The free sample brief

You shouldn’t have to trust a sales page — so here’s the whole thing. This is the free sample brief, built from a real, closed Riverside County sale (TC-223, auction April 28, 2026) — nobody can bid on it, and we have nothing to sell you against it. Everything below is the actual output of our pipeline on the real 946-parcel list. The only thing missing versus a live-sale analysis is that this sale is over. The format, the rigor, and the honesty are identical.

The full machine-readable shortlist (all 175 survivors) is the companion file riverside-tc223.csv. Every number on this page traces back to one pipeline run — filter-config hash 28c23eb7243637fe, source PDF SHA-256 on file.

1. The reduction: 946 → 175

The county published 946 parcels. Six deterministic rules (string matches, threshold checks, a blocklist lookup — no AI, fully reproducible) dropped 771 of them. Each drop records exactly one reason, so it sums cleanly:

One filter did almost all the work. “Does this parcel have a street address?” removed 687 parcels by itself — about 89% of the entire cull. The most valuable rule in the stack is also the dumbest one. Anyone selling you a neural network to find paper lots is selling you a neural network to read a field that already says SITUS ADDRESS: NONE.

The 175 survivors carry minimum bids from $10,106 to $195,409 — median $25,082, average $35,985. That’s a set a human can actually work through.

2. What the survivor set looks like

A brief doesn’t just hand you 175 rows — it tells you the shape of what survived so you know where to spend attention.

By minimum bid: the median ($25,082) sits well below the average ($35,985), so the set is right-skewed — a cluster of modest openers around $25K with a tail of larger ones pulling the mean up. The normal shape of a residential-leaning survivor set.

By city (top of the spread): Riverside (19), Hemet (16), Menifee (11), Perris (10), Palm Springs (9), Corona (8), Moreno Valley (8) — plus 20 parcels where the county PDF didn’t print a parseable city (they render as ”—”; their street address survived the address filter, but the city field didn’t extract — a thing to confirm on the county GIS, not a defect in the parcel).

By assessed-owner type (we publish the type, never the name — the pattern is what matters, not the person): individual 84, family-group 49, trust 31, LLC 6, estate 4, government 1.

By risk flag — and here is the single most important honesty caveat in this brief:

Why three of the six flags are empty: they’re derived from property data (an automated valuation, last-sale date, square footage) — and we have only run that enrichment on 5 of the 175 parcels. A flag that needs a Zestimate can’t fire on a parcel that doesn’t have one yet. So the three structural flags that come straight from the parcel record (Mello-Roos from the TRA code, commercial from the owner name, bulk-owner from the sale-wide count) are populated; the three that need enrichment are not. On a live paid brief we run enrichment across the full surviving set first, so all six flags are live. We’re showing you the half-finished state truthfully rather than implying a complete valuation pass that didn’t happen.

3. Per-parcel reasoning (the part you’re actually paying for)

This is what a brief is: not a verdict, but a short, honest read on a parcel and the specific next thing to check. We never say “bid.” We say “worth a closer look” or “probably skip — here’s why.”

3a. The five we could value — and why “too good to be true” is the correct instinct

These are the only 5 survivors currently carrying a real automated valuation (AVM). Read what the naive version of this looks like, then read the honest version.

The naive read: “A $32,000 bid on a $704,000 house? That’s 95% off. Back up the truck.”

The honest read — and this is the whole point of the product: a bid-to-value ratio this low is either the genuine tax-sale value proposition (the min bid is just the back taxes, which on a long-defaulted property can be a small fraction of value) or a flashing red light. You cannot tell which from the number. Before you believe any of these, verify:

• Is the AVM even this parcel? AVMs are matched by address/APN geocoding, and the match is wrong more often than vendors admit. Item 6 (Eastvale) is the live example: it reports year built 1920 for a parcel in Eastvale, a city that didn’t incorporate until 2010. A 1920 structure there is implausible — that AVM almost certainly matched the wrong record, which means its $155,300 value (and the tidy 11.7% ratio) is untrustworthy until you confirm the address on the county assessor’s record. Treat the other four AVMs with the same suspicion until verified.
• Is it fee-simple title? Tax sales sometimes convey a fractional or possessory interest, not the whole fee. A “95%-off house” that’s really a 1/6 undivided interest is not 95% off.
• What survives the deed? A senior federal tax lien, a surviving special assessment, or an occupancy/eviction problem can erase the apparent margin. Pull title before you fall in love with a ratio.

Net: these 5 are the first parcels to research, not the first to bid on. A 4.5% ratio earns a title order and an assessor-record check — that’s the work. (Item 4 is the one with no Mello-Roos flag — its TRA is baseline — so it carries one less thing to confirm than the others.)

