How the Unvelo Pendulum Works
The Unvelo Pendulum is a weekly housing-market leverage index for 95US metros plus the national market. Scores run from -100 (peak buyer's market) to +100 (peak seller's market). Each score blends four signals from Redfin, Realtor.com, and Zillow: inventory growth, days on market, price cuts, and sale-to-list ratio. This page documents the full methodology.
By Alessandro Bordignon, creator of the Pendulum · Last updated Jun 15, 2026
What is the Unvelo Pendulum?
The Pendulum is a free market-leverage gauge built for US real estate agents. It answers one question per market: who holds the negotiating power right now, buyers or sellers? One score per market, refreshed every Monday, no signup required.
Agents walk into listing appointments with national headlines and a gut feel. Headlines describe the country. Your sellers live in one metro. The Pendulum gives you a defensible, sourced read on your market that you can show a client and stand behind.
Open the live Pendulum to see today's readings.
What does the score mean?
Every market gets a score from -100 to +100. Negative means buyers hold the leverage. Positive means sellers do. Zero means the market is balanced. The further from zero, the stronger the tilt.
| Score range | Label | Read it as |
|---|---|---|
| above +25 | SELLER-TILTED | Sellers firmly in control of the pace |
| +11 to +25 | LEANING SELLER | Sellers ahead, edge softening |
| -10 to +10 | BALANCED | Neither side holds a strong edge |
| -25 to -11 | LEANING BUYER | Buyers gaining room to negotiate |
| below -25 | BUYER-TILTED | Buyers firmly in control |
Example: a +30 market moves faster than the national pace, with fewer price cuts and stronger closing prices. A -30 market has swelling inventory and homes that sit. Both can exist in the same month, two states apart.
Which four signals drive the score?
Four signals, equal weight, 25% each: inventory growth year over year, median days on market, the share of listings with a price cut, and the sale-to-list ratio. Together they cover supply, speed, seller behavior, and closing power.
| Signal | What it measures | Favors | Range | Weight |
|---|---|---|---|---|
| Inventory growth (YoY) | Active listings change year over year, percent. | buyers when higher | 30% | 25% |
| Days on market | Median days on market for sold homes. | buyers when higher | 60 days | 25% |
| Price cuts | Share of active listings with at least one price reduction, percent. | buyers when higher | 25% | 25% |
| Sale-to-list ratio | Median sale price divided by last list price, percent. | sellers when higher | 3% | 25% |
Each signal is observable in public data, and each one moves when negotiating power moves. Rising inventory hands buyers options. Longer days on market drain seller patience. Price cuts are sellers blinking first. Sale-to-list shows who wins the final negotiation.
How is the score calculated, step by step?
Blend, compare, normalize, average. Each signal is blended across sources for every month. The blended value is compared with the same month's national value, divided by a fixed range, and capped. The four results are averaged and scaled to the -100 to +100 score.
- Step 1: Blend the sources
For each month, every signal is a weighted average across sources: Redfin 50%, Realtor.com 30%, Zillow 20%. Price cuts blend 60%/40% from Redfin and Realtor.com. Sale-to-list comes from Redfin only. When a source has not published a month yet, the remaining weights renormalize.
- Step 2: Compare with the national baseline
Each blended value is compared with the same month's blended national value. Markets are measured as distance from the national pulse. The national score uses a fixed balanced baseline instead, because a market cannot be measured against itself.
- Step 3: Normalize and clamp
Each difference is divided by that signal's normalization range: 30 points for inventory growth, 60 days for time on market, 25 points for price cuts, and 3 points for sale-to-list. Results are capped between -1 and +1. Signals that favor buyers count negative.
- Step 4: Average and scale
The four normalized scores are averaged and multiplied by 100, then rounded. The result is the published Pendulum score, from -100 (peak buyer's market) to +100 (peak seller's market).
As a formula: position = round(100 × mean(clamp(±(value − national) / range)))
| Signal | San Francisco, CA | Reference | Normalized |
|---|---|---|---|
| Inventory growth (YoY) | -22.4% | -4.8% | +0.587 |
| Days on market | 21 days | 46 days | +0.417 |
| Price cuts | 22% | 25% | +0.120 |
| Sale-to-list ratio | 109.1% | 99.3% | +1.000 |
Reference: the national averages (May 2026). Mean of the four normalized scores: +0.5308. Multiply by 100 and round: +53. That is exactly the published score for San Francisco, CA.
“A market is not a buyer's market because homes sit 60 days. It is a buyer's market because homes sit longer than the rest of the country. The Pendulum measures distance from the national pulse, not distance from an opinion.”
