An independent archive of Bargain Hunt
At its simplest, this project aims to record every episode of Bargain Hunt: when the episode aired, who presented it, which experts were involved, what items the teams bought and where they bought them, where the items were sold, and who won.
Beyond that, the project is an attempt to build a usable archive of the programme as data. That means asking broader questions: which item types perform best, which buying locations produce the strongest results, how auction houses compare, how experts perform over time and what patterns emerge once hundreds of episodes are looked at together rather than in isolation.
The Gavel Index aims to track episode-level and item-level information across the programme's history.
Episodes, series, presenters, experts, buying locations, auction houses, auctioneers, teams, item descriptions, purchase prices, competition prices, sale outcomes and selected item classifications.
Estimates, bonus-buy detail, specialist item classification, and certain contextual notes where the programme does not consistently state them.
Team challenges, buying locations, auctioneers and auction houses in clip shows or compilation episodes, or in earlier eras of the programme before certain formats were introduced.
Data is compiled from programme videos or transcripts and then standardised into a structured database. The key principle is consistency rather than pretending the programme is more precise than it really is.
Expert statistics describe results for teams supervised by the expert across indexed episodes, rather than attributing every purchase to the expert personally. Episode types and currencies are kept separate; no currency conversion is applied. A complete monetary total is unavailable when one or more contributing team totals cannot be calculated.
Hammer returns use items with both purchase and hammer prices recorded in the same currency. ROI is total profit divided by total purchase cost for that matched sample, expressed as a percentage, rather than an average of individual item percentages. Missing values are not zero; a zero total cost does not produce a calculable ROI.
Median profit is the middle result when calculable profits are ordered, or the mean of the two middle results for an even sample. Lower and upper quartiles mark the middle half of results and are shown for at least four calculable items. They use linear interpolation at positions (n − 1) × 0.25 and (n − 1) × 0.75 in a zero-based ordered sample. Loss rate is the percentage of calculable items with a negative hammer profit. Sample sizes are shown alongside the results; small samples should be interpreted cautiously.
Buying profiles describe category, primary material and period. Shares use all items in the selected scope as the denominator, including unclassified items, so recorded classifications may total less than 100%. Classification coverage is shown separately. Ties for the most-purchased classification are retained.
Series rows cover indexed appearances, ordered by series number, rather than a complete career chronology. Purchase-price bands are 0 to under 25, 25 to under 50, 50 to under 100, 100 to under 200, and 200 or more, in the relevant purchase currency. Each lower limit is included and each upper limit excluded. Items without a purchase price cannot enter a price band.
“Calculable Competition Items” counts selected items whose competition contribution can be calculated. “Calculable Hammer Items” counts items with a recorded purchase price and hammer price in the same currency, regardless of whether the item was selected for the competition. Category Breakdown uses hammer results. Competition and hammer figures can differ for protected values, stolen items, disqualifications and rejected bonuses; an unknown hammer price is not treated as zero.
Most-Purchased Category is determined by the number of items bought. Best-Selling Category is determined by total recorded hammer receipts, not competition returns or profit. Total Hammer Return includes every recorded hammer price; Hammer ROI uses only the matched sample with both prices recorded in the same currency. As a result, total hammer return and total spending need not reproduce the displayed ROI when their available samples differ.
Core and bonus rows describe items from the expert’s supervised teams, regardless of who bought the bonus. Hammer results include rejected bonuses where calculable; competition profit includes only selected items with calculable contributions. Unknown item roles remain separate.
Category comparisons require at least ten calculable items in both samples. The archive benchmark includes the expert’s own teams and matches episode type, competition currency, purchase currency and hammer currency. The difference is the team average hammer profit minus the archive average. This is a descriptive comparison, not a ranking or a measure of the expert’s individual influence.
Ideally, I would list episodes and series as they appear on the BBC's Bargain Hunt page.
The Beeb, with its rather arbitrary approach to its own indexing, has other ideas.
Take Series 27.
At the time of writing, the series page lists 32 episodes, each with a sub-heading marking it as the xth episode of 32.
All seems in order.
But look closer.
Not only are episodes 9 and 31 missing from the list, two episodes are labelled as episode 15, and another two as episode 18.
