How can you predict which multi-fam building will actually sell

[Thomas Thuillier]

For those of you buying distressed buildings: more violations, more fines, more liens = more motivated seller, right?

We tested this. We tracked every multifamily building in New York City over 30 months. 172,000 buildings. 22 public data sources. 37M+ parameters. We labeled every building that changed hands and compared their characteristics to buildings that didn't sell on a Point-in-Time basis.

The result was not what we expected.

**Deeply distressed buildings almost never sell.** In 30 months, only 31 buildings in severe distress actually transacted. Most are stuck. Litigation holds them in place. Liens block clean title. Enforcement actions freeze everything. The owner can't sell even if they want to.

**Early-stage distress is where the deals are.** 1,799 buildings sold from a "watch" state. These are buildings where problems are accumulating but haven't yet triggered enforcement. The ratio: 58 to 1.

For every deeply distressed building that trades, 58 buildings in early distress change hands.

The profile of buildings that actually sell looks like this:
- Small (median 6 units, not 50+)
- Lower assessed value (around $500K, not $5M)
- Recent mortgage (owner bought within 2 years and is cutting losses)- Brooklyn heavy (55% of early distress sales happen in Brooklyn)- Violations accumulating but no litigation yet

Compare that to deeply distressed buildings that DON'T sell:
- Larger (median 36 units)
- Higher value ($1.5M assessed)
- Older mortgage (owner has been fighting for years)
- Active litigation blocking the transaction

The takeaway for investors: stop competing over the 31 deeply distressed buildings everyone knows about. Start tracking the 1,799 buildings where an owner recently bought a problem and complaints are accelerating. That's your deal flow.