Deep dive · NYSE: HVT
Haverty Furniture Companies, Inc. stock analysis
Everything behind the call on Haverty Furniture Companies, Inc. 6 sections, ordered the way risk stacks up: survival first, quality and price in the middle, context last. The Beacon gives you the read. This is where you check the work.
The five reads
Global tapeOverseas lean-2.73%
Frankfurt25F-2.73%⏱
Where else this company trades. A closed venue shows its last close, marked ⏱ — not a live price. Overseas lean is the average move across the 1 overseas venue listed above, excluding the home listing — a read on overnight sentiment rather than a signal.
The bottom linestart hereAs of 2026-09-17
Chart pattern
Above both averages
The 50-day sits above the 200-day
50-day $26.80 · +1.8%
200-day $23.90 · +14.1%
200-day $23.90 · +14.1%
Earnings
In 40 days
28 Oct 2026
beat 5–6 of the last 8
cohort P(beat) 64% · n=45,184
cohort P(beat) 64% · n=45,184
Valuation
10.4× forward
95th of its own five-year range
5yr median 9×
Dividend
4.84%
Forward annual yield at today's price
Risk flags
2 on watch
5 checks run, 0 clear
beta 0.82 · vol 22% ann.
worst drawdown -25%
worst drawdown -25%
New to this? Every metric carries a plain-English read beside it, and an ⓘ next to a label opens the definition. Anything unfamiliar is also defined at the foot of the page under Glossary.
NYSE · USDHome Improvement Retail
$27.28+0.03 · +0.11%CLOSED · vol 291.7K
52-week range↑ 37% off low · ↓ 8.5% off high
$19.89now $27.28$29.83
Quote Sep 18, 2026 · time not on file
Tape 2026-09-17
Factors 2026-09-17
Tape 2026-09-17
Factors 2026-09-17
01
Survival & solvency
and 0 of 5 forensic checks clear.
↳ Plain EnglishBefore anything else: can this company pay its bills? These blocks check the balance sheet, whether your slice of ownership is being diluted, and whether the accounting looks honest.
Piotroski F-score
F 5/7WATCH
Piotroski F-score 5 of 7. Nine pass/fail tests covering profitability, leverage and operating efficiency; seven or more is strong.
Balance sheet
Your ownership stake
Buyback yieldGROSS
1.28%
Repurchases as a share of market value, before new issuance is netted off.
Stock-based comp / revenueIN LINE
1.0%
Stock comp is the dilution pressure a buyback has to outrun. We read up to 5% of revenue as modest and above 12% as heavy — our own reading band, not a measured peer statistic.
Accounting forensicsSEC filings
Red-flag x-rayGRADE D
50 / 100
0 critical · 2 watch · 0 clear of 2 scored tests
Red-flag x-ray · every check, passes included
CheckStatusValueWhat it looks for
Bankruptcy distressNAn/aAltman-Z not available · fundamentals
Earnings manipulationNAn/aBeneish-M not available · fundamentals
LeverageNAn/aLong-term debt / equity not available · fundamentals
Profitability qualityWATCHF 5/7Piotroski F-score 5 of 7 · fundamentals
Filing forensicsWATCH1 flag1 filing red flag on the latest 10-K · forensic
5 checks, 2 watch. The passes are shown because a score built from five checks means something different from one built from three.
02
Business quality
The quality blocks below render whatever the filings support; anything absent says so in place.
↳ Plain EnglishMargins tell you how much of each sales dollar the company keeps. Return on capital tells you whether reinvesting that money creates value. Both are read against what the filings support, not against a target.
12-1 momentum
+28.8%EX LAST MONTH
Return over the last year excluding the most recent month — the standard momentum factor, which drops the short-term reversal.
Profitability
Data unavailableNo parsed income statement is on file, so margins, returns on capital and cash generation cannot be computed for this name.
Growth & relative strength
Trailing return by period
1 month-3.6%
3 months+12.0%
6 months+29.0%
Year to date+15.6%
1 year+16.8%
Relative-strength rankTOP 19%
81 / 100
Where this name's 3-month return against the S&P 500 ranks across the market — a quarter's standing, not the year's.
Around earningsAug 4, 2026
P(beat next quarter)COHORT BASE RATE
64%
survivor-tilted universe (currently-listed names; large-cap-skewed coverage)
Beat rate · this nameBEAT 5–6 OF THE LAST 8
6 of 8
Its own record against consensus. 8 prints is a small sample, which is why the cohort rate above carries the probability.
Fell after a beatP(FELL | BEAT)
24%
A beat is not a plan: this share of them still closed lower on the day.
Rose after a missP(ROSE | MISS)
20%
The mirror of the line above, over its own denominator.
Typical earnings moveMEDIAN ABSOLUTE
±6.6%
Unsigned magnitudes from its own prints. Direction is not part of this read.
Reaction: beat vs missASYMMETRIC
+3.6% / -3.6%
The average day-one move after a beat and after a miss. A miss dropping the stock further than a beat lifts it is the asymmetry worth sizing for.
Dividends & income
Dividend yieldforward annual
4.84%
The cash return the payment alone provides at today's price.
Inventory ledger
Data unavailableNo inventory detail is tagged in this company's filings. The ledger needs the XBRL inventory components, which a business without physical goods does not report.
Segments & geography
Data unavailableRevenue by segment and region is not on file for this name, so product and geographic concentration are not shown.
Pattern reliability · platform-wide cohortSnapshot · Sep 18, 2026
Uptrend · leans bullish
HorizonHigherMedianObservations
10d47%-0.1%1,125,547
21d47%-0.2%529,218
63d48%-0.9%169,468
slope +0.07 over 40 bars
Long legged doji
HorizonHigherMedianObservations
10d48%0.0%171,393
21d49%0.0%80,724
63d49%-0.3%25,610
indecision, long shadows
Double bottom · leans bullish
HorizonHigherMedianObservations
10d49%0.0%1,242,409
21d50%0.0%584,002
63d51%+0.2%184,289
two troughs ≈ equal — reversal if neckline breaks
Inverse head and shoulders · leans bullish
HorizonHigherMedianObservations
10d49%0.0%491,675
21d50%0.0%231,291
63d51%+0.1%73,497
low below two ≈-equal shoulders — bullish reversal
Descending triangle · leans bearish
HorizonHigherMedianObservations
10d47%0.0%504,664
21d48%0.0%237,571
63d49%0.0%74,935
flat support, falling resistance
Near support · leans bullish
HorizonHigherMedianObservations
10d51%+0.1%533,688
21d53%+0.2%250,417
63d54%+0.4%78,400
testing recent support
Bollinger squeeze
HorizonHigherMedianObservations
10d43%0.0%757,951
21d45%0.0%355,725
63d47%0.0%112,112
bands at a multi-month tight — breakout pending
0395TH OF ITS 5-YEAR RANGE
Price & timing
Price $27.28, 10.4× forward earnings, RSI 49.
↳ Plain EnglishA multiple is what the market charges for a dollar of this company's earnings, cash or revenue; the trend and the options tape say what the price is doing around it. Together they are the timing picture, not the quality one.
Trend
Above 50 & 2002 OF 3 AVERAGES
50-day $26.80 · 200-day $23.90. Price above both is the structural definition of an uptrend.
The setup right now
Price · 504 sessions2024-09-13 → 2026-09-17
price50-day200-day
How to read this
Up 0.11% on the session.
Follow the solid line: each point is a session close. Compare it to the 50-day average — while price stays above it the short-term trend is up, and a drop below is the first structural warning.
Momentum (RSI-14)COOL
49
030 oversold70 hot100
Where momentum sits on a 0–100 scale. Above 70 is overbought, below 30 oversold; the middle says the move has room either way.
Typical day (ATR-14)±2.6%
$0.71
The average true range — how far this stock moves on an ordinary day. It sets the scale a stop has to clear.
MACDBELOW SIGNAL
-0.08
The short-term trend against its own signal line. The sign is the read; the size is not a target.
Bollinger position37% OF RANGE
Mid band
Where price sits inside its own volatility band. The edges are where a move is stretched, not where it reverses.
20-day VWAPPRICE BELOW
$27.43
The volume-weighted average of the last month. Price above it means recent buyers are, on average, in profit.
Chandelier exitLONG INTACT
$26.51
A volatility-based trailing stop, set a few average ranges below the recent high.
28 Oct 2026 · in 40 daysNext earningsThe next scheduled catalyst. The forward guide usually moves the tape more than the printed number. Date as the vendor last stated it on 7 Aug; companies confirm about three weeks ahead, so it can still move.
2026-09-25 · in 7 daysConsumer SentimentA scheduled macro release. It moves this stock independently of anything the company does.
Setup & oddsNightly
Direction oddsnext 12m
50 / 50cohort n=13,084
Effectively a coin flip. A split this close is not an edge you can size on, whichever side is nominally ahead. These are the cohort's odds — comparable names over the window — not this stock's own history.
Pullback riskoverstretch read
Low0.73× segment rate · n=16,498
Historically, mid names in this tape state saw a ≥10% pullback within 63 trading days in 58% of cases (n=16,498).
Overstretchvs the 20-day mean
-0.6σ
How many standard deviations price sits from its 20-day mean. Beyond 2σ is genuinely stretched; inside it is ordinary.
Bottom zoneif it breaks down
$26 – $273 supports
Where prior demand appeared: S1 at $27.08, Pivot low at $26.99, Pivot low at $26.59.
Slide scorefalling-knife check
12 / 100Neutral
A composite of consecutive down days, gap risk and volume expansion. Low means any decline is orderly.
Signal
The odds are a coin flip.
Historically, names tripping 1 validated red flag finished lower 12 months later in 50% of cases (n=13,084) — a survivor-biased floor, not a forecast.
Valuation
Earnings yield vs cashABOVE CASH
0.01pp
The earnings yield clears the risk-free rate, so the current return alone competes with cash.
Multiples
Forward P/Eearnings expected over the next 12 months10.4×5yr median 9×
Trailing P/Eearnings already reported19.9×95th of its 5yr range
PEG ratioP/E divided by growth; near 1 reads fair for the growth1.96
1 of the 3 priced multiples above sits in the richer half of its band. Each priced multiple is read against this stock's own history where a five-year range is on file, so the question is whether today is dear for this name rather than for the market. The multiples are computed separately from the quote above and are not re-priced as it moves, so they can lag it.
Price horizon & timingSep 17, 2026
Entryzone of pivot low + S1
$27.11In zone
Where the composer's entry sits, on its own stated basis.
Target52-week high
$29.47$29–$30
A zone, not a price. It is where the composer expects supply, not a forecast.
Stop2× ATR below entry
$25.70risk $1.41/sh
The level that says the setup is wrong, on the basis named beside it.
Risk / rewardtarget vs stop
1.67:1
Reward against risk on the composer's own levels.
Measure from a price$
To the 50-day
-1.8%
short-term trend line · $26.80
To the 200-day
-12.4%
primary trend line · $23.90
To the composer's target
+8.0%
52-week high · set Sep 17, 2026 · $29.47
To the composer's stop
-5.8%
2× ATR below entry · set Sep 17, 2026 · $25.70
To the 52-week high
+9.3%
highest print of the past year · $29.83
To the 52-week low
-27.1%
lowest print of the past year · $19.89
A typical day (ATR)
$26.57 – $27.99
±$0.71 average true range
Type any price to re-measure every distance from it. Each row is a level this page already states somewhere else, so nothing here is a new number — and none of it is a recommendation to trade at that price.
Signal
The levels line up.
Entry $27.11 · stop $25.70 · target $29.47, a 1.67:1 ratio. The composer flagged: fair-value reference.
Gamma map
Data unavailableThis name has no listed options chain, so there is no dealer gamma to map. Regime, flip level, walls and the implied move all need a chain to exist.
Options tape
Data unavailableNo options chain is connected for this ticker. Max pain, the put/call ratio, the implied move and the open-interest ladder all need the chain feed.
0413F quarterly · Form 4 dailyNET SIGNAL +3
Ownership & flow
Institutions hold 89%, short interest is 0.05% of float.
↳ Plain EnglishWho is buying and selling, and what they had to disclose. Institutions file 13F 45 days after quarter end, executives file Form 4 within two days, and Congress discloses under the STOCK Act with a 30–45 day legal lag. Every one of those is a look backwards.
Net signal
+3LEANING · 30d
1 bull / 0 bear factors. The points sum to the net beside the label; nothing here is weighted after the fact.
The QTick signalNightly
Earnings4 QUARTERS · TIMING
+3
latest EPS beat by 9%
Signal
1 bull / 0 bear factors
Independent factors scored and netted to +3 over 30d, a leaning reading. 1 of 9 factors carried a direction; the others scored zero and do not vote. The signal describes where the data leans, not what to do about it.