That isn't a numbering error.
The BBC's Genome Project, compiled from the Radio Times listings, shows that, yes, two episodes were listed as the fifteenth of the series, and two as the eighteenth.
What's more, the listings also show gaps in the schedule where episodes 9 and 31 should be.
The numbers don't add up.
The BBC says Series 27 contained 32 episodes.
The schedule suggests 34.
What, then, is The Gavel Index to do?
Implement its own logic is what.
Series align with the BBC designation – but episodes are ordered by their broadcast date.
The first episode broadcast in the series is episode 1, the second is episode 2, and so on.
What the BBC says becomes irrelevant.
This occasionally produces episode counts which differ from the BBC's own numbering.
Some series gain episodes.
Some series lose episodes.
That is a problem for Aunty to solve.
Assigning a type is the easier of the two: it is simply what an item is.
A lamp is a lamp.
A chair is a chair.
A spoon is a spoon.
Classifying a category is more difficult for some items than others.
A lamp falling under the Lighting category is easy enough.
So too is categorising a chair as a piece of furniture.
Harder is what to do with the spoon.
After much thought, I decided to arrange items into the categories where most of the item's value lies.
A stainless steel serving spoon, therefore, falls under the Kitchen & Domestic category, whereas a silver serving spoon is listed within the Silver & Metalware category.
This is not a perfect system, but it is a consistent one.
Period, primary material and additional material provide further classifications where that information is recorded.
Item categories and types are standardised to keep the database usable. That means very similar objects may be grouped under a common type for analysis. The full wording and nuance of the object is in the item description.
Sometimes Bargain Hunt does not state a figure or fact: an item might not receive an estimate or it may not be fully described. In situations where the show does not provide information, the database does not invent it; it leaves the relevant field blank. Not all blank values mean the information is impossible to find elsewhere, mind; they can also mean the information is not currently confirmed for the purposes of this dataset.
Best-selling means highest total recorded hammer receipts. Hammer return includes all recorded hammer prices; ROI uses only items with both prices recorded and matching currencies. Competition results include selected bonuses only. Missing prices are excluded; ties are shown together.
No. The Gavel Index is an independent database and analysis project. It has no affiliation with the BBC or with Bargain Hunt.
The hammer price records the auction sale. The competition price records the programme’s scoring value, which can differ because of disqualification or a programme-stated adjustment. A selected disqualified item counts as a zero sale, so its competition loss equals its purchase price. An unselected item, including a rejected bonus, is excluded entirely: neither its purchase cost nor its sale contributes to the team result.
No. It records what happened within the programme. Bargain Hunt is a television competition with a compressed and highly simplified view of the antiques market. Auction outcomes shown on the programme are useful data points, but they are not a definitive guide to 'true value'.
Financial results use accounting notation: profits appear without a plus sign, and losses appear in brackets, for example £25 and (£25). This also applies to ROI percentages. Profits are green, losses red, and break-even uses the normal text colour; brackets identify losses independently of colour.
Actual auction profit or loss = Hammer Price − Purchase Price. Competition profit or loss = Competition Price − Purchase Price for selected items; unselected items contribute zero. Item and venue auction statistics include unselected purchases where their actual results are known. Team winners use competition totals. Amounts must be in compatible recorded currencies; the site does not convert them automatically.
The team rejected it. In those cases the item may have sold at auction in reality, but its competition contribution is zero under the rules of the programme.
It is a programme-specific adjustment to the value used for scoring. The archive records the programme’s final competition price, including any protection already applied. It then subtracts the purchase price once when calculating a selected item’s competition profit or loss. It does not reconstruct an adjustment from an estimate.
Item descriptions can change throughout an episode. An expert might disagree with the information written on the item's ticket, and the auctioneer might disagree with the expert. When it comes to deciding whose assessment to favour, The Gavel Index leans towards the auctioneer's description as they have more time to research the item, while reserving the right to standardise descriptions for database consistency.
No. The archive’s actual auction profit is the recorded hammer price minus the purchase price, without fees, commissions or taxes. An item bought for £40 and sold for £41 therefore shows a £1 actual auction profit in the database. That is not a calculation of the seller’s net proceeds or a guide to the wider antiques market.