Every number above comes from the published dataset. You can download the dataset (JSON) and reproduce any score by hand. The file includes the scoring constants, every market's blended numbers, and the per-month national baseline.
Where does the data come from?
Three listing-data sources and one macro source. Redfin Data Center is the primary anchor. Realtor.com Research and Zillow Research blend in as second and third opinions. FRED supplies national macro context like mortgage rates. All four publish public, citable data.
| Source | Role | Blend weight |
|---|---|---|
| Redfin Data Center | Primary anchor. All four signals, monthly history. | 50% |
| Realtor.com Research | Inventory, time on market, and price cuts, monthly history. | 30% |
| Zillow Research | Inventory and days-to-pending (a time-on-market proxy), monthly history. | 20% |
| FRED (St. Louis Fed) | National macro context: mortgage rates, housing starts. Not scored. | 0% |
Sources publish on their own schedules. Zillow typically lags Redfin by about a month. When a source has not published a month yet, the remaining weights renormalize, and the score still ships. The latest reading often blends Redfin and Realtor.com until Zillow catches up. Nothing is interpolated and nothing is guessed.
How is the national score different?
Metros are scored against the same month's national values. The national market cannot be scored against itself, so it uses a fixed balanced baseline: flat inventory, 45 days on market, 20% price cuts, and 100% sale-to-list.
Distance from the national pulse is the point. 46 days on market is the national median right now. A metro at 70 days is slow in absolute terms, but what matters for your negotiation is that it is slower than the country. The national score answers a different question: how far is the whole country from a balanced market?
Which markets does the Pendulum cover?
96 markets today: 95 metros plus the national reading. A market is included when the primary source publishes complete data for all four signals. Coverage grows as source data allows. The goal is every US market.
When a metro's source data is incomplete in a given month, the metro is skipped rather than shipped with a partial scorecard. That is why the count can move slightly between weeks. The roadmap runs from major metros to all metros, then to state and county levels as the sources support them.
How often is the Pendulum updated?
Every Monday at 10:00 UTC. The pipeline pulls fresh source data, recomputes every score, and republishes the dataset. The underlying sources publish monthly with a lag, so a June refresh typically reflects May 2026 closings. Each market shows its own data month.
Two dates matter. The data month (May 2026 right now) is the month the numbers describe. The refresh date (Jun 15, 2026) is when the pipeline last ran. Weekly snapshots are archived, so the 12-month trail behind every gauge is real history, not a backfill.
What the Pendulum can't tell you
The Pendulum is a leverage gauge, not a crystal ball. Know its edges before you quote it.
- It reads metro-level data. Your ZIP code or neighborhood can lean differently than the metro around it.
- It is not a price forecast. Leverage describes today's negotiation, not next year's prices.
- Sources publish with a lag of roughly one month. A sudden shift shows up in the next data month, not the same week.
- Markets without complete primary-source data are skipped entirely rather than estimated.
- Source definitions differ slightly. Zillow's days-to-pending proxies days on market, which is one reason it carries the lowest blend weight.
Frequently asked questions
Is the Unvelo Pendulum free?
Yes. The Pendulum is free, with no signup and no paywall. Open the live gauge, pick your market, and read the score. It refreshes every Monday.
Is the Pendulum a home-price forecast?
No. The Pendulum measures current negotiating leverage, not future prices. A buyer-tilted score means buyers hold the edge in negotiations right now. Prices can still rise in a buyer-tilted market and fall in a seller-tilted one.
What does a score of +30 mean?
+30 is seller-tilted territory. Homes in that market move faster than the national pace, with fewer price cuts and stronger sale-to-list ratios. Sellers hold most of the negotiating power. Any score above +25 reads as seller-tilted.
Can I cite the Pendulum in client presentations?
Yes. Agents use Pendulum readings in listing appointments and buyer consults. Credit the Unvelo Pendulum and note the data month shown with the score. Every published score can be checked by hand against the public dataset.
Why does the Pendulum differ from other market indexes?
Most indexes track a single source or report raw averages. The Pendulum blends three independent listing datasets, scores each market against the same month's national values, and publishes the math. When two indexes disagree, check what each one actually measures.
What is the difference between the Pendulum and the Weekly Pulse?
The Pendulum scores market leverage from monthly source data. The Weekly Pulse tracks faster-moving national indicators, like mortgage rates and new listings, week by week. The Pulse tells you what changed this week. The Pendulum tells you who holds the power.
How many markets does the Pendulum cover?
95 US metros plus a national reading, as of May 2026. A market is included when the primary source publishes complete data for all four signals. Coverage grows as source data allows. The goal is every US market.
See where your market sits this week.