arrives with its hit rate and the scorecard behind it, so an agent can discount the ones that have not earned attention.

See exactly what it read

Whatever the agent concludes, the rows it worked from are the rows on your dashboard. Open the page and check its reasoning against the same numbers.

Connect with one URL

https://leaderslaggards.com/api/mcp

Add it as a custom connector in Claude or ChatGPT, or with `claude mcp add` / `codex mcp add`, and sign in with Google when the tool asks. No key to copy: it reads exactly what your account holds.

Pricing

$x.xx/month

One plan. Cancel anytime.

  • Every sector and its heaviest names, ranked every night
  • Predictions on the 1-hour, 4-hour and daily charts
  • 4 Shallow stock analyses a month

Deeper reports, per report: Medium $x.xx · Deep $x.xx · Extra Shallow $x.xx

Get started