The archive records the pre-auction appraisal, rather than the supervising expert’s buying-stage estimate. Appraisers and auctioneers have separate roles in the database, even where the same person performs both. An estimate can be a range, a single value or an upper threshold, depending on what was stated.
A blank estimate field does not necessarily mean missing data. In almost all cases where an item doesn't have an estimate, it's because the auctioneer did not give one. James Lewis and Charles Hanson are the worst offenders for this. James Lewis claims not to provide estimates for any item he values at less than £20 – ignoring all the times he has gone on to do just that. Charles Hanson, for his sins, becomes so excitable when discussing the provenance of an item that the guide price often becomes an afterthought.
A few reasons. The first is that not all charity episodes feature celebrities; some on occasion have featured members of the public. More to the point, whereas celebrity episodes follow the same rules as competitive episodes, charity episodes are something of a free-for-all. Rules are applied more leniently in charity episodes – or outright ignored. Teams submit more items than the regulated four. Pudsey runs around the auction room encouraging people to bid. In all the excitement items end up selling for far higher than their guide price. Separating celebrity and charity episodes from one another – and from competitive episodes more generally – also allows the database to analyse item performance without these outliers skewing the results.
No. That's a different show. If you want those numbers, build your own database.
The Gavel Index is a living database. Statistics such as win rates, highest profits and best-performing categories will invariably adjust as more entries are added to the database. Changes may also occur from correcting errors, standardising item types, or reclassifying items into more appropriate categories.
You can – although I reserve the right to accept or reject any contribution that I cannot verify. If you want to suggest a change, you can find contact details at the bottom of this page.
Series and episode directories include known catalogue episodes, including clipshows. Public statistics use fully indexed episodes: verified episode and item records, an auction, and confirmed complete team and item lists. Missing optional details such as estimates do not by themselves prevent an episode being indexed. Unused reference entries are excluded from public statistics.
Expert appearances count supervising team assignments. Presenting, appraising or auctioneering does not count as an expert assignment. Equal team competition totals are a draw. Win rate is wins divided by wins plus losses plus draws; unresolved outcomes are excluded from that denominator.
Competitive episodes form the Records rankings. Statistics can be viewed separately for other episode types. Competition currencies define separate groups, with actual auction amounts separated further by auction currency where necessary. Programme-stated converted purchase amounts are recorded in the currency of those amounts; no historical exchange rate is invented.
One episode-specific auction remains one auction, even when several auctioneers share the rostrum. Each person has a joint appearance, but their appearance counts must not be added together to count distinct auctions. The programme does not identify which person sold each lot in these cases, so individual financial auctioneer records use solo auctions only.
An unknown value is not zero. It can indicate an unstated or unconfirmed amount, an outcome that cannot yet be calculated, or no eligible result in the selected group. Averages use known calculable values. Financial totals used for Records require complete relevant amounts so that a partial total is not ranked as complete. Optional details may be omitted from profiles.
Actual ROI is actual auction profit divided by purchase cost, expressed as a percentage. A zero purchase cost has no defined ROI. Expert ROI uses total actual profit divided by total purchase cost across the expert’s supervising team items, including unselected items. The category-at-house record uses average actual profit. Auctioneer profit records use solo auctions. Record holders with equal values are shown together; no minimum sample-size threshold is currently imposed.
Summary names combine the recorded period, primary material and type, omitting missing parts. The item page shows the full description. Its URL is assigned once using the item ID and description; later description edits do not automatically change that URL.
Home selects from eligible facts at random on each page load. The same fact can appear again. The underlying results only change when the indexed data changes.
Realistically, I have no way of stopping you from manually using facts that are visible on a public page. All I can do is request that if you do use the data that is publicly available, you cite the site.
But I cannot stress enough the importance of treating the data with the right level of caution. The Gavel Index is useful as a record of what happened on Bargain Hunt. It is not a universal pricing oracle for antiques and collectibles and any intent to use it as such is done at your own risk.
If you want to report a correction, ask a question about the methodology, or discuss the project, you can contact me at: