He was not angry when he said it. That was the striking part.

The buyer who brought this to us is 41, a civil engineer, married with two children, and looking for a family home in Sydney's inner west that he intends to hold for twenty years. He is methodical to the point of being slightly apologetic about it. He keeps a spreadsheet.

That spreadsheet is how he knows that over eleven months he has inspected forty-one properties, made offers on nine, and been to four building inspections at roughly $600 each. Three of those four sold for more than $200,000 above the top of their advertised guide. The fourth reached a number inside the guide, and was passed in anyway.

So he did the arithmetic that most people never sit down to do, and arrived at a question we now get asked constantly:

"Am I making decisions based on a number that was never meant to be accurate?"

This week that question stopped being rhetorical. The NSW government has passed legislation to force a price or price guide onto every property advertisement in the state. It is a significant reform and it will help. But it cannot fix what is actually broken here — and the reason why is the single most useful thing a buyer or investor can understand about how property gets priced.

What just changed in NSW

The new laws do four things. They mandate a price or price guide on all advertising. They require agents to publish a statement of information showing comparable sales and the suburb's median sale price. They significantly increase underquoting penalties — agents can now be fined $110,000, five times the previous penalty or three times their commission, whichever is greater. And penalties for businesses engaging in dummy bidding at auction double from $55,000 to $110,000.

There is also a genuinely clever provision: agents must revise the advertised price throughout a campaign, and the advertised figure cannot be set lower than a written offer the vendor has already rejected, or lower than the highest unsuccessful bid at a previous auction. That closes a real loophole. If someone offered $1.4m and the vendor said no, the guide cannot keep saying $1.3m.

Commencement is not yet set — consultation comes first.

Now the number that reframes everything

An analysis of every published price guide for residential sales over a recent six-month period found that fewer than 4% of agents list a property within 3% of what it eventually sells for.

Read that again slowly, because the instinctive interpretation is wrong.

Ninety-six per cent inaccuracy is not ninety-six per cent dishonesty. If it were simply lying, the errors would run one way and they would be roughly the same size. They don't, and they aren't.

Across the capitals, the typical above-guide sale overshoots by 4–9%. The typical below-guide sale undershoots by only 3–5%. The error is lopsided — which is exactly the signature you would expect from a mix of optimism, competitive tension on the day, and, in some cases, deliberate bait.

Then it gets genuinely strange.

The same rule, two opposite errors

In Sydney, when a property sells above its guide, it sells for a median $117,500 above the top of the advertised range. Brisbane is close behind at around $100,000. In Perth, 76% of capital city sales exceed the guide ceiling; in Adelaide, 73%. Across NSW, roughly half of all sales land above guide, about 40% below, and only around one in ten actually at it.

Melbourne does the reverse. It is the only capital where more properties sell below their guide (53%) than above (37%). In regional Victoria it is more pronounced again — 71% sell below guide.

Melbourne has had mandatory price disclosure for roughly a decade. It is the most tightly regulated pricing regime in the country, complete with a requirement that agents justify the three comparable properties they used. And its guides are wrong in the opposite direction to everywhere else.

That should stop you. If the problem were purely agent dishonesty, the strictest regime in Australia would produce the most accurate guides. Instead it produced a different error.

So what is actually broken?

Not honesty. Resolution.

A price guide is an estimate of what one specific address will fetch on one specific day, produced from a small number of comparable sales in the surrounding area, filtered through a professional judgment call. The new NSW statement of information makes this explicit by requiring agents to publish comparable sales and the suburb's median sale price alongside the guide.

The suburb median is a fine number. It is also, for this purpose, close to useless — because nobody buys a suburb. They buy one address, on one street, with one aspect, one set of neighbours and one ceiling on what anything in that position can ever be worth.

This is the part we spend our working lives inside, and the numbers are not subtle. Within a single suburb — one postcode, one council, one school catchment, one set of comparables that any agent in the area would legitimately cite — we consistently measure a 20–30% spread in effective yield between the best and worst streets. Not between suburbs. Inside one.

The drivers are the ones that get averaged away the moment you zoom out: rent actually achieved rather than advertised, how long a property genuinely sits vacant, true days-on-market rather than the relisted version that resets the clock, and the ceiling the street itself imposes regardless of what you do to the house. A boundary you cannot see on a listing can be worth more than everything you can — we have measured two near-identical houses separated by $380,000 because of which side of a catchment line they sat on. That last one routinely costs people six figures — we have watched a buyer spend $210,000 on a renovation that added nothing because the street had a ceiling nobody checked for.

