An auction estimate looks like a price. It isn’t. It’s the one number that trips up almost everyone new to bidding, and misreading it is how people either overpay or walk away from something they could have won. Here’s how to actually read an estimate, why the hammer so often lands somewhere else, and what you really pay once the gavel falls. The examples are real, from Phillips, Koller, and SBI Art Auction.
TL;DR: the whole thing in four lines
- The estimate is a marketing tool, not a prediction. Houses often set it low to pull a crowd, so the hammer can land well above it.
- The hammer price isn’t your total. Add the buyer’s premium, roughly 20 to 27 percent at the major houses, plus tax and shipping, before you set a ceiling.
- Strategize for the specific piece, not the category. Condition, completeness, and provenance decide whether a lot blows past its estimate or lands right inside it.
- If you ask AI to help with a bid, don’t be shocked by a higher number. It’s reading the same public hammer prices or other comps you should be reading, rather than the house’s deliberately low estimates.
Disclosure
I collect art and watches, including work by one of the artists mentioned here (Hiroshi Nagai). I’m not selling any of it, and nothing here is a recommendation to buy. These are teaching examples, and the lessons about estimates apply to almost anything you’ll meet at auction.
Frequently Asked Questions
Why is an auction estimate often lower than the final price?
An estimate is a marketing tool, not a forecast. Houses often set it low to attract more bidders and build momentum, and for a strong or fresh-to-market piece the hammer can land well above the high estimate. An estimate also describes a typical example, so an exceptional specimen can blow past it.
What is a reserve price?
The reserve is the confidential minimum the seller will accept. It usually sits at or just below the low estimate, and it’s never published. If the bidding doesn’t reach it, the lot goes unsold, which the trade calls being “bought in.” So a low estimate doesn’t mean you can grab the piece cheaply; the hidden reserve still has to be met.
What does “estimate on request” mean?
For very high-value lots, a house sometimes prints “estimate on request” instead of a number. It’s not a mystery, just a nudge to contact the specialist directly, usually because the piece is valuable enough that they’d rather discuss it than post a figure. Ask, and they’ll give you a range.
What is the hammer price, and is it what I actually pay?
The hammer price is the winning bid the auctioneer calls when the gavel falls. It isn’t your total. On top of it you pay the buyer’s premium, a fee charged by the house, plus any sales tax and shipping. Always budget for those before you bid.
How much is the buyer’s premium?
It varies by house and price level, but at the major houses it currently runs around 20 to 27 percent of the hammer on the first tier, tapering on higher amounts, and it can be lower at smaller regional houses. Plan on at least 20 to 25 percent on top of any bid.
How do I place a bid at auction?
Most houses let you bid four ways: in the room, live by phone with a staff member, live online through the sale’s platform, or by leaving an absentee (or “commission”) bid in advance that the house executes up to your maximum. You register beforehand either way. If you can’t attend, absentee and online bidding are how most private buyers take part.
Why do two similar pieces sell for very different prices?
Condition, completeness, and provenance. A pristine, full-set example with original papers or an original-owner history can multiply the price of an otherwise identical piece. In thin markets that are still finding a price, the result also depends heavily on who happens to be bidding that night.
What an estimate actually is
Start with what the number is for. An estimate isn’t a prediction of the sale price; it’s a marketing figure. Especially for a desirable lot, houses set the low estimate deliberately low, because a low number pulls a crowd. People see a tempting range, pile in hoping for a deal, and that early energy is the point. It doesn’t set the final price. The real price is set by whoever’s still bidding at the top. So a low estimate tells you the house wants a busy room, not that the piece is cheap.
One more thing sits behind the estimate: the reserve. That’s the confidential minimum the seller will accept, usually set at or just below the low estimate, and you never see it. If the bidding doesn’t reach it, the lot simply doesn’t sell, and the trade calls that being “bought in.” So the low estimate isn’t a floor you can grab the piece at. It’s a lure, and the real floor is a number you can’t see.
The hammer isn’t what you pay
When the gavel falls, the auctioneer calls the hammer price: the winning bid. That’s the headline number you see in results, but it isn’t your total. On top of it you pay the buyer’s premium, a fee the house adds to every sale, and then any sales tax and shipping. At the major houses the premium currently runs around 20 to 27 percent of the hammer on the first price tier, tapering on higher amounts, and it can be lower at smaller regional houses. It adds up fast. On the Rolex below, a $203,200 hammer becomes roughly $258,000 once you add a premium in the mid-20s percent, before any tax. So decide your true all-in ceiling first, then work backward to the highest hammer you can actually afford.
Why results diverge from the estimate: read the specimen
Two lots that look identical on paper can sell worlds apart, and the reason is almost always the specific example in front of you: its condition, whether it’s complete, and where it has been. Two watches make the point better than any rule.
