Overview
Week of March 9
What changed, what it means, and how strong the evidence is. Ordered by significance, not recency.
Markets scored
1,847
of 4,061 evaluated
Withheld
2,214
failed a publication gate
Median confidence
64
MODERATE BAND
Patterns detected
37
All unvalidated
Data freshness
74
WEIGHTED MEAN
Watchlist, week over week
MoM sparkline, 12 periods| Market | Opp | 1W | Conf | Trend | Driver |
|---|---|---|---|---|---|
| PLACEHOLDER CO A | 62.4 | ▲ +3.1 | 71 | Price cuts up, DOM flat | |
| PLACEHOLDER METRO B | 57.9 | ▼ −1.8 | 68 | Pending to active fell | |
| PLACEHOLDER CO C | n/a | n/a | 38 | Withheld conf < 40 | |
| PLACEHOLDER ZCTA D | 49.1 | ▲ +0.4 | 52 | No material change |
Opportunity distribution
SCORED MARKETS, n=1,847Bins of 5 points, 25 through 85. Your watchlist markets marked.
Patterns detected this week
Pattern Increasing Buyer Leverage
PLACEHOLDER CO A · COUNTYStrength
75
Persistence
3 mo
Signals
3 / 3
Confidence
40
No historical validation yet
A described combination of current signals, not a demonstrated predictor. Confidence is capped at 40 until backtesting produces evidence.
Pattern Exit Liquidity Deterioration
PLACEHOLDER METRO B · CBSAStrength
88
Persistence
4 mo
Signals
4 / 5
Confidence
40
Shown at equal weight to favorable patterns
This market also shows improving acquisition conditions. A warning pattern is never subordinated to a favorable one in the same market.
Discover
Screen markets
Ranked within a peer set, never across peer sets. Withheld markets are counted, not hidden.
Filters apply to your preference profile. Best market nationally and best market for you are different questions.
Ranking
1,847 SCORED · 2,214 WITHHELD| # | Market | Opp | Conf | Acq | Exit | Reno | Carry | Econ | Risk | DOM | Cuts | Inv YoY | Pat | 12M |
|---|
2,214 markets withheld from this ranking
EVERY EVALUATED MARKET IS ACCOUNTED FOR| Gate | Markets | Share | What would change it |
|---|---|---|---|
| G3 Confidence below 40 | 1,402 | 63% | Listing metrics suppressed by the publisher at this population size |
| G1 Under 4 pillars, or acq/exit missing | 596 | 27% | Exit liquidity needs listing data unavailable here |
| G4 Stale inputs | 153 | 7% | QCEW construction wage older than two quarters |
| G2 Peer set under 30 | 63 | 3% | Too few comparable markets at this level |
Market · County · FIPS 06067
PLACEHOLDER COUNTY A
Peer set: top population tertile, 1,033 counties. Rank 1 of 1,033. Oldest contributing observation: ACS 5-year, 396 days.
Flip Opportunity
62.4
▲ +3.1 1W · ▲ +5.6 4W
Prior weights, unvalidated
Confidence
71
HIGH BAND
Median DOM
44
▲ +13 YoY · P78 OWN 5Y
Price cut share
21.4%
▲ +6.2pp YoY · P81 PEER
Repeat-sale price index & transaction volume
FHFA HPI, QUARTERLY, NSA · PLACEHOLDERCandles show quarterly open, high, low, close of the index. Bars below show repeat-sale pair count, which drives the sample adequacy component of confidence.
Confidence components
7 COMPONENTS| Component | Score | Wt | Level |
|---|---|---|---|
| Coverage | 79 | .25 | |
| Freshness | 74 | .20 | |
| Agreement | 71 | .15 | |
| History | 100 | .10 | |
| Sample | 81 | .15 | |
| Precision | 100 | .10 | |
| Stability | 62 | .05 |
Indicators
TOGGLE SERIES · SHADED = LOW CONFIDENCE PERIODWhat raised and lowered this score
SIGNED CONTRIBUTION VS PEER MEDIANFactors that could not be measured
- Distressed inventory and foreclosure activity. No free national source exists. Missing for every US market. Costs each one roughly 6 coverage points.
- Local renovation unit cost. National materials index only. The renovation pillar uses county construction wages and permit volume as proxies, both unvalidated.
- Current insurance cost. Latest available data ends in 2022.
Provenance
EVERY NUMBER TRACES TO AN ARTIFACT| Metric | Lvl | Age | Lic |
|---|---|---|---|
| Days on market | CNTY | 18d | Unver |
| Price cut share | CNTY | 18d | Unver |
| Sale to list | CNTY | 18d | Unver |
| HPI repeat sale | CNTY | 74d | PD |
| Unemployment | CNTY | 41d | PD |
| Construction wage | CNTY | 152d | PD |
| Effective tax rate | CNTY | 396d | PD |
| Hazard index | CNTY | static | PD |
Export is blocked for figures whose lineage includes an unverified source.
Watchlist
Followed markets
Change columns show movement since the previous score run. Pin markets to compare them.
Positions
4 MARKETS · 1 WITHHELD| Market | Lvl | Opp | 1W | 4W | 12W | Conf | DOM | Cuts | S2L | Inv YoY | Patterns | 52W |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| PLACEHOLDER CO A | CNTY | 62.4 | ▲ 3.1 | ▲ 5.6 | ▲ 9.2 | 71 | 44 | 21.4% | 97.8% | +19% | Buyer leverage | |
| PLACEHOLDER METRO B | CBSA | 57.9 | ▼ 1.8 | ▼ 4.2 | ▼ 6.1 | 68 | 61 | 27.8% | 96.1% | +31% | Exit deterioration | |
| PLACEHOLDER CO C | CNTY | n/a | n/a | n/a | n/a | 38 | n/a | n/a | n/a | n/a | Withheld | |
| PLACEHOLDER ZCTA D | ZCTA | 49.1 | ▲ 0.4 | ▼ 0.9 | ▬ 0.0 | 52 | 38 | 16.2% | 98.9% | +7% |
PLACEHOLDER ZCTA D is a mixed-resolution view
Price and listing metrics are native to ZIP. Employment, income, and property tax arrive from the containing county through a crosswalk at an allocation weight of 0.86. That is why its geographic precision component is 68 rather than 100.
Patterns
Pattern library
Three patterns implemented. The remaining twelve in the specification are documented as library entries marked not implemented, which is more honest than shipping thin versions of all of them.
Implemented
3
ALL UNVALIDATED
Documented only
12
NO DATA OR NO SPEC
Retired
0
KEPT VISIBLE WHEN IT HAPPENS
Detections this run
37
ACROSS 1,847 MARKETS
Pattern Increasing Buyer Leverage Unvalidated
v1.0.0 · CBSA + COUNTY · MIN CONF 60Supply is building, homes are taking longer to sell, and sellers are cutting prices, while transaction prices and local employment have not yet broken.
Signals
| Signal | Req | Threshold | Obs | Fired |
|---|---|---|---|---|
| Active listings YoY | YES | ≥ own 20Q P70 | +19.0% | 3 / 3 mo |
| Median DOM YoY | YES | ≥ own 20Q P70 | +42.0% | 3 / 3 mo |
| Price cut share YoY | YES | ≥ own 20Q P70 | +6.2pp | 2 / 2 mo |
| Median sale price YoY | CORR | between −2.0 and +4.0% | +5.1% | Not fired |
| Unemployment 12M change | CORR | ≤ +0.5pp | +0.7pp | Not fired |
Signal persistence
Filled cells are periods where the signal met its threshold. Persistence is the single most effective false-positive control in the framework.
What would invalidate this
- Unemployment rising more than 1.5 points over 12 months. That reclassifies the pattern as demand collapse rather than buyer leverage.
- Median sale price falling more than 6 percent year over year. The acquisition window has become a falling knife.
- Active listings falling back below the own-history median for 2 consecutive periods.
What to watch next
Sale-to-list ratio, pending-to-active ratio, months of supply. If those deteriorate alongside, the read shifts from buyer leverage to demand deterioration.
Historical behavior
No historical validation yet. This pattern has not been backtested. Until validation produces at least 30 instances across at least 8 markets and 3 calendar years, PatternIQ will not tell you what usually happens next, because it does not know.
Known failure conditions
This fires identically at the start of a normal seasonal inventory build and at the start of a genuine correction. Year-over-year framing handles most seasonality, but a market with an unusual seasonal profile can still trip it.
Deal Analyzer
Underwrite a property
Arithmetic on your own assumptions. No model output appears here until calibration passes, so every number below is a consequence of what you entered.
Result
ArithmeticExpected net profit
$38,420
COMPUTED FROM YOUR INPUTS, NOT A FORECAST
| Total project cost | 326,580 |
| Gross profit | 65,420 |
| Return on cost | 11.8% |
| Return on cash | 31.2% |
| Annualized return | 27.4% |
| Break-even resale | 326,580 |
| Max advisable purchase | 251,300 |
| Margin of safety | 10.5% |
Market context
AT TIME OF ANALYSISOPP 62.4 · CONF 71 HIGH · 1 PATTERN
Exit liquidity is the weakest pillar in this market at 44 of 100. Your 150-day holding assumption sits below the county median DOM of 44 plus a typical renovation window. Consider testing 180 and 210 days.