An estimate built on suburb-resolution inputs will be wrong at address resolution. Sometimes by $117,500. That is not a scandal. It is arithmetic.

The problem nobody is legislating for

There is a second failure mode here that gets almost no attention, and it is the one that should worry a serious buyer more.

Everyone focuses on the property that sells far above its guide. The more revealing case is the property where bidding reaches a number inside the advertised range — and it is passed in anyway.

That is our engineer's fourth building report. The guide said one thing; the reserve, sitting invisibly somewhere above it, said another. He was never bidding on a property that was for sale at the number he was shown.

The reserve is allowed to sit outside the price guide. A buyer can do everything right, land squarely within the published range, and discover the range was never the question. No amount of penalty for underquoting addresses that, because technically nothing was underquoted.

The argument on the other side, which is not stupid

It is worth being fair to the objection, because it is a real one and it is being made loudly.

The industry's position, put bluntly by its NSW representative body, is that a vendor is also a consumer — and asks whether there is anything wrong with a vendor wanting to maximise their price. The concern is a market where you scan a barcode on a house and read off its value, when every property has attributes worth more to some buyers than others.

He is right about that last part, and it is the same point we are making from the other direction: properties are not interchangeable, which is precisely why suburb-level numbers cannot price them. Where we part company is the conclusion. "Every property is different" is an argument for better information about that specific property. It is not an argument for less information.

And the compare-and-contrast is instructive. Queensland went the other way entirely — since 2014 it has been illegal to give any price guide on an auction property. One state is about to mandate guides; another bans them outright; both governments say the goal is clarity for buyers. Queensland's own auctioneers describe buyers there as flying blind, doing their own due diligence with the same data reports available to agents.

Which is, quietly, the whole answer.

What we told him

Treat the guide as marketing, not as data. Even after these reforms, it is one professional's opinion of a range, produced from comparables that may not include a single property on the street you are actually buying in. Useful signal. Not a valuation.

Do the comparable work yourself, at street level. The new statement of information gives you the agent's comparables and the suburb median. That is a starting point you should interrogate, not accept. Which comparables were excluded? Are they on this street or four streets away where the ceiling is different? What did they actually achieve, and how long did they sit? This is the same discipline we walked through with a buyer weighing a national price forecast against what was happening at his own address.

Set your own number before you see theirs. Decide what the property is worth to you on the data, then look at the guide. Done in that order the guide cannot anchor you. Done in the reverse order it always will.

Ask the reserve question early and in writing. You will not always get an answer. The refusal is itself information, and it is far cheaper to collect before you commission a building report than after.

Price the cost of being wrong. Our engineer spent about $4,000 and eleven months chasing numbers that were decorative. Independent street-level research on a shortlist costs a fraction of one wasted building inspection, and it is the only thing that tells you whether a guide is optimistic, conservative, or roughly right.

What it means for you

The reforms are a genuine improvement, and NSW buyers will be better off for them. Publishing a guide on every listing, banning a guide below a rejected offer, and lifting penalties to six figures will all remove real bad behaviour from the market.

But the legislation is trying to make an estimate honest. It cannot make an estimate precise — because the inputs it mandates are suburb-resolution inputs, and the question every buyer is actually asking is an address-resolution question. Fewer than four in a hundred guides land within 3% of the sale price, and in one state they miss high while in another they miss low. That is not a problem with the people producing them. It is a problem with what they are being asked to do.

Which is, in the end, an unusually good piece of news for investors — because it means the gap between the advertised number and the real number is not random. It is knowable. It is knowable by anyone willing to look at the street rather than the suburb, at achieved rents rather than advertised ones, at the true time a property sat rather than the reset version.

Everyone at that auction is working from the same guide. Almost nobody is working from the same research. In a market where fewer than 4% of published estimates are close to right, the buyer who has independently priced the address is not competing on nerve or luck. They are the only person in the room who knows what the number should have been — and that has always been where property rewards the people who did the work, whatever the headline says about the market that week.

The guide tells you what someone wants you to think. Your research tells you what you are buying. Only one of those has ever made anybody money.

This article is general information only and does not constitute financial, legal or investment advice. It does not take into account your objectives, financial situation or needs. Property pricing, disclosure and underquoting laws differ materially between states and territories and are subject to change, and the NSW reforms described here are subject to consultation and are not yet in force. You should seek advice from a licensed professional and conduct your own due diligence before making any property decision.