Research › Hypotheses Tested
Artul.ai Research LibraryStudy No. 55Hypotheses TestedUpdated 2026-08-28

A Secure Base, Several Live Doors, and No Free Lunch: A Call Study

By Artul.ai Research Group · n = 92 earnings calls · First published 2026-08-28
Abstract

This study profiles 92 earnings calls from 2015-2024 that answered YES to the hypothesis 'Secure base, several live doors' out of a 342-call corpus (26.9%, 95% CI 22.5%-31.8%). These calls show modestly more confident language than the baseline: confidence 7.80 vs 7.32, specificity 7.83 vs 7.63, and stress 1.79 vs 2.40. Management teams raised guidance 33.7% of the time vs 22.5% in the base. Overrepresented motifs include 'Early Products Growing Fast' (1.57x) and 'Skeptic Reassured' (1.34x); underrepresented motifs include 'Pricing Recovering' (0.60x) and 'The Question Left Hanging' (0.48x). Among 44 calls with measured post-call returns, the median was -9.3%, essentially matching the baseline median of -10.9%, with 36.4% beating versus 37.6% in the base.

Key findings
  • The pattern matched 92 of 342 calls (26.9%, 95% CI 22.5%-31.8%) across 2015-2024.
  • Guidance was raised on 33.7% of these calls versus 22.5% in the base, while 51.1% maintained guidance.
  • Language deltas favor the pattern: confidence 7.80 vs 7.32, specificity 7.83 vs 7.63, and stress 1.79 vs 2.40.
  • Post-call returns for 44 matched calls show a median of -9.3% versus -10.9% in the base, with 36.4% beating versus 37.6%.
  • 'Early Products Growing Fast' appears 1.57x more often than baseline, while 'Pricing Recovering' appears 0.60x as often.

1Introduction

Analysts who follow earnings calls often hunt for a particular texture: a company standing on firm ground while multiple growth options stay open. Calls with this 'secure base, several live doors' quality are easy to root for, which makes them worth studying rather than assuming. If such calls carry distinctive language, guidance behavior, or motif patterns, that tells us something about how this posture shows up in the transcript record. This study examines 92 calls from 2015-2024 that matched the hypothesis, drawn from a 342-call corpus, and compares their language profiles, guidance actions, recurring motifs, and measured post-call returns against the rest of the sample.

2Data & methodology

The corpus comprises 342 earnings-call transcripts published between 2015 and 2024, each scored independently by a large language model on an identical 37-field battery: seven categorical business verdicts, eight 0–9 behavioral meters, and twenty yes/no judgments. The study group is defined as calls that answered YES to the research hypothesis "Secure base, several live doors" (n = 92; 26.9% of the reference set, 95% Wilson interval 22.5%–31.8%). Baseline figures use the set of calls on which this question was tested. Market outcomes join a fixed sample of 22,449 calls with twelve-month total returns in excess of SPY, measured from the first close after each call; this sample skews toward liquid U.S. names and is reported as descriptive history only.

3Results

The pattern's calls read differently: confidence scored 7.80 versus 7.32 in the base, specificity 7.83 versus 7.63, and stress 1.79 versus 2.40, with evasion slightly lower at 2.45 versus 2.70. Guidance behavior skews constructive: 33.7% raised guidance versus 22.5% in the base, and no calls in this group withdrew guidance, against 1.2% in the base. Motif lifts are notable: 'Skeptic Reassured' appears on 94.6% of these calls versus 70.8%, a 1.34x lift, and 'Early Products Growing Fast' on 56.5% versus 36.0%, a 1.57x lift. The annual share of matching calls ranges from 0.17 in 2015 to 0.00 in 2020 and 2025, with no steady direction. Measured returns do not separate the groups: median -9.3% versus -10.9%, and beat rates of 36.4% versus 37.6%.

