It looks bad. By February the foreigners held 42.5% of the car market (cars, not trucks) for this model year (marching to that magic 50% mark). Of course, we all know they are kicking Detroit hard in the car business, but then we've got trucks, meaning SUVs and pickups.

But now Honda has proven it can build a better minivan than DaimlerChrysler Corp. (which, by the way, has developed a severe case of Minivan NIH, Not Invented Here. Still no power sliding doors or built-in VCR). And that big Toyota Tundra pickup, with the 32-valve V-8, already is in production. And those foreigners keep bringing out more utility vehicles, from BMW's X5 to Nissan's Xterra, more than I care to count.

Fear not. The team from Detroit can win. They are good, those Germans and Japanese, Koreans and Swedes, but they have a weakness. They are slow. Slow to react to the market. Slow to change direction. And Detroiters can be very fast.

So the answer is to throw a rapidly changing group of innovative products on the market. Lots of new models, all the time. Build variations of what exists out there. Niche vehicles. Sell them at the top dollar. And when the market gets crowded, get out and come up with new variations.

Why get out? Because Detroit doesn't want to spend the money and manpower updating marginal products. That gets expensive. Dump them and come up with new stuff.

It's a modern variation of the annual model change.

Ford Motor Co.'s got the idea. Look at all the variations coming. The SUV-pickup variations, the Explorer Sport Trac, the CrewCab (both next spring), and the Lincoln Blackwood (coming as soon as Ford can round up enough 300-hp V-8s, say two years).

Then there is the small, high-volume, car-based SUV (that's really a new vehicle, not a variation), plus an activities vehicle variation off the Focus small car, and the big Excursion SUV (a variation off the Ford super pickup). That's six vehicles coming next year and the year after, and only one is really new.

This building of variations is what made General Motors Corp. the giant it is. With two platforms (they were called bodies in the old days), GM covered the market from Chevrolet to Cadillac. They were brilliantly done, and really different cars, even if they were variations.

Somewhere in the 1960s and 1970s and 1980s, Detroit forgot all this. Whenever a company wanted a new model, it started from scratch: an all-new design, an all-new platform. It became enormously costly to redo all these vehicles, and they began looking alike and driving alike, too.

Now the idea of variations from a single body is coming back.

It's not just Ford. Chrysler's PT Cruiser coming next spring is a variation of the small Neon that could make the Neon profitable. That's because Chrysler will get $5,000 to $6,000 more for a PT. Contrast this Chrysler idea, building a variation vehicle so desirable that customers will pay extra for it, with GM's Project Yellowstone, to build new factories and create a new manufacturing system to make its small cars profitable.

Of course, the Chrysler half of DaimlerChrysler has another terrible problem: Brain Drain.

The best executives seem to be leaving. Indeed, the three men who fought hardest to create the PT - Bob Lutz, Francois Castaing and Chris Theodore - are gone. Mr. Theodore, basically the chief engineer, went over to Ford.

And why shouldn't the best ones leave? The Germans are giving themselves the top jobs, even though the Chrysler side of DaimlerChrysler is bigger and more profitable, and the Chrysler paychecks are going to be squeezed in the future. The rumor mill even has Bob Eaton bailing out soon.

An auto company is really people. Unless Juergen Schrempp, the Daimler chief who thought up the merger, changes his management structure, he's going to find that his Chrysler partner is nothing but a bunch of old factories with no exciting product. This would be in about five years.

But right now the old gang has left enough good product to last a few years.

So let's get back to variation strategy. Is it possible that GM may be relearning the variation game? The Aztek, a crossover vehicle headed for Pontiac, is a variation off GM's minivan platform. It's just a start, but hopefully GM will understand the significance of fast and frequent model changes.

Lots of fresh metal, lots of new types of vehicles, lots of variations. This new annual model change can keep GM, Ford and the Chrysler side of DCX on the winning track despite the foreigners' gains.

Keep innovating, keep evolving, keep changing. That's the way to win. - Jerry Flint is a columnist for, and former senior editor of, Forbe's magazine.