What is news? More precisely, when is news? Lately, the division has become starker between news that reports the details of past events, and "news" that is a forecast, a best guess, of the outcome of past events, or even of future events.

For some time, I have found fiction of increasingly little interest why should I care about made-up people, when real people have far more interesting lives and are, to boot, real? Much of what passes for "news" I find similarly uncompelling, because it is mere speculation. Worse by far, though, is the "news" that is speculation about speculation.

An example: A hurricane threatens--let's call it Cassandra. The TV news anchor, who has flown into the town most likely to be destroyed cuts to a guy in a sports coat with a computer screen. He is an "expert" on such matters, who confirms that when the storm hits in a couple of days, an apocalypse will ensue, with insured losses that will cause the collapse of insurance and reinsurance companies and disrupt shipping for generations.

When Cassandra takes a slight turn and misses the populated areas altogether, there's nary a change in anyone's routine. As Hurricane Katrina approached New Orleans late in August, the catastrophe modeling agencies fell over each other to forecast the estimated cost of insured losses from a storm that had not yet hit anything. On Friday, Aug. 26, three days before Katrina made its second landfall, RMS forecast insured losses at $1 to $2 billion. AIR Worldwide's estimate that day was a maximum of $600 million. Guess who made the evening "news"?

On Sunday afternoon, RMS upped the ante: "greater than" and possibly "far more devastating than" $9 billion. Monday morning, as the storm hit, EQECAT forecast losses for U.S. insurers from Hurricane Katrina of $30 billion.

By Monday evening, Aug. 29, after the worst of things, estimates were "between $10 billion and $25 billion." The hurricane, in other words, followed a more or less predictable line, while the financial forecasts were all over the map.

What public service was fulfilled by these estimates? And why were they so wildly inaccurate? The only answer to the latter question is a losing proposition: hurricanes are erratic, and impossible to forecast with any accuracy, right up to and after the event. That being the ease, why issue speculative reports containing more hot air than the hurricane itself?

The agencies' forecasts were picked up by news organizations and parlayed into dire predictions of the washing-away of the city of New Orleans by a 28-foot tidal wave. RM8 provided a statistic: New Orleans had $40 billion of insured values, and the surrounding parishes some $110 billion.

The modeling agencies shouted "Fire!" in a crowded media room. The agencies have been very kind to me in terms of access in the past, but really, they must desist from this sort of behavior. The fact is that we really won't know the true losses for weeks, maybe months.

To all the instant experts, prognosticators, seers, pundits and other purveyors of opinion who masquerade as news analysts by forecasting, I would paraphrase Bambi's mother: "If you don't have anything factual to say, don't say anything at all."