Table 1. Mean behavioral scores (0–9 scale), study group versus baseline
MeterStudy groupBaselineΔ
Candor6.966.97-0.01
Evasion2.452.70-0.25
Specificity7.837.63+0.19
Stress1.792.40-0.61
Promotion5.235.06+0.17
Confidence7.807.32+0.48
Table 2. Guidance actions, study group versus baseline
ActionStudy groupBaseline
Raised33.7%22.5%
Maintained51.1%51.8%
Lowered9.8%12.6%
Withdrawn0.0%1.2%
Table 3. Co-occurring battery signals ranked by lift (group prevalence ÷ baseline prevalence)
SignalLiftIn groupBaseline
Early Products Growing Fast1.57×56.5%36.0%
Skeptic Reassured1.34×94.6%70.8%
The Question Left Hanging0.48×21.7%45.0%
Results Worse Than Direction0.50×23.9%48.2%
Pricing Recovering0.60×10.9%18.1%
Volume About to Step Up0.67×17.4%26.0%
Underused Fixed Costs0.67×27.2%40.6%
20150.17%
20160.09%
20170.10%
20180.07%
20190.01%
20200.00%
20210.04%
20220.10%
20230.06%
20240.04%
20250.00%
Figure 1. Share of all analyzed calls matching the study definition, by year.
Table 4. Twelve-month excess total returns versus SPY (descriptive history, not a signal)
StatisticStudy groupReturns sample
Median excess return-9.3%-10.9%
Interquartile range-28.2% to +6.8%
Share beating SPY36.4% (95% CI 24%–51%)37.6%
Observations44133
Table 5. Most recent calls matching the study definition
TickerQuarterCall dateCall grade
WRBYQ1 20242024-05-09A
LINCQ1 20242024-05-06B+
ROCKQ1 20242024-05-01B+
LTRXQ3 20242024-04-29C
SNVQ1 20242024-04-18B
NICEQ4 20232024-02-22B+
METQ4 20232024-02-01B+
BRBRQ4 20232023-11-21B+

4Discussion

A careful reader should conclude that calls matching this hypothesis have a recognizable linguistic signature: more confident, more specific, less stressed, and more likely to raise guidance. They also talk about early products growing fast and manage to reassure skeptics. What should not be concluded is that this pattern causes better outcomes or predicts them. The measured returns are statistically indistinguishable from the baseline, and the motif shares and language scores are descriptive comparisons within one corpus. The guidance differences are associations observed after the fact, not evidence of an edge, and the returns sample covers only 44 of the 92 matched calls.

5Limitations

All fields in this study are AI-read from transcripts and are noisy measurements of tone, not ground truth. The returns sample of 44 matched calls and 133 baseline calls is drawn from a broader universe of 22,449 calls skewed toward liquid names, so small-sample medians and beat rates are imprecise. Our own forward tests falsified directional prediction for these patterns. LLMs partially remember famous stocks' histories, which contaminates language scores and any backtest-style comparison. Motif lifts describe this corpus only and should not be extrapolated to future calls or other datasets. See the full methodology, including the C1 pattern’s forward-test failure and the LLM-memorization finding.

Companion page: every company matching this hypothesis is listed at the question’s own page.

Cite this study Artul.ai Research Group (2026). “A Secure Base, Several Live Doors, and No Free Lunch: A Call Study.” Artul.ai Earnings-Call Research Library, Study No. 55. https://artul.ai/research/hypothesis-secure-base-several-live-doors

Related studies

The Numbers Are Fine; Everything Else Is Pending: FlRoom to Run and Nowhere to Hide: Calls With UncontesLight at the End of the Tunnel Is Often a Train: RecDon't Stick to the Script: Earnings Calls Where AnswHave Your Cake and Expand the Base Too: Repeat GrowtPressed Harder, Answered Straight: A Profile of 183
Not investment advice. Artul.ai publishes AI-generated earnings-call quality grades and expected-volatility estimates — never buy or sell recommendations. We tested over 1,600 predictive hypotheses against 165,000 transcripts; the honest result, including what failed, is documented in our methodology.