Inference reasoning from the pillar breakdown, not a prediction
Sensitivity
NET PROFIT · RESALE × RENOVATION| Reno ↓ / Resale → | 330k | 347k | 365k | 383k | 400k |
|---|---|---|---|---|---|
| 52,000 | 13,900 | 30,000 | 47,000 | 64,000 | 80,100 |
| 62,000 | 5,300 | 21,400 | 38,420 | 55,400 | 71,500 |
| 72,000 | −3,300 | 12,800 | 29,800 | 46,800 | 62,900 |
| 82,000 | −11,900 | 4,200 | 21,200 | 38,200 | 54,300 |
Break-even
PROFIT VS RESALE PRICEZero line crossed at 326,580. Everything left of it is a loss.
Portfolio
Projects, planned against actual
Your own record is the only data in PatternIQ that measures what you actually care about.
Completed
2
SOLD
Active
2
1 LISTED, 1 RENOVATING
Realized net profit
$63,500
2 PROJECTS
Mean reno overrun
+13.1%
n=2, NOT A PATTERN
Projects
PLANNED VS ACTUAL| Project | Status | Purchase | Reno plan | Reno actual | Var | Days plan | Days actual | Var | Net profit |
|---|---|---|---|---|---|---|---|---|---|
| PROJECT 1 | Sold | 212,000 | 54,000 | 61,300 | +13.5% | 140 | 176 | +36d | 29,400 |
| PROJECT 2 | Sold | 188,500 | 47,000 | 52,900 | +12.6% | 130 | 151 | +21d | 34,100 |
| PROJECT 3 | Listed | 244,000 | 68,000 | n/a | n/a | 160 | n/a | n/a | n/a |
| PROJECT 4 | Renovating | 201,000 | 58,000 | 41,200 | to date | 145 | n/a | n/a | n/a |
Variance by renovation category
ACTUAL VS PLANNEDEvery row is below the 3-project floor. None of these is presented as a pattern.
Personal patterns
SAMPLE GATES ENFORCEDNothing to report yet
Personal patterns need at least 3 completed projects in a category before an early signal is shown, and at least 6 before it is presented as a pattern. You have 2. PatternIQ will not draw a conclusion from that.
| Completed | Presentation |
|---|---|
| 1 to 2 | Nothing shown |
| 3 to 5 | Early signal, low confidence, n in the headline |
| 6 to 9 | Pattern with a confidence interval on effect size |
| 10+ | Eligible to adjust estimates, with your approval |
Run Analysis
What does my history tell me
Findings ranked by financial significance, each with sample size and confidence in the headline rather than the footnote.
Inference Holding periods ran longer than planned in both completed projects
FINDING 1 OF 2| Project | Planned | Actual | Variance | Carry cost |
|---|---|---|---|---|
| PROJECT 1 | 140 | 176 | +36d | $2,480 |
| PROJECT 2 | 130 | 151 | +21d | $1,610 |
Why this matters. Holding period compounds. It raises interest, taxes, insurance, and utilities simultaneously, and it delays the capital you need for the next acquisition. It is the most commonly underestimated line in flip underwriting.
Sample size 2. This is an observation, not a pattern.
Two projects cannot establish a tendency. Financial significance if it holds: roughly $2,100 in additional carrying cost per project at your current financing terms. PatternIQ will revisit at 6 completed projects.
Fact Renovation costs exceeded budget in both completed projects
FINDING 2 OF 2PROJECT 1: 54,000 planned, 61,300 actual, 13.5 percent over. PROJECT 2: 47,000 planned, 52,900 actual, 12.6 percent over.
Recommended question rather than recommended action: is your contingency percentage set below your own observed overrun rate? You currently use 12 percent.
Learn
Median days on market
Every metric carries the same ten-part explanation, versioned so a score from March shows the explanation current in March.
Explanation
EDUCATION v1.0.01. Definition
The middle value of the number of days that listings sold in a period spent on the market before going under contract.
2. How it is calculated
The publisher takes every listing that went pending in the period, computes days between list date and pending date, and reports the median. PatternIQ does not recompute it.
3. Why investors track it
It is the most direct available measure of how long it takes to convert a finished property into cash. For a flip, holding period is a cost that compounds across interest, taxes, insurance, and utilities at once.
4. What rising values can mean
Demand is softening relative to supply, pricing is ahead of what buyers will pay, or the mix of what is selling has shifted toward harder-to-sell property. Rising days on market usually precedes falling prices rather than following them.
