The new season lands in shops in August, which is roughly six months before anyone actually needs it and about nine months before you will find out whether it was a good idea.

There is no reliable way to predict fashion. There is a reasonably reliable way to predict your own regret, which is more useful.

The four questions

Before buying anything described as new-season, ask these in order. A no to any of them is not a veto, but two is.

1. Does it work with three things I already own? Not one — three. A garment that needs a whole outfit built around it is not one purchase, it is four, and the other three usually never get made.

2. Would I have wanted it last year? This separates a genuine gap in your wardrobe from a gap that was manufactured in July. If you were reaching for something like it before it was in the shops, that is real.

3. Is the shape new, or just the styling? A familiar shape in a new colour or fabric is low-risk — it slots into what you already do. A genuinely new silhouette changes what has to go with it, and it is far more likely to look dated in eighteen months.

4. What does it cost per wear if I am honest about the number of wears? Not the optimistic number. The honest one. The arithmetic is in cost per wear, and it settles most arguments quickly.

What reliably survives

The things that stay tend to share a property: they solve a problem rather than make a statement.

  • Outerwear in a classic shape. A good coat is never out.
  • A new neutral. If a season introduces a brown or a green that turns out to work with everything, it stays for years.
  • A better version of something you already wear. A denser knit, a heavier cotton, a shoe with a proper sole.
  • Anything that makes an existing outfit warmer or more comfortable.

What reliably does not

  • Extreme proportions. A silhouette at the far end of anything — very cropped, very oversized, very long — is the most dateable thing in a wardrobe, because the eye adjusts and the extreme becomes visible as a period marker.
  • Hardware and embellishment. Buckles, chains, contrast stitching, logos. These date faster than shapes because they are decorative rather than structural.
  • A colour that only works with itself. If the shop styles it exclusively with other pieces from the same range, that is a signal.
  • Anything requiring a second purchase to function. Trousers that need a particular shoe. A top that needs a specific layer.

The timing bit nobody mentions

New-season stock is at full price in August and September because that is when the demand is manufactured, not when the need is.

If a piece is genuinely a classic — the coat, the boots, the knitwear — buying early makes sense, because sizes go and you will wear it for five months this year. If it is a trend piece, waiting until November costs you six weeks of wear and frequently thirty per cent of the price.

The useful discipline is to decide in September what you actually want, then buy the classics immediately and put the rest on a list. Deciding before the discounts start is what stops the discounts doing the deciding.

Before you commit

The pieces most worth checking are the mid-price ones, where the spread between shops is widest and the difference in the garment is often negligible. DUPI takes a photo of something you have seen and finds the closest matches at prices that are actually listed — which is a better basis for a decision than the first price you were shown.