One more prediction

Leaves are falling, footballs are flying, and the trick-or-treaters have come and gone.

Pre-season editions of basketball magazines are on the newsstands and the first wire service polls have been tabulated and published.

In the immortal words of Bob Dylan, "You don't need a weatherman to know which way the wind is blowing."

College basketball is just around the corner and expert predictions are as common as turkeys on Thanksgiving.

It's hard to justify one more prediction, but I'm going to make one anyway.

My prediction is that all the expert predictions are -- WRONG.

I feel safe in making this prediction because the weight of history is on my side.

Last year, The Sporting News, which had the courage to rank all Division I teams from top to bottom, rated Syracuse No. 26, Wisconsin No. 41 (fourth in the Big Ten) and Wake Forest No. 53 (sixth in the ACC).

Of course, Syracuse did somewhat better than that by winning the National Championship.

What you may not remember is that Wisconsin won the Big Ten title earning a Top 20 ranking, and Wake Forest won the ACC title and finished the year ranked in the Top 10.

The polls did no better.

Arizona was ranked No. 1 for all but a few weeks of the season.

During the other weeks, Duke, Alabama, Florida and Kentucky each received 15 minutes of fame in the top spot, but none of those teams won the National Championship.

In fact, none of them played in the title game and none of them made the Final Four.

Only Kentucky and Arizona even managed to reach the Elite Eight.

Experts are like the Supreme Court Justice who claimed he knew pornography when he saw it -- they believe they know a good team when they see one.

Why then, is it so difficult for experts to predict which teams will be the best in the country?

Predictions are a classic case of men working without tools.

In the absence of facts, experts extrapolate from last season's results.

They subtract out the effect of "graduations" and defections and add in a guess about the impact of new recruits.

However, a team is more than a roster of talent. Imponderables and intangibles, like team chemistry, off-the-court distractions, leadership (or the absence thereof), injuries, and serendipitous scheduling, all play major roles in a team's destiny.

The success Syracuse had last spring was due as much to great (and greatly improved) team chemistry as it was to the performance of an extraordinary freshman named Carmelo Anthony.

This is the shady area where experts are paid for diagnosis and prognosis but it is the area in which they generally fail.

Polls are essentially rolling, weekly predictions.

Based on the shaky foundation of pre-season predictions, each team is assigned a position in the pecking order and that position changes in small increments with each win or loss.

As a result it is difficult to overcome an inaccurate initial ranking. Last year Alabama was overrated but took half the season to fall out of the polls.

Wisconsin was underrated and took virtually the entire season to climb into the polls.

Alabama was never as good as its ranking and Wisconsin was always better than its ranking, but the polls aren't a good mechanism for reflecting that reality.

The crux of the matter is that experts are infatuated with the idea that winning has meaning beyond the result or outcome of an individual contest.

They adore the binary-ness of it, and individual results can be aggregated into a won/lost record and used as a measuring stick for teams, coaches and players.

It's a convenient model, but not a logical one.

College basketball teams perform erratically. Some days they play "over their heads" and upset better teams, while on other days they under-perform and lose games they shouldn't.

We also know that when teams meet twice in the same season, as is typical in conference play, the home team prevails more often than not.

Jeff Sagarin, who invented the Sagarin Ratings (www.sagarin.com), calculates home court advantage as having been worth a whopping 4.11 points per game during the 2002-2003 season.

So, obviously, individual games are influenced by factors other than the innate and inherent abilities of the two teams.

The following is my list of factors that influence individual outcomes, in descending order of significance:

1. The innate and inherent ability of team "A". 2. The innate and inherent ability of the opposition, team "B". 3. Location of the game: home, away or neutral court. 4. The emotional state of each team (being "up" or "down"). 5. The availability of each team's full complement of players (some may be injured or ineligible). 6. The quality of the coaching (including strategy, preparation, motivation and game management). 7. Officiating, i.e. good calls and bad calls. 8. The bounce of the ball (or pure luck).

Individual wins are like "The Perfect Storm" -- a convergence of immeasurable factors that create an outcome the winner may not be able to duplicate under any other circumstances.

When we aggregate these individual outcomes into a record, our intent is to even-out the breaks, the officiating, and the effects of game locations.

Records should reward good teams for consistently superior coaching, for keeping players healthy and eligible and for frequently reaching a high emotional pitch.

But comparing records is tricky business.

The most comparable records are those of intra-conference play, because teams in a conference normally play round-robin, home-and-home schedules that are fair to all teams in the conference.

Inside a conference, a team with a 13-3 record is logically better than a team with a 9-7 record.

However, it is illogical to say that a team with a 13-3 record in one conference is necessarily better than a team with a 12-4 record in another conference, solely on the basis of conference records.

Yet, we combine conference records with non-conference records--against disparate, and usually weaker, opponents -- and want to believe that a better record means a better team.

A record certainly tells us a lot about a team's character and its ability to "close the deal."

But a record doesn't tell us whether one team is better than all teams with less impressive records.

At a minimum, a record must be coupled with an assessment of schedule strength to arrive at a measure of any team's intrinsic quality.

The trick in measuring strength of schedule is in knowing which opponents are good and which are not.

Voters in the polls merely use their intuition to factor strength of schedule into their rankings.

Many rating services, on the other hand, use the records of teams played to determine strength of schedule.

They basically answer a question (how difficult is this schedule?), with another question (how difficult are the schedules of the teams on the first team's schedule?).

All of which brings us full circle to the original question -- how innately, intrinsically, inherently good is any team, and if wins and records don't reliably answer that question, what does?

I'll take a shot at that answer in the next column.

Don't get me wrong, polls and predictions are fun, but I wouldn't bet the Ferrari on their accuracy.