CHARLESTON, W.Va. -- Stop reading now if your only desire is to learn which party will win the 2010 House of Representatives election. I don't know and neither does anyone else, regardless of how many newspapers carry George Will's predictions, or how much Charlie Cook charges for his political insights.

Predicting the future about a fabricated event (there is no House election, but rather 435 separate elections) more than a year away isn't even a worthy carnival game, so how does it wind up featured in our newspapers and on our best (which doesn't mean good) political analysis shows?

The nonsense and bloviating in political punditry would not be tolerated in other sections of our newspapers. That is, we would not treat as worthy predictions of next Election Day's weather in all 435 congressional districts based on aggregate meteorological trends -- and our sports columnists would not indulge predictions about how the Yankees will fare against the Red Sox in 2010 when we don't know who will be on either team.

Actually, we know a great deal about predicting generic individual House races because we know a handful of big things:

First, incumbents almost always win, because of large advantages enjoyed by incumbents, mostly having to do with better name-recognition, voter loyalty for services brought home to the district, and election-enhancing advantages such as staff time and easier access to campaign funds.

Second, the party that wins the White House usually loses some seats in the next election because of the phenomenon known as "surge and decline." This occurs because the presidential race attracts a disproportionate number of those less interested in politics who stay home during the midterm elections. Decline happens when these marginal voters skip the succeeding election, the result being lost seats for the president's party.

Third, the party with more seats at risk will, all else being equal, lose more seats because of basic probability. For example, if you have an 80 percent chance to win each of 60 seats, you will on average win 48 and lose 12 -- whereas if you have an 80 percent chance to win each of 40 seats, you will on average lose only eight while winning 32.

Fourth, voters vote their party identification, and independents are basically coin-flippers, although incumbency can affect this behavior in House elections (see the first reason).

CHARLESTON, W.Va. -- Stop reading now if your only desire is to learn which party will win the 2010 House of Representatives election. I don't know and neither does anyone else, regardless of how many newspapers carry George Will's predictions, or how much Charlie Cook charges for his political insights.

Predicting the future about a fabricated event (there is no House election, but rather 435 separate elections) more than a year away isn't even a worthy carnival game, so how does it wind up featured in our newspapers and on our best (which doesn't mean good) political analysis shows?

The nonsense and bloviating in political punditry would not be tolerated in other sections of our newspapers. That is, we would not treat as worthy predictions of next Election Day's weather in all 435 congressional districts based on aggregate meteorological trends -- and our sports columnists would not indulge predictions about how the Yankees will fare against the Red Sox in 2010 when we don't know who will be on either team.

Actually, we know a great deal about predicting generic individual House races because we know a handful of big things:

First, incumbents almost always win, because of large advantages enjoyed by incumbents, mostly having to do with better name-recognition, voter loyalty for services brought home to the district, and election-enhancing advantages such as staff time and easier access to campaign funds.

Second, the party that wins the White House usually loses some seats in the next election because of the phenomenon known as "surge and decline." This occurs because the presidential race attracts a disproportionate number of those less interested in politics who stay home during the midterm elections. Decline happens when these marginal voters skip the succeeding election, the result being lost seats for the president's party.

Third, the party with more seats at risk will, all else being equal, lose more seats because of basic probability. For example, if you have an 80 percent chance to win each of 60 seats, you will on average win 48 and lose 12 -- whereas if you have an 80 percent chance to win each of 40 seats, you will on average lose only eight while winning 32.

Fourth, voters vote their party identification, and independents are basically coin-flippers, although incumbency can affect this behavior in House elections (see the first reason).

And fifth, at the margins, where close elections are won and lost, quality challengers, money, scandals and national tides can swing outcomes in individual races (and, often, this information becomes clear only in the weeks immediately preceding Election Day).

Predictions of a forthcoming Republican House wave made by Cook, Will and others are at best signs of profound ignorance and at worst sophistry. Again, this does not mean that Republicans will not retake the House of Representatives in 2010. Very unlikely events do occur (someone usually wins the Powerball lottery), although such an outcome would be contrary to fact-based expectations, given that the quality of challengers, future political events, and fundraising are all unknown.

The problem with outlandish political predictions is not only that being wrong has no negative consequences, but also that making predictions can actually affect outcomes (which, in Will's case, may be part of the purpose of the prediction itself).

Commentators forecasting a potential Republican wave mention some aggregate trends gleaned from selective data. For example, they cite selected facts that seem, on their face, relevant: Several dozen Democrats (including both Democrats in the West Virginia House delegation) represent districts where McCain and Bush outperformed Obama and Kerry; some polls show Republicans doing disproportionately well on a national generic ballot, particularly among independents; President Obama's approval ratings are down; Democratic gubernatorial candidates in New Jersey and Virginia are struggling; and the Republican base, by some measures including news viewership and Tea Party activity, has been more energized and less dispirited than the Democratic base.

Most pundits do not, however, share one of the dirty secrets about American congressional races: More than 75 percent (including all three in our state) are currently expected to be cakewalks, with little chance of a surprise result. According to CQ Politics, a well-respected nonpartisan political site, 335 of 435 congressional races next year fall into this category, with most of these seats (194) held by Democrats. This means that Republicans must win 40 of the 63 remaining Democratic seats, most of which are currently rated as "Democrat-favored." The remaining 30 or so Democratic seats are listed as "lean Democrat" or tossups.

Elementary math tells us that even if the odds in each of the 63 non-safe Democratic races were 50-50 like a coin toss (and they're clearly not, at least at this point), the odds of winning 40 of 63 races would be about 2 percent. Moreover, all of this also assumes that Republicans will also win all 40 relatively competitive races on their side, so it would be more like going 80 for 103, the odds of which would indeed be similar to hitting the aforementioned Powerball lottery (about 7 in 100 million).

Again, this does not mean that Republicans will not take back the House, but it does mean that we should ask more questions of those who make predictions, particularly when they have a stake in the outcome or when the predictions are inherently illogical, like the forecast of a wave the Republicans will ride to victory at a time when Republican self-identification is at contemporary lows and turnout in the midterm election should be about a third less than it was in presidential year of 2008.

We should require of our pundits what we require of our sportswriters and elementary school math students: They should discuss individual matchups explicitly, handicapping the races and discussing the odds of each challenger winning; and they should show their work.

Dr. White is a political science professor at Concord University.