Is there anything enviable about having to taste scores of the new 2004 Bordeaux from the top châteaux?

Is there anything enviable about having to taste scores of the new 2004 Bordeaux from the top châteaux? You might think so, and a UK wine merchant gloated about his forthcoming trip in an e-mail to me earlier this month. But nobody could be envious of a week more mouthpuckering than mouthwatering, tasting hundreds of unready young wines. By coincidence, the same wine merchant was right in front of me on a plane taking half the UK wine trade to Bordeaux. I didn't come across him that week but I couldn't help wondering how green he was feeling by the end.

The party from Britain decamped to France earlier this month to do the rounds of the châteaux for the annual tasting of each new vintage. It's a unique event. No other region in the world sells its fine wines quite so long before they are ready to be bottled. The en primeur system - Bordeaux wine merchants buying from the châteaux, and UK and other national wine traders buying from them and selling to far-sighted, economy-conscious or speculating members of the wine-buying public - works to everyone's advantage in good years. Château owners receive an instant injection of cash, Bordeaux and UK wine merchants take their slice, and consumers benefit by paying less than the wine will cost in two years' time when it has been bottled and delivered to the UK. So the theory goes. In practice, for the system to work, the wine has to be very good indeed and the price reasonable enough to justify the outlay upfront. The 2000 and 2003 vintages are examples of it working well, but everyone comes unstuck if the wine is too expensive from the start, as happened with the 1997.

Last year's Bordeaux harvest was late but abundant, making up for the relatively small vintages of the two previous years. Quantity and quality are not synonymous though. To get their grapes to ripen, producers who could were secateuring unripe bunches like mad to try to concentrate the fruit in the remaining grapes. "We spent 19,000 hours cutting bunches," says Christian Moueix, whose family company controls 16 châteaux in St Emilion and Pomerol. Following an indifferent July and August, a dry September eased anxieties. The weather favoured producers with the nerve not to pick too early, especially in the Médoc, where the late-ripening cabernet sauvignon prevails.

To avoid unripe, diluted wines producers had their work in the cellar cut out. Some overcompensated by concentrating the juice or adding flattering amounts of oak, or both. Result: some truly undrinkable wines. Many lesser châteaux have made doughnut wines - initially pleasing, but with little fruit in the middle and a drying aftertaste. Top châteaux, mostly in the Médoc but a handful in St Emilion and Pomerol, are in the best position to work hard in the vineyard and cellar, select the right vats for blending and sacrifice unsatisfactory wines. It is these that have made classic wines, preserving vibrant fruit with the flesh and backbone to age. Lafite, Margaux and Léoville Las Cases, for instance, have put only around two-fifths of their production into their final blend.

2004 seems to confirm the logic of Bordeaux's complex and arcane classification. The vintage has tended to favour châteaux with the best locations and resources. Bearing in mind that tasting unfinished wines makes predicting their future only a little more reliable than a village fête palmist, it doesn't look as if the 2004 vintage overall has the necessary consistency. Bordeaux needs to cut back to 2002 prices or below. Only then will it be worth finding out which châteaux have delivered the goods.