All ideas below are paper trades — tracking signals, not deploying capital.
Credit markets this week offered a deceptively calm surface. With no macro data in the feed, the signal comes from the filings themselves: seven distinct borrower clusters flipped to non-accrual in a single cycle, spanning first lien, subordinated debt, preferred, and equity instruments. That breadth — not any single name — is the story.
Non-accrual clustering at this density (15+ filing-level events resolving to roughly nine underlying borrowers) typically precedes NAV markdown cycles 60–90 days out. The `v3_borrower_stress` signal, which has fired 65 times with a 46.2% hit rate and a mean abnormal return of −13.05% over 60 days, remains the highest-magnitude signal in the scorecard this week — directionally bearish with real teeth. The `exit_vs_peer_contradiction` signal (70% hit rate, +3.3% avg abnormal return) is the lone constructive setup in the book. Spreads are not in the feed, but when first-lien paper at multiple BDCs stops accruing simultaneously, the smart prior is that marks haven't caught up yet.
Runway Growth Finance disclosed three Blueshift Labs positions flipped to non-accrual: a December 2023 senior secured loan (SOFR+8.25% PIK, 13.25% floor), a December 2024 revolver, and an associated warrant. That's first-lien debt originated across two separate tranches plus equity — a full stack impairment on a single application-software borrower. The combination of PIK structure and full-stack non-accrual is a red flag for realized loss rather than temporary dislocation. This is proven wrong if Blueshift completes a refinancing or sale process within 60 days that clears par. RWAY profile →
Fidus Investment disclosed Virtex Enterprises (component manufacturing, subordinated debt, S+4.00%, 2.50% floor, 9.77% cash rate, maturity 3/31/2027) moved to non-accrual. Sub debt non-accruals at BDCs almost always trail a first-lien impairment that was either held elsewhere or quietly marked down in prior quarters. The 2.50% SOFR floor and modest spread suggest this was originated as a "safer" credit — which makes the flip more surprising and the eventual mark more punishing. Thesis fails if Virtex cures within one quarter via asset sale or sponsor equity injection. FDUS profile →
Oaktree Specialty Lending's non-accrual on Senior Loan Fund JV I, LLC (Multi-Sector Holdings, sub debt) is structurally distinct from a direct borrower flip — JV vehicles carry leverage inside the vehicle, meaning losses can be amplified relative to the face value of OCSL's investment. JV impairments are notoriously slow to resolve because they require co-manager coordination. At book equity of $100,293 in the platform portfolio, any NAV erosion at OCSL at current levels has limited cushion for multiple simultaneous JV markdowns. This is wrong if the JV's underlying loans are already marked and the non-accrual is a reclassification rather than new impairment. OCSL profile →
ACI Group Holdings, Inc. generated five distinct non-accrual events this cycle: First Lien 1, 2, 3, Drawn revolver, and a preferred position held by a separate reporting entity. Multi-tranche, multi-instrument non-accruals on a single borrower that show up across what appear to be separate BDC filers is the textbook signal for a club deal gone wrong — meaning the loss is syndicated across multiple vehicles. Identifying which BDCs hold ACI and sizing their exposure relative to NAV is the trade here. This resolves as a non-event only if ACI is already fully reserved in prior marks. ACI research →
Streamland Media Midco LLC (1st Lien/Last-Out Unitranche, S+6.50% including 5.50% PIK, maturity 04/02/29) flipped to non-accrual. The structure is notable: a unitranche with a PIK component that large implies the borrower was already cash-constrained at origination — 5.50% PIK in a first-lien is a yield-enhancement structure, not a standard term loan. Last-out positioning within a unitranche means recovery in a restructuring is subordinated to any "first-out" slice. The combination of PIK dependency, last-out position, and entertainment sector cyclicality makes this a high-impairment-probability credit. Wrong if Streamland closes a media asset sale that restores cash flow coverage. Streamland holder search →
The `debt_to_equity_conversion` signal was passed on entirely this week — 7 events, 28.6% hit rate, −2.3% average abnormal return is a losing trade on both frequency and magnitude. The `exit_at_loss` signal (168 events, 49.4% hit rate, +0.18% avg return) is essentially noise at current resolution; we will not trade coin flips. Showing restraint when the scorecard itself says "don't" is the discipline.
Next week, the question is whether Q2 filing season accelerates this non-accrual wave or reveals that marks have already absorbed it — NAV revisions across RWAY, FDUS, and OCSL will be the tell.
Screening research, not investment advice. All data from public SEC filings.