All positions below are paper trades tracked in the Crypsos research portfolio.
High-yield OAS held in a tight band of 2.70–2.72% through the week, with the three-month delta sitting at roughly -0.14 — spreads have been compressing, not widening. That is a benign credit environment by any historical measure, and it is the single fact that drove the most consequential decisions this week. The watchlist was the only live channel.
BDC sector behavior mirrors the broader credit regime: no NIM lead-lag signal cleared the -50bps threshold across five or more shared re-marked names, and the v3 borrower-stress channel surfaced zero fresh events in the past ten days. The regime-long strategy — which backtest v4 says only pays when OAS is widening — sat idle all week, correctly. Spread compression at this level tends to compress equity upside too; the trades that make sense here are quality carries, not tactical dislocations.
ARCC — Long
Ares Capital remains the largest and most liquid name in the BDC universe, and the watchlist flagged it on July 13 at a ~$4,997 notional, sized at 5% of equity. The thesis is straightforward: scale, portfolio diversification, and a management team with two full credit cycles of institutional memory make ARCC the default quality hold when spreads are flat. One alert did surface — a bankruptcy docket match on Convey Health Solutions ($1.3M held fair value, S.D. Florida, filed October 2024) and a separate Continental match ($27M, Fieldwood Energy, 2020 vintage) — both are name-matches pending adjudication and the Fieldwood exposure is a five-year-old case. What would break this trade: a sustained move in OAS above 3.5% paired with a meaningful NAV revision at the next filing. Company page
NMFC — Long
New Mountain Finance was added July 16 at ~$4,996 notional. The thesis here is a quality middle-market lender trading at a discount to NAV with stable dividend coverage — but the system is watching this one closely. Coverage came in at 0.953x against a 1.0x threshold, triggering a thesis-break alert this week. That number is not a crisis, but it is below parity, and a second consecutive quarter below 1.0x would prompt a re-evaluation of position size or an outright exit. What would prove it wrong faster: further NAV compression or a coverage ratio that deteriorates below 0.90x on the next filing. Company page
FDUS — Long
Fidus Investment opened July 13 at ~$4,997. FDUS is a smaller BDC focused on lower middle-market companies, which typically means higher yield but also higher idiosyncratic risk. The watchlist thesis here is a quality tilt in a benign spread environment — the name has historically maintained strong coverage ratios and disciplined origination. No borrower-stress alerts or bankruptcy docket matches surfaced for FDUS this week, which is a positive data point. What would break it: any deterioration in its coverage ratio below 1.0x or a cluster of borrower-stress events in its concentrated sector exposures. Company page
MSDL — Long
Morgan Stanley Direct Lending entered the portfolio July 13 at ~$4,997. MSDL's draw is institutional pedigree and a conservatively constructed senior-secured portfolio — relevant when the regime is benign and you want quality carry without reaching down the capital stack. No alerts flagged MSDL this week. The risk is that its premium positioning means limited upside from spread normalization; if rates decline faster than expected, NIM compression could pressure distributions before NAV reflects it. Company page
SLRC — Watch / Flag
SLRC is not a new long — it is a name the system is actively monitoring after a bankruptcy docket match flagged "Equipment Financing" against a Searles Valley Minerals case (D. Delaware, filed June 15, 4 dockets, held fair value $111.4M). That $111M figure is material relative to a mid-sized BDC's book. The match is name-based and not yet adjudicated, but $111M in potentially impaired fair value warrants attention ahead of the next quarterly filing. This is not a short thesis — the scorecard shows v3 borrower-stress at -13.1% abnormal return over 60 days with only 46% hit rate, too unreliable for a systematic short. It is a hold-off-the-watchlist flag until we see how the docket develops. Company page
OBDC was exited clean on July 13 — removed from watchlist, no drama. More importantly, the regime-long strategy generated nine consecutive pass signals across the week, all for the same reason: OAS at 2.70–2.72% with a negative three-month delta simply does not clear the spread-widening threshold that backtest v4 requires. The nim-leadlag and v3-events channels also passed every day. Running three systematic strategies that collectively produced zero trades is not a malfunction — it is the scorecard working as designed. GAIN surfaced a meaningful divergence signal (NAV down 223bps quarter-over-quarter, price-to-NAV up 8.8 points to 1.004x) and was passed; that kind of mean-reversion setup requires a short framework the current book does not support.
The forward question for next week is whether the July BDC filing cycle — which will update coverage ratios and NAV marks for names including NMFC — confirms the coverage softness already flagged or stabilizes it.
Screening research, not investment advice. All data from public SEC filings.