Cross-category read
Data unavailableFewer independent evidence families are ready than the engine's minimum, so there is nothing to check for agreement. A cross-category read over two categories is not a read.
Wall Street
Data unavailableNo sell-side firm publishes research on this name. There is no consensus rating, no price target and no estimate revisions to weigh — for a company this size that is the norm rather than a gap in our data.
Who owns it13F quarterly · Form 4 daily
Institutional ownershipINSTITUTIONALLY HELD
89.0%
The share held by 13F filers. It describes last quarter's positioning, not today's — filings land up to 45 days after quarter end.
Insider ownershipOFFICERS & DIRECTORS
15.0%
The share held by the people running the company.
Short interest (float)LIGHT
0.05%
Falling — shares short changed -23.1% over 30 days. There is no short base to squeeze here.
Free float91% OF SHARES
13.61M
Shares actually available to trade, excluding insider and restricted holdings.
Dollar volume (ADV)54th PCTILE
$3.40M/day
Average dollars traded per day, and where that ranks cross-sectionally. Thin liquidity is a cost you pay on the way out, not just on the way in.
Borrow feeData unavailableWe do not license securities-lending data, so the annualized cost of borrowing shares to short is not something we can show. Short interest and days-to-cover above are the closest read we can source.
The story · news & sentimentRolling window
Why it moved todayNONE
+0.11%
HVT was little changed +0.11% since the last close — no notable catalyst, and the move itself was modest.
Congress
Data unavailableNo congressional trades in this name have been disclosed under the STOCK Act. Lawmakers' disclosed trades cluster heavily in large caps, so an empty list is the usual state.
050 graded callsORDINARY VOLATILITY
Risk & track record
Beta 0.82, 22% annual volatility, a -25% worst drawdown, and 2 of 8 risk surfaces clear.
↳ Plain EnglishWhat owning it does to a portfolio, what the model votes against, and how often our own calls have been right. Every call is graded when it resolves, and the ones still open stay visible.
A very bad day (EVT)
−5.6%1-IN-200 DAY
Fitted to the tail of the return distribution rather than assuming a bell curve. Size the position so a day like this is survivable.
What it does to a portfolio5-year window
Beyond the tail (ES)1-IN-1,000 DAY
−10.8%
The average loss on a 1-in-1,000 day or worse — deeper in the tail than the 1-in-200 figure above and further than the measured history reaches, so it reads as the scale of a crash, not a forecast of one.
Tail understatementNEAR-NORMAL
1.3×
The tail runs close to what a normal curve would predict, so volatility is a reasonable proxy for risk here — which is not true of every name.
Max drawdownWORST FALL
-24.7%
The deepest peak-to-trough fall in the measured window. Only hold a position you can watch do this.
Beta (vs the market)IN LINE
0.82
Moves about 0.8× the market over the past year.
Annualized volatilityORDINARY
22%
How far it swings over a year, from its own realized returns.
Tactical phaseNIGHTLY READ
Neutral
Selling pressure56
Buying pressure0
Overheat1
State: neutral — No dominant short-horizon pressure in the tape.
What votes against
Data unavailableNo factor in the model currently votes against this name. That is a statement about the model's inputs today, not a claim that nothing could go wrong.
Failure modes · by company type
Data unavailableFailure modes are routed per company type, and only the bank-run and crypto-treasury rulebooks are validated so far. This name routes as growth, so there is no cohort-specific checklist to run. A generic list would read as analysis without being any.
The risk case · why not to buy8 surfaces · nightly
Suggested position capmax defensible size
3.2%tail cap binds
max defensible size 3.2% - tail binds
Avoid score6-month distress
1.0%0.1× its cohort · mid-volume names
Observational 6-month large-drawdown risk ranking — not a forecast, not alpha. Historical distress rate for this mid-volume names segment: 8.7%. This name reads 0.1× that base rate.
