Did my buyer's agent just save me from a deal or kill a good one, still not sure which

The probate comp is a real issue and dropping it is defensible, but I'd want to know what the remaining comps look like without it before you decide she's wrong. One bad data point can skew a small sample hard, especially in a pocket where only four or five things sell in a quarter.

What's the tenant situation actually look like - month to month, lease in place, how long have they been there?

The probate comp question is the right one to press. Probate sales frequently transact below market because estate executors prioritize certainty and speed over price optimization, and including one without adjustment in an ARV stack pulls the number in a direction the open market would not reproduce. That is a legitimate methodological objection, not wishful thinking on your part.

But the assumption doing the most work in your original 14% is the ARV itself, and if the seller's agent sourced those comps, you should treat them as advocacy, not analysis. Agents representing sellers do not pull comps to be neutral.

The face she made is worth taking seriously. You described the Peoria flip yourself. The question to ask is whether your objection to the probate comp is analytically grounded or whether it is the move you make when a number does not produce the return you wanted.

Here is how I would separate those two things. Pull the October probate comp and look at its condition at sale, days on market, and whether there was a buyer concession. If it closed significantly below the other comps without a clear condition explanation, filtering it out or adjusting it is defensible. If it was in comparable condition and still went low, that tells you something about that pocket's buyer pool in that season, and your agent may have caught a real signal.

The unverified tenant income is the risk you did not fully price. If you cannot confirm rent, you are underwriting a vacancy assumption whether you intend to or not. That affects your actual hold-period return independently of the ARV argument.

The roof cost is a known variable. What is your current estimate on that, and did your 14% figure include it at full replacement cost or at a deferred-maintenance discount?

Seven months is long enough that a probate comp from October is borderline usable anyway, especially in Sacramento where Q4 last year was soft and Q1 picked up. That one comp pulling your ARV down 18k suggests it wasn't a small outlier, it was doing a lot of work in her model. If her two comps averaged to land you there and one of them is a probate from a pocket that runs 8-12% below arm's length, you can almost do the math on what her actual adjusted number looks like without it.

The thing I have never seen work is negotiating the ARV with the seller's agent's comps as your starting point and then splitting the difference. What happens is you end up buying at a price that required a rosy number to pencil, and six months in you're refinancing into equity that isn't there. Your agent's face might be the right read, not because she's anchored, but because you said the tenant income is unverifiable, and that detail alone changes the hold math in ways the ARV argument kind of papers over. What's the rent on that unit versus market, and did anyone pull an estoppel or is this all verbal from the seller?

What market is this in Sacramento exactly, because that city has pockets where October sales genuinely run 12-15% below spring and pockets where seasonality barely shows up in the data.

The face she made is doing a lot of work in your telling of this story, and I think you know that.

Unverifiable tenant income on a roof job at 9% is three separate problems before you even get to the comp dispute, so even if she's wrong about the probate sale dragging value, you're still buying a complicated situation on thin margin.

What's the tenant situation, month to month or is there a lease you can actually read?

The face she made is the part I keep coming back to, because a good agent flags the bad comp and fixes it, they don't make a face.

What is the roof quote, because that number swings this more than the 18k ARV gap does.

Confirm with whom the probate comp closed and at what discount before anything else.

A title company or the county recorder can pull the deed transfer and sale price in a day, and if the probate sale came in more than 10 to 12 percent below the median for that pocket, most appraisers would exclude it from an ARV bracket on a standard sale. That is the number that decides whether her sheet holds. The October date is less important than the sale type, and you already know that, so the question is whether she filtered for it or just sorted by proximity and date. Ask her directly: was that comp filtered by sale condition, or pulled raw from the MLS? Her answer will tell you whether the face was judgment or defensiveness.

The tenant income piece is the separate problem your numbers have not priced yet. An unverified tenant on a property with a deferred roof is not a 9 percent deal and it is not a 14 percent deal, it is a deal where the actual hold cost is unknown, and neither ARV matters until you know whether that tenant can stay and on what terms. What does the lease say about verification rights on renewal?

The unverifiable tenant income is doing more work in this decision than the ARV gap is, and I don't see you talking about it much.

If she can't show a lease or bank statements you're essentially buying a vacancy with extra steps, and 9% on a guess at occupancy is not 9%.

What's the roof quote looking like, because that plus an unverified tenant is where I'd be focused before I even got back to arguing comps.

The probate filter matters, but the unverifiable tenant income is what actually breaks the underwriting at 9%, because any lender running DSCR will haircut or exclude that rent entirely, and if you're paying cash the income you're underwriting may not survive a lease audit at closing.

Ask her to pull the October non-probate sale's price per square foot against the seller's comps and see where the gap actually lives. If the spread is tight, the probate sale is doing most of the dragging and her ARV deserves a revision. If the spread holds without it, she's right and you're anchoring to a number that felt good first.

What's the verified rent on the other units, if there are any, and is the unverifiable tenant on a lease or month-to-month?

The comp question and the tenant question are separate problems, and collapsing them is costing you clarity.

Take a property at a $300k ARV with a roof replacement at roughly $12k and a tenant whose income you cannot document. At 14% return on the seller's comp set, you are probably working with a purchase price somewhere around $260k to make that pencil. Drop ARV to $282k using your agent's number, hold the purchase price, and the return compresses to roughly 9%, which is where you landed. But here is what that spread is actually telling you: the two comp sets imply a $18k disagreement on value, and the entire gap between a workable deal and a dead one lives inside that $18k. That is a fragile deal regardless of which ARV is right.

The probate filter matters, but the tenant situation is the thing that changes the character of the risk. An unverifiable income tenant on a property where the ARV is already contested means your stabilized value assumption has two soft legs. Lenders using rental income to qualify you on a small multifamily will want documented rent, and if the appraisal comes in closer to your agent's number, you could face a financing gap that the seller will not bridge.

Ask for the full comp sheet with closed dates, price per square foot, and days on market for every sale she used, not just the two that pulled the value down. A probate sale in October at a depressed price is a legitimate exclusion if you can show it transacted below market and document why. Has the tenant been there long enough that you can pull a payment history from the current owner, even informally?