SAFT

Safety Insurance Group, Inc.

NASDAQ · Common stock

104.13

+0.28 (+0.27%)

Close, Oct 20, 2026

Watch

Structural summary

What is the structure, what changed, and what would invalidate it

Viridia Intelligence

Context across dimensions, not a rating

StructureHigher highs, higher lowsIntermediate-degree swings on the daily chart.

WaveUnresolvedZigzag, wave C up (Pattern Confidence 71); the weekly count agrees.

FibonacciNo zoneNo confluence zone near price.

MomentumStrong3 months +43.6% (ahead of SPY by 40.7 points); 6 months +37.0%.

RegimeUptrend+0.5% vs the 50-day average, +24.6% vs the 200-day.

RiskNormal20-day volatility 2% vs 45% over the year; 1-year drawdown −15.2%; beta -0.18.

SAFT setup historyNot replayed yet; the track record fills in most-traded securities first.

Current structure

Price, swings and the preferred count on the chart

SAFT · 1DLoading bars…

SourceMassive, end of dayUpdated—MethodDaily bars, split-adjusted

Candidate interpretations

Every count shown passes every hard rule

Candidate interpretations

Ranked by Pattern Confidence

Counts ending at the latest intermediate swing low, 70.81 on Jul 23, 2026. 64 labelings tested, 54 broke a hard rule, 10 remain.

Hard-rule evidence: why 54 labelings were eliminated

21Wave 2 always retraces less than 100% of wave 1
9Wave 3 always travels beyond the end of wave 1
9Wave 4 always retraces less than 100% of wave 3
5Price after the last labeled point has not crossed a rule-based invalidation level
4Contracting triangle: C ends short of A's end, D short of B's end, and E short of C's end
2Wave B retraces at least 90% of wave A

Method Every chain of 3–6 alternating pivots ending at the latest confirmed pivot is tested as each pattern; only counts with zero hard-rule failures are kept, then ranked by Pattern Confidence. Rules: rulebook.Engine pivots-1.0.0+rules-1.1.0+candidates-1.0.0+fib-1.0.1+rank-1.2.0+setup-1.0.1 (computed on load), bars to Oct 20, 2026

Fibonacci map

Confluence zones above and below the closeClose 104.13

No price band within 35% of the close has two or more independent Fibonacci relationships.

Method Targets come from the Fibonacci relationships the sources describe for the wave in progress, Elliott trend channels, the prior fourth wave of lesser degree, and 38.2–78.6% retracements of the latest swing. A zone is where two or more independent relationships fall within max(¼ ATR, 0.4%) of each other; strength counts them, weighting primary ratios and higher degrees. It is not a probability.

Evidence

The setup the preferred count implies, and how that kind of setup has done

Setup

Intermediate degree · from the preferred count

No scenario at intermediate degree right now

The count has no unreached Fibonacci target beyond the current price. Viridia shows nothing rather than a guess.

NoteResearch scenario generated from Viridia's structural model. Not a recommendation.MethodA setup restates the preferred count as a trade: the entry (the close, or the level where the correction in progress most often ends), the stop where the count is wrong (a rule level, or the end of a finished pattern), and the count's own Fibonacci target. Reward:risk is measured from the middle of the entry range. A setup needs reward:risk of at least 1 : 1. It is research output from the wave count, not a recommendation or a forecast.

Engine details

Analysis engine: pivots-1.0.0+rules-1.1.0+candidates-1.0.0+fib-1.0.1+rank-1.2.0+setup-1.0.1
Bars analysed to: Oct 20, 2026
Price history: Automatic, updated nightly
Daily bars: 516 since Sep 30, 2024
Avg volume, 50 days: 128.5K
Dollar volume, 20 days: $12.3M
SEC CIK: 0001172052
Provider symbols: nasdaqtrader: SAFT · massive: SAFT
Listing changes: None since Sep 24, 2026

SourcePrices: Massive (end of day, split-adjusted). Listings: Nasdaq Trader, SECUpdatedOct 22, 2026, 8:00 AM EDTMethodSecurity master, daily snapshot diff

Structure at a glance

Daily chart, as of Oct 20, 2026

Since the count's last pivot (70.81, Jul 23, 2026) price has moved +47.1%, the way the count expects.10 of 10 rule-valid counts expect the current move to be up.Weekly count: zigzag complete ↑, same direction

Close call.The top two counts are within 1 points. Treat the structure as unresolved until one is invalidated.

Preferred count

Zigzag up · wave C ↑

Price is in wave C of a three-wave zigzag correction up.

A-B from 67.93, Jun 3, 2026

Pattern Confidence71High

3 of 3 checks met

Key levels, preferred count

Invalidation−34.8% from close

67.93

Price below this level breaks the count.

Target, wave C

None

No unreached Fibonacci target.

Nearest confluence zone

None

No zone within 35% of price.

Alternate count

Zigzag down · complete, next move up ↑

A three-wave zigzag correction down looks complete, so the larger trend would normally resume up.

A-B-C from 81.49, Feb 6, 2026

Pattern Confidence70High

5 of 6 checks met

Why this count is preferred · 3 met, 0 not met

Met

Wave B retraces 38.2% to 85.4% of wave A
Wave B retraces under 90% of wave A (deeper B waves fit a flat better)
The count starts at the price extreme of the prior swing (a new trend begins where the old one ended)

Not met

None.

MethodPattern Confidence = (guidelines met + 2) ÷ (guidelines measured + 4), from the rulebook's guidelines, Fibonacci tendencies and wave-personality checks (wave 3 speed and volume, wave 5 volume, alternation in time, a start at the prior extreme), alternate waves in Fibonacci ratio, and the rarity of running flats. The alternate is the best-ranked count that tells a different story. It ranks the rule-valid counts against each other; it is not a probability of any outcome.

Structural summary

Viridia Intelligence

Current structure

Candidate interpretations

Candidate interpretations

Fibonacci map

Fibonacci map

Evidence

Setup

Engine details

Structure at a glance

What changed