For almost 30 years, Sam Katz and I have been great friends, even when we were on opposing sides of an issue. We had feisty public-policy disputes when his friend Bill Green ran against my friend John Heinz for the U.S. Senate and argued when working for competing business groups here in the city.

Over the decades, we came to agree more than disagree on political issues, and our families have also become close. My wife and daughter both worked on his gubernatorial campaign, members of my family have contributed chunks of money to his campaigns, and I enthusiastically endorsed him for mayor many months ago.

So why am I so unhappy that a few weeks back my newspaper published its poll showing my friend running a competitive race for mayor, and, for all I know, our poll scheduled to be released tomorrow might show the same thing?

The answer is simple.

I've argued strongly against the paper ever taking and reporting on its own polls because I believe the Daily News should report the news, not make it.

We do more than merely make news whenever we break a poll story. Sometimes we also break the back of a candidate who's trying to break out of the pack.

That's what happened back in the Democratic primary when a Daily News poll showed that Dwight Evans' campaign had almost fallen off the polling radar screen. From that point forward, Evans' campaign was over because his money evaporated. No money, no message, no candidacy.

A positive poll can do the opposite. On the morning the Daily News poll showed that Katz and Street were running neck-and-neck, Katz had his single most successful fund-raising breakfast ever and collected more than $400,000 in pledges.

How ironic that one sound bite, a positive poll paid for by one branch of the media, could instantly help finance more sound bites in the form of 30-second TV ads.

A Daily News poll has that kind of power even though no one, including our own pollster, knows how accurate it really is. The problem lies in the assumptions any pollster must make in conducting a poll. How many "probable" voters are there? Will blacks, whites, Hispanics and other groups turn out in equal numbers? Will any, or all, of those people be honest with the pollster on the phone?

Minor misses by the pollster on any of those guesses can skewer the results. And even if the pollster guesses right, so what?

The poll the Daily News will report on tomorrow will be based on phone interviews taken a few days ago. And this week, both candidates will pour another million bucks into TV ads, send maybe a million pieces of direct mail and place hundreds of thousands of phone calls to the homes of swing voters.

More money, energy and effort will go into this last week of the campaign than anyone has seen in this city in decades. So even if one candidate is up or down a few points in tomorrow's Daily News poll, it probably doesn't mean anything.

Not this year. Not in this election. This one's too close for any poll to call, and that's before you factor in the effect of rain on the Election Day turnout.

How ironic it would be if, after $25 million, dozens of debates, hours of 30-second TV ads and millions for mailings and phone calls, the final result were to be dictated by something neither candidate can control. Rain.

So, please, forget the polls this year - especially any from the Daily News. At best, they are nothing but good guesses.

This race is up to you.

Please vote.