3b. The commercial-environmental wildcard

Item 1 · APN 102083027 · Corona · min bid $19,532 · owner type: LLC · flags: COMMERCIAL_RISK, MELLO_ROOS_POSSIBLE

The owner is an LLC, which is why this carries COMMERCIAL_RISK. Commercial and industrial parcels are the trap that turns a $20K bid into a six-figure loss, because a tax deed does not wipe out CERCLA environmental liability — it attaches to you as the new owner by federal statute, not as a recorded lien the sale extinguishes. Before bidding on anything with a commercial fingerprint, the move is a Phase I Environmental Site Assessment (~$3,000): a records review that also helps preserve the bona-fide-prospective-purchaser defense. $3,000 to de-risk a $20,000 bid is cheap insurance; if you’re not willing to spend it, you’re not yet capitalized for commercial tax-deed parcels. (This is also one of the 6 parcels where, on a live brief, you’d expect the “thin Zillow data” caveat — commercial parcels return sparse valuation coverage by their nature, not by a bug.)

3c. The bulk-owner edge

Item 673 · APN 639272007 · Desert Hot Springs · min bid $25,082 · owner type: individual · flags: MELLO_ROOS_POSSIBLE, BULK_OWNER_NEAR_THRESHOLD

This owner appears on several parcels in this sale — close to, but under, the threshold where we’d auto-drop the whole cluster as speculator inventory. Desert Hot Springs is exactly the geography where the post-2021 cannabis-zoning speculation collapse left clusters of distressed lots. The flag isn’t “skip” — it’s “look at this owner’s other parcels before you bid on this one.” If the rest of their holdings are no-address desert lots, you’re likely looking at the edge of a failed-thesis cluster and this parcel inherits that context.

3d. The Mello-Roos overlay (170 of 175 parcels)

MELLO_ROOS_POSSIBLE fires on 170 of the 175 survivors — because in this part of Riverside County a non-baseline Tax Rate Area is the norm, not the exception. Do not read this as “170 parcels have Mello-Roos.” The TRA code encodes the combination of taxing agencies on a parcel; a non--000 suffix can mean a Mello-Roos CFD, but it can equally mean a different school district, a 1915-Act assessment, a fire/water/lighting district, or an annexation boundary — with no special tax at all. It’s a “confirm the specifics on the secured tax bill” flag, not a detector. The dollar amount is never derivable from the TRA code alone. On a $700K house a Mello-Roos overlay might be a rounding error; on a thin-margin rental it can wreck the carry. Confirm before you set a maximum bid — don’t skip on the flag alone.

3e. The 170 we have not valued yet (structural-only reads)

For most of the survivor set we currently have only the structural signals — bid, situs city, owner type, TRA flag. That still tells you something. Two honest examples:

• Item 6’s cohort in established cities (Riverside, Hemet, Menifee) with individual/family owners and modest bids around the $25K median are the “ordinary residential default” backbone of the list — the candidates most likely to reward a county-GIS + comp pass. They have no red flags because we haven’t looked hard enough yet, not because they’re cleared.
• The 20 “city = —” parcels had a street address (they passed the address filter) but the city didn’t parse from the PDF. Before treating any of them as a candidate, confirm the situs on the county GIS — the missing city is a parsing gap to close, and occasionally a sign the address line itself is unusual.

On a live paid brief, these 170 are where enrichment runs and the per-parcel notes get their value layer. Here, we’re being straight that the work isn’t finished.

4. What this brief does not establish (and what’s still on you)

Survival means a parcel cleared the structural filters. Nothing more. We did not pull title, did not inspect anything, did not enrich most of the set, and we never tell you what to bid. Before bidding on any tax-sale parcel — these or any others — the standing checklist applies in full:

• Pull the parcel on the county GIS; confirm zoning, lot size, frontage, and that the situs is what the AVM thinks it is.
• Order a title chain and look for liens that survive a tax sale — federal tax liens especially. (In California a tax deed extinguishes the pre-sale HOA back-dues lien, but the recorded CC&Rs carry through, so you owe assessments going forward; Mello-Roos can survive conditionally. The full legal mechanics, checked against primary sources, are in the research library.)
• Get a title report from a service that issues policies on tax deeds.
• Physically inspect, or at minimum review recent satellite and street-view imagery.
• Determine occupancy — the deed is yours; the eviction is still your problem.
• Check environmental records (EnviroStor, GeoTracker) for anything with a hint of commercial use — and run a Phase I before bidding on a commercial parcel.
• Confirm the minimum bid hasn’t changed at the auction platform on the morning of the sale.

The cut gets you from “946 parcels, any of which could waste your money” to “175 that clear the structural bar, here’s the reasoning on each.” The decision stays with you — that’s the only place it belongs.

5. What you’d get on a live sale

This sample is a closed sale, on purpose, so you can judge the work with zero pressure. For an upcoming California county sale, the same pipeline runs — but with enrichment across the full surviving set (so all six flags are live and most rows carry a valuation), published in full and free, before bidding opens. Subscribe and you’ll get the next one the day the county list drops — no paywall, no course.

Download the full machine-readable shortlist: riverside-tc223.csv (all 175 survivors, the columns above).

The Riverside TC-223 sale closed in April 2026 and is presented here as a methodology sample, not a recommendation. LotBrief surfaces signals and explains them; it does not make recommendations, is not a title search, and does not guarantee the county data is correct. Automated valuations are third-party estimates and are frequently wrong — verify everything (title, liens, zoning, occupancy, environmental history) with the appropriate licensed professionals before bidding. We are not licensed real estate brokers, agents, or attorneys in any jurisdiction. Nothing here is legal, tax, or investment advice. The decision is yours.