And watch price trends. Most successful artists show a slow price increase in the first phase of their work, a steeply upward price trend in the second and a flat or downward price trend in the third. Decreasing returns set in, then negative returns. It is an S-shaped curve. You may be told that an artist’s work has tripled in value over the past 10 years, and that may be true. But it masks the reality that prices peaked two years earlier, that they have now dropped by a quarter and that the work now appears only sporadically, in secondary auctions. Plot prices as best you can, and when the steep rise ends and the flat phase seems to be starting, sell.

An interesting insight comes from David Galenson, a professor of economics at the University of Chicago who has studied the relative value of paintings. Galenson suggests that a pattern holds for the majority of successful artists: Their most valuable work is produced either early in their career, as in the case of Andy Warhol, or very late, as with Jackson Pollock. The younger innovators—think Picasso—are conceptual artists who make breakthroughs by introducing new ideas of painting at an early age. The older innovators—think Cézanne—develop their art through a lifetime of work, with their best-recognized contributions coming later, as the result of trial and error. This also suggests that for investment purposes, you should ignore older artists, because the market will have found them first, and midcareer artists, because they will already have had increases in value. The greatest return comes from discovering the innovators, the ones who will help shape art for decades to come. Try first to identify trends, and then identify the young innovators. That is how someone like Charles Saatchi collects.

When you visit galleries and auction houses, do not expect to like very much of the contemporary art you see, and don’t be put off by disliking everything in a show—everyone dislikes most of what they see. There is an art world saying that you have to see hundreds of works you don’t like before you begin to understand what you do like. This is a good argument for attending art fairs. Look at 1,000 works before you buy anything, and expect to see an additional 500 for every additional purchase. According to MoMA, the average patron looks at a work of art for seven seconds. When you enter an art fair or gallery room, try the “which one as a raffle prize” approach, or else look around and think, “If I were going to steal one, which would it be?” Most people have thought more about imaginary theft than about connoisseurship. Once you have circled the room, spend 30 seconds looking at the object of your criminal lust, and try to understand why you chose it.

When it comes to the work of Western artists, what kind of painting will appreciate most? There are general rules. A portrait of an attractive woman or a child will do better than one of an older woman or an unattractive man. An Andy Warhol Orange Marilyn brings 20 times the price of an equal-size Richard Nixon.

Colors matter. Brett Gorvy, cohead of contemporary art at Christie’s, claims that the grading from most salable to least is red, white, blue, yellow, green and black. When it comes to Warhol, green moves up. After all, green is the color of money.

Bright colors do better than pale colors. Horizontal canvases do better than vertical ones. Nudity sells over modesty, and female nudes for much more than male. A Boucher female nude sells for 10 times the price of a male nude. Representations of people do better than landscapes. A still life with flowers is worth more than one with fruit, and roses are worth more than chrysanthemums. Calm water adds value (think of Monet’s “Water Lilies”); rough water brings lower prices (think of maritime pictures). Ship-wrecks bring even less.

Purebred dogs are worth more than mongrels, and racehorses more than cart horses. For paintings that include game birds, the more expensive it is to hunt the bird, the more the bird increases the value of the painting; a grouse is worth three times as much as a mallard. There is an even more specific rule, offered by New York dealer David Nash: Paintings with cows never do well. Never.