LowDistress screenThe distress model did not score this name in the latest run, so this surface is unmeasured rather than clear.Not scored
MedPosition size, not conviction, is the binding constrainthalf-Kelly 55.0% · tail cap 3.2% · category cap 54.6%. The tail cap binds, so 3.2% is the number (n=140,733). We show which one binds rather than the friendliest.Sizing
LowLottery-ticket behaviourIts best 21-day return sits at 3.4% against a 10.8% cohort 90th-percentile threshold — below the line where forward returns have historically turned negative.Behaviour
LowCash runwayNot scored for this name.Not scored
MedPost-earnings driftMove already priced: after jumps this large, measured drift edge is zero.Timing
LowDividend riskNot scored for this name.Not scored
LowRestatement historyFiling history is not deep enough to check for restatements.Not scored
LowCrowded shortNot in a warn state: the crowded-short model only publishes when short interest and thin liquidity coincide, and neither condition is met here.Not scored
Signal
2 of 8 surfaces clear.
8 surfaces checked, 2 clear. The binding constraint is position size — the tail cap is the one that binds. Each row states which surface it is and what it measured; a surface that could not be scored says so rather than counting as clear.
Who should skip it
Skip it if any of these is you
You'd be buying into earnings
A report lands in 40 days.
Our track record0 resolved
Track recordBUILDING IN THE OPEN
12 logged
12 calls logged on this name, 0 resolved so far. Calls are logged the moment they are made, so the open ones are public before anyone knows how they turn out, and a hit-rate is not published until enough of them resolve.
Calls on HVTTHIS NAME
12
This name's own call count. Too few have resolved to publish a rate over them, so none is shown rather than a rate over 0.
0610-K disclosed linksHOME IMPROVEMENT RETAIL
The wider context
5 mapped peers, 3 connections.
↳ Plain EnglishCompanies do not trade alone. This is the peer set, the supply chain that has to deliver, and the macro variables that move the stock regardless of company news.
Rate sensitivity
-39.0% per +1ppSECTOR
This is the SECTOR prior, not a measurement of this name — the regression did not clear its own bar. The verdict beside it (headwind) comes from the prior too, so the two agree by construction rather than by evidence.
Versus peersLatest session
Peer set · 5 names
SymbolRelationshipPriceSession
HDPeer$299.98-0.84%
LOWPeer$192.49-0.88%
FNDPeer$47.28-0.27%
HVT-APeer$27.25-5.61%
TTSHPeer$2.25-1.32%
The mapped peer set and how each traded in the latest session. A peer moving with this name says the move is the sector's; one moving against it says the move is this company's.
Time machineHistory from 2023-06-01
Signal
3 of 4 backward windows are positive.
Look-back returns describe what happened, not what is likely. They are here to set the scale a future move would be measured on.
Macro driversRolling window
Next macro catalystIN 7 DAYS
Consumer Sentiment
2026-09-25. A scheduled release that moves this stock independently of anything the company does.
Events · what moves this stockSep 11, 2026
Retail salesUsually helps when shoppers spend more than expected
This report is the sector's own sales figure, printed monthly.
Typical for the Consumer Cyclical sector
On the day 0.0% · A week later -0.7% · Two weeks later -0.7% · Rose on 18 of 39 report days
Across the last 39 Retail sales reports since Jun 15, 2023, HVT moved 0.0% on average on the day and was 0.7% lower two weeks later. The S&P 500 moved +0.1% on those days.
Initial jobless claimsUsually hurts when more people file for benefits than expected
Fewer paychecks means less spending at stores, restaurants and dealers.
Typical for the Consumer Cyclical sector
On the day 0.0% · A week later +0.1% · Two weeks later +0.1% · Rose on 91 of 168 report days
Across the last 168 Initial jobless claims reports since Jun 8, 2023, HVT moved 0.0% on average on the day and was 0.1% higher two weeks later. The S&P 500 moved 0.0% on those days.
Housing startsUsually helps when more homes are started than expected
New homes get fitted out.
Typical for the Home Improvement Retail industry
On the day -0.5% · A week later -0.4% · Two weeks later -1.3% · Rose on 18 of 39 report days
Across the last 39 Housing starts reports since Jun 20, 2023, HVT moved -0.5% on average on the day and was 1.3% lower two weeks later. The S&P 500 moved 0.0% on those days.
Industrial productionUsually helps when factory output comes in stronger than expected
A stronger economy means more consumer spending.
Typical for the Consumer Cyclical sector
On the day 0.0% · A week later -0.7% · Two weeks later -0.7% · Rose on 18 of 39 report days
Across the last 39 Industrial production reports since Jun 15, 2023, HVT moved 0.0% on average on the day and was 0.7% lower two weeks later. The S&P 500 moved +0.1% on those days.