5. What falling values can mean
Demand is firming, or inventory has thinned so what remains is the more desirable stock. A fall driven by scarcity is different from one driven by demand, and the two are distinguished by looking at new listings alongside.
6. Interactions
Read alongside price reduction share and sale-to-list ratio. Days on market rising while price cuts stay flat suggests sellers are waiting rather than capitulating. Rising alongside cuts suggests capitulation has begun.
7. Common interpretation mistakes
- Treating it as a measure of the whole market. It only describes properties that sold. Listings that never sold are invisible in it.
- Comparing it across publishers. Realtor.com and Redfin define it differently and the levels are not interchangeable.
- Ignoring relistings. A listing withdrawn and relisted restarts the clock, which understates true time to sell in slow markets, exactly when you most want the truth.
8. Example
A county reports 44 days this month against 31 a year ago. That is a 42 percent increase, at the 78th percentile of its own five-year history. Combined with price cuts at 21 percent of listings, that is two of three required signals for Increasing Buyer Leverage.
9. Data source
Realtor.com Residential Inventory Core Metrics, county level. License status unverified, so figures derived from it cannot be exported.
10. Update frequency
Monthly, published roughly one to two weeks after month end.
This metric in your markets
CURRENT LEVEL VS OWN 5Y RANGEMetric index
- Active listing count
- Confidence Score
- Effective property tax rate
- Flip Opportunity Score
- House price index, repeat sale
- Median days on market
- Months of supply
- Peer set
- Pending to active ratio
- Price reduction share
- Sale to list ratio
- Unemployment rate
Data & Sources
Where everything comes from
Including what is missing, what is stale, what is legally unresolved, and what the backtest found. Negative results appear here at the same size as positive ones.
Sources in use
9 DATASETS · 2 UNVERIFIED LICENSE| Source | Supplies | Level | Cadence | Age | Freshness | License |
|---|---|---|---|---|---|---|
| FHFA HPI | Repeat-sale price index | ST/CBSA/CNTY/ZIP5/TRACT | QTR | 74d | Public domain | |
| BLS LAUS | Unemployment | ST/CBSA/CNTY | MON | 41d | Public domain | |
| BLS QCEW | Employment, construction wage | CNTY | QTR | 152d | Public domain | |
| Census ACS 5-yr | Income, tenure, tax paid, vacancy | CNTY/TRACT/ZCTA | ANN | 396d | Public domain | |
| Census BPS | Building permits | ST/CBSA/CNTY/PLACE | MON | 28d | Public domain | |
| FEMA NRI + OpenFEMA | Hazard risk, disaster history | CNTY/TRACT | STATIC | n/a | Public domain | |
| Freddie Mac PMMS | 30-year mortgage rate | NATION | WKLY | 4d | Attribution | |
| Treasury FIO | Homeowners insurance premium | ZCTA | ONE-OFF | 1,320d | Public domain | |
| Realtor.com Core | DOM, listings, price cuts, pending | CBSA/CNTY/ZIP | MON | 18d | Unverified | |
| Redfin Data Center | Sale price, sale to list, above list | CBSA/CNTY/ZIP | MON | 18d | Unverified |
Known gaps
Distressed inventory
No free national source. County recorders hold the primary records across 3,000+ jurisdictions, many prohibiting bulk automated access.
EFFECT: −6 COVERAGE PTS, EVERY MARKET
COST TO CLOSE: $599/MO OR ATTOM CONTRACT
Local renovation cost
Materials indexes are national only. No free equivalent of the RSMeans city cost index exists.
EFFECT: RENO PILLAR WEIGHTED .06
COST TO CLOSE: RSMEANS, UNPRICED
Current insurance cost
The best free ZIP-level data ends in 2022, in the middle of the largest premium repricing in decades.
EFFECT: FRESHNESS EFFECTIVELY ZERO
COST TO CLOSE: NO KNOWN SOURCE
Property-level data
Free data supports market analysis, not address-level underwriting. The Deal Analyzer runs on your assumptions.
EFFECT: NO COMPS, NO AVM
COST TO CLOSE: RENTCAST $74/MO
Backtest results
Information coefficient
AWAITING FIRST RUNThis panel will show the rank IC time series with its Newey-West statistic, the decile spread table, comparison against the four required baselines, and per-pattern lift with FDR-adjusted p-values.
Status
No backtest has run yet
The Opportunity Score is using declared prior weights that have never been tested against realized outcomes. Until the harness runs and clears its sample gates, treat every score in this product as a structured summary of current evidence and not as a demonstrated predictor.
| Gate | Required | Have |
|---|---|---|
| Instances | 30 | 0 |
| Distinct markets | 8 | 0 |
| Calendar years | 3 | 0 |
| Genuine vintages | some | 0 |