The Rolex that blew past its estimate
At Phillips, a Rolex Day-Date “Malachite” (ref. 18238, circa 1995) carried an estimate of $40,000 to $80,000 and hammered at $203,200. Most examples of this reference trade around $60,000 to $80,000, and the estimate was built around that typical example, which usually shows wear and often has the dial cracking these references are known for. This one was different: fresh to the market (never before offered at auction) from the family of the original owner, a full set with its 1995 warranty and box, and in exceptional original condition with no cracking. The house estimated the type. The specimen was extraordinary, and collectors chased it well past the high estimate.
The AP where the estimate already knew
Compare an Audemars Piguet Royal Oak Day-Date moonphase at Koller. Comparable examples without their original papers, the certificate and warranty that document a watch’s authenticity, change hands around $40,000. This one was in excellent condition, a discontinued reference, and, crucially, came with those papers. It sold near the top of what these command, but that price sat inside the auction’s estimate range. Same story of condition and completeness driving value, but here the house had already priced the premium in, so the result landed within the estimate rather than rocketing past it. The lesson isn’t that estimates are always low. It’s that the estimate reflects what the house thinks the specific piece is, and your job is to check whether they got it right. Watches have their own maze well before any auction; we cover buying at the counter versus the secondary market in our piece on why cash can’t buy you a Rolex.
Thin markets move differently
Everything above assumes a deep market, like Rolex or AP, where comparable sales are plentiful and estimates are well informed. Newer names are a different game. Hiroshi Nagai, the Japanese city-pop painter, has only recently moved into blue-chip territory, and his market is still in price discovery. Look at two of his paintings at SBI Art Auction: both untitled, both from 2021, both acrylic and nearly the same size, both with deliberately low estimates. One hammered at about $26,000. The other, months later, hammered at about $127,000. Same artist, same house, a five-times spread, because the market hasn’t settled and the result came down to who was in the room. For an established name, study the comps, the prices comparable pieces have actually fetched, and you get a tight read. For an artist still finding a price, the comps are a starting point and not much more.
| Lot | Estimate | Hammer |
|---|---|---|
| Rolex Day-Date “Malachite” (ref. 18238), Phillips | $40,000 to $80,000 | $203,200 |
| Nagai, Untitled (2021), SBI, Apr 2025 | ~$7,000 to $10,000 | ~$127,000 |
| Nagai, Untitled (2021), SBI, Sep 2024 | ~$4,000 to $6,000 | ~$26,000 |
| Phillips (The New York Watch Auction: XIV) and SBI Art Auction results. Figures are approximate and are hammer prices; add the buyer’s premium and tax for the true cost. The Audemars Piguet example above sold within its Koller estimate. | ||
If you ask an AI, expect a higher number
Here’s a useful tell. Ask ChatGPT or Claude what a given lot should sell for, and it’ll often quote a range above the auction house’s printed estimate. That’s not the AI being wrong. It’s reading the same public hammer prices you should be reading, rather than the house’s deliberately low come-on. When the machine and the catalogue disagree, the machine is usually closer to what the piece actually costs, hammer plus premium included.
How to set your own bid
1. Start from suspicion, not the estimate.
The estimate is the house’s marketing number, not a valuation, and it isn’t on your side. Treat a low estimate as a lure and a high one as an anchor, then set the printed range aside. Your number should come from the market, not the catalogue.
2. Research comparable sales.
Find what closely comparable examples actually sold for, matched on condition, completeness, and originality. A full set with papers or an original-owner history is worth real money over a bare example, and the estimate may or may not reflect it. Auction result databases and the houses’ own past results are public, and our acquisitions record lists pieces at the prices we actually paid.
3. Set a ceiling you’re comfortable with, all-in.
Decide the total you’re genuinely happy to pay, including the buyer’s premium and any tax, then work backward to the highest hammer you can bid. Do this before the sale, especially if you bid absentee and won’t be awake to do the math live.
4. Hold the line, and be ready to walk away.
A live room and a ticking clock will both push you past your number. The discipline that protects you is simple: set the ceiling in advance, and be willing to lose the lot. There’s almost always another piece. Overpaying because you got caught up is the one mistake you can’t take back, so if the bidding passes your number, let it go.
The bottom line
An estimate is a tool the house uses, not a valuation you should trust. Read the specific piece, price in the premium, and set your ceiling on comps rather than on the catalogue. Do that and auctions stop feeling like a rigged game. Whether it’s a pristine Rolex, a papered AP, or a painter the market is still figuring out, the discipline is the same: know what you’re looking at, and know what you’ll pay before the gavel moves.
Figures here are approximate and drawn from public auction results as of mid-2026. Buyer’s premiums and taxes are extra, markets move, individual results vary widely, and nothing in this piece is investment advice.
New to buying at this level? Our guide on where to start walks through first purchases and real budgets. Or if you’re eyeing a specific lot and want a second read on the estimate before you set your number, comparing notes is free.
Get in Touch