Producer prices (PPI)Usually hurts when producer prices run hotter than expected
Higher prices squeeze household budgets and higher rates make borrowing dearer; both are bad for stores, restaurants and carmakers.
Typical for the Consumer Cyclical sector
On the day +0.2% · A week later +0.9% · Two weeks later +0.5% · Rose on 20 of 39 report days
Across the last 39 Producer prices (PPI) reports since Jun 12, 2023, HVT moved +0.2% on average on the day and was 0.5% higher two weeks later. The S&P 500 moved +0.3% on those days.
Consumer sentiment (UMich) is due in 6 days.
Consumer sentiment (UMich)Usually helps when consumers feel better than expectedPeople spend more when they feel confident.
Typical for the Consumer Cyclical sector
On the day +0.1% · A week later -0.8% · Two weeks later +0.2% · Rose on 21 of 38 report days
Across the last 38 Consumer sentiment (UMich) reports since Jun 26, 2023, HVT moved +0.1% on average on the day and was 0.2% higher two weeks later. The S&P 500 moved 0.0% on those days.
Core PCE (the Fed's gauge) is due in 11 days.
Core PCE (the Fed's gauge)Usually hurts when inflation runs hotter than expectedHigher prices squeeze household budgets and higher rates make borrowing dearer; both are bad for stores, restaurants and carmakers.
Typical for the Consumer Cyclical sector
On the day -0.7% · A week later -1.2% · Two weeks later -0.6% · Rose on 12 of 38 report days
Across the last 38 Core PCE (the Fed's gauge) reports since Jun 28, 2023, HVT moved -0.7% on average on the day and was 0.6% lower two weeks later. The S&P 500 moved +0.2% on those days.
Core inflation (core CPI)Usually hurts when inflation runs hotter than expected
Higher prices squeeze household budgets and higher rates make borrowing dearer; both are bad for stores, restaurants and carmakers.
Typical for the Consumer Cyclical sector
On the day +0.1% · A week later +0.9% · Two weeks later +0.4% · Rose on 19 of 38 report days
Across the last 38 Core inflation (core CPI) reports since Jun 12, 2023, HVT moved +0.1% on average on the day and was 0.4% higher two weeks later. The S&P 500 moved +0.3% on those days.
Customer momentum
Data unavailableCustomer momentum needs listed customers that disclose a revenue dependence on this company, and none is on file. The connections below are a different map: competitors and partners, most of them curated rather than filed.
Connections3 links · none confirmed
Direct connections · 3 edges
CompanyRelationSessionProvenance
FDX · Suppliersupplier-0.73%Rumored · manual
AMZN · Customercustomer+1.00%Rumored · manual
TGT · Competitorcompetitor-1.03%Rumored · manual
0 of 3 edges are named in a filing; the rest come from reporting we have not matched to one yet, and are labelled as such.
Glossary · every term on this page
Survival
Altman Z
Bankruptcy-risk score. Above 3.0 is the safe band.
Piotroski F
Nine pass/fail health tests; 7 or more is strong.
Beneish M
Accrual and growth pattern screen. Clear below −2.22, watch to −1.78, elevated above; a year of very fast revenue growth lifts it by construction.
Merton DTD
Standard deviations of asset value between the company and its debt.
EDF
Default frequency calibrated on what actually happened to similar names.
Price & tape
RSI-14
Momentum gauge 0–100. Above 70 is overbought, below 30 oversold.
ATR
Average true range — the size of a typical day.
VWAP
20-day volume-weighted average price — the average paid over the last month; the page shows no intraday VWAP.
Supertrend
A trend-following stop line. A close through it flips the read.
Chandelier
A trailing stop set a few average ranges below the recent high.
Options & flow
Gamma / GEX
Dealer hedging pressure. Negative gamma amplifies moves both ways.
Call / put wall
The strikes with the heaviest open interest; they act as ceiling and floor.
Max pain
The strike where most options expire worthless.
13F · Form 4
Institutional filings, 45 days after quarter end. Insider trades, within two days.
Scoring
Composite signal
Independent factors scored and netted into one number.
Hit-rate
Share of past calls that resolved in the direction stated.
Brier score
Calibration measure. Lower means stated probabilities track reality.
HHI
Concentration 0–10,000. Above 2,500 means one segment carries the company.
EVT
Fits the tail rather than assuming a bell curve, so bad days are not understated.