One statistician's eight-point plan to clean up the system

Golf World, February 19, 1999 issue, The Bunker.

Long before David Duval won last month's Bob Hope Chrysler Classic with a final-round 59 and didn't become the No. 1 player in the world, golfers were moaning about the World Golf Ranking. The reason officials are suddenly listening is that the ranking, conceived by IMG in 1986 and long the subject of the same fan indifference as the NFL's quarterback rating system or baseball's slugging percentage statistic, suddenly matters. Besides determining the field at next week's WGC-Andersen Consulting World Match Play and the WGC-American Express Championship in November, the ranking helps decide who gets to play in three of the game's four Grand Slam events (the PGA Championship being the sole holdout).

Enter Dean Knuth, who spent 16 years working at the USGA developing the association's Course Rating and Slope Rating Systems and earning a reputation as one of the game's foremost statisticians. Today, as head of the San Diego office of the Inter-National Research Institute, he writes high-tech software for the U.S. Military. Recently, Golf Digest asked Knuth to study the system used to compute the World Golf Ranking. His findings, released this week, include eight major criticisms:

1. The ranking isn't current. The system takes into account players' finishes over the last two years, but except for doubling points awarded over the previous 12 months, it makes no adjustment for time. Knuth's recommendation: Decrease the current importance of events by one-twelfth each month you go back. In other words, points earned for finishes in last week's event are multiplied by 2.0; points for finishes earned in February 1997 (the beginning of the ranking period) are multiplied by 0.083.

2. Major championships aren't given enough emphasis. They only account for 10 percent of the total points awarded in a given year. Knuth's recommendation: Increase their value to 15 percent of the year's total.

3. The system's "strength of field" factor is out of whack. It is based heavily on the current ranking of the players in a particular tournament's field, which tends to unfairly favor players with higher rankings. Knuth's recommendation: Determine each tournament's strength of field factor by averaging its strength of field over the previous five years.

4. Points are distributed unevenly. Players earn points for top-20 finishes only. Thus, finishing 21st and missing the cut are worth the same amount of points: zero. Knuth's recommendation: Award points to all players who make the cut, similar to the way the purse is distributed at a PGA Tour event.

5. Too many tournaments are counted. The system counts all finishes in all events. Because points are awarded for top-20 finishes only, too much emphasis is placed on a player's top-20 finishes. Knuth's recommendations: Toss out each player's lowest 10 tournaments or count all major championships plus the best half of his other finishes.

6. Japanese and Australasian events count for too many points. Basically, their fields are too weak to justify their value. Knuth's recommendation: Cut their points value in half.

7. European events count for too few points. Basically, their fields are too strong to justify their value. Knuth's recommendation: Make their value comparable to a PGA Tour event.

8. Unofficial events are too official. Small-field, off-season, invitation-only events such as the World Match Play and the Million Dollar Challenge are worth too many points. Knuth's recommendation: Cut their value in half.

The result: Under Knuth's proposed system, Duval would have been ranked ahead of Woods until the latter's victory at the Buick Invitational last weekend. But the controversy over which player is No. 1 isn't what Knuth believes is most important: It's the controversy over who belongs in the 30 slots between Nos. 40 and 70, where, he says, the current system is most error-prone, and the difference between players is often less than a tenth of a point. A more practical explanation is this: Under the current ranking, Carlos Franco, Frankie Minoza, Craig Stadler and Nick Faldo get to play in the next week's Andersen Match Play. Under Knuth's system, that quartet drops out and is replaced by Per-Ulrik Johansson, Frank Nobilo, Jay Haas and David Toms.

Golf Digest will present Knuth's report to the International Federation of PGA Tour officials next week.

-- Compiled by Geoff Russell