After two months of painstaking video review — believed to be the largest effort of its kind in road racing history — the Honolulu Marathon Association has closed the books on its 2007 race results, concluding that nearly 2,000 finishers were not included in its original results and that less than 8 percent of all finishers were correctly timed.

Using video collected by Sports Media Productions for the MyMarathonDVD custom race DVD service, marathon officials and specially trained contractors reviewed some seven hours of finish-line video to confirm the number and identities of the finishers.

The unprecedented effort took an estimated 1,200 to 1,500 man hours at a cost of about $50,000, according to Honolulu Marathon Association President Jim Barahal.

Despite close examination (reviewers spent as many as four hours scanning each minute of video), the marathon association said it was not possible to "reconstruct a completely accurate, standardized timing data base" for the Dec. 9 race.

While many runners received a certificate verifying their "chip" time — the most accurate reading based on when they first crossed the start line and when they crossed the finish line — many others were only able to get confirmation of their "gun" time, meaning that their start time was the same as the official start of the race regardless of when they actually made it across the start line.

Many runners also received inaccurate or incomplete "split" times, which indicate their progress at different points along the route. The roughly 2,000 people who were not recorded at the finish line received no split times.

Final certificates were mailed this week, and the marathon association posted a message to all finishers on its Web site (www.honolulumarathon.com) explaining the situation.

"I'm never going to be satisfied with the results of this race because it was an extraordinary failure," Barahal said. "But in that context, we did all that was possible.

"When people finish a race, they want their time confirmed and they want to be rewarded with a certificate that shows their time," Barahal said. "That's what was most important to us. Given where we were after the race, I'm happy that we were at least able to identify every finisher."

NEW TIMING SYSTEM

According to the marathon association, the problems stemmed from undetermined glitches in a new timing system developed by Michigan-based SAI Timing and Tracking.

The marathon previously had used the ChampionChip timing system, which uses a microchip-implanted plastic disc that attaches to runners' shoes.

SAI's system, which uses a smaller, lighter microchip-implanted tag was expected to enhance the quality of data collected while reducing overall cost. Instead, the system's failure resulted in the largest timing fiasco in the race's 35-year history.

Concerns about the accuracy of the race results arose within hours of completion of the event. The next day, Barahal told The Advertiser that as many as 3,000 runners may have been missing from the final results based on the number of finisher's shirts claimed after the race.

Over the next few days, hundreds of runners came forward to say that their results were inaccurate, incomplete or missing.

The marathon association verified that there were 22,839 finishers this year, not 20,865 as originally reported by SAI.

SAI's original data also indicated that there were only 20,582 runners who started the race.

Because of the density of people at the start line, the association would not have been able to get an accurate count of exactly how many people started the race, which was held under rainy and humid conditions.

However, by subtracting the number of race bibs that were not claimed at the pre-race packet pickup from the total number of verified entrants, the association determined that the maximum number of possible starters was 24,769.

According to the association, only 1,798 finish times reported by SAI were accurate.

FINDING THE MISSING

The video review, which used specially designed software developed by SMP's Holly Huber, was directed toward identifying the 3,070 "missing" runners, who finished the race but did not have recorded start times. That proved a more complex process than anyone had imagined.

"We had to find the found before we could find the missing," Barahal said. "It took an unbelievable amount of cross-referencing to be able to do that."

According to Ryan Lamppa, media director and researcher for Running USA, no other race has ever undertaken such an involved effort to confirm its timing results.

"They didn't have to do it, but to their credit, they did," Lamppa said. "It was good to see them take the time to get official results. I'm sure everyone who ran the race is appreciative of that."

According to marathon officials, the effort also went a long way in soothing the concerns of the marathon's Japanese sponsors.

Japanese runners make up the majority of the marathon field each year, and the race's biggest sponsors are all based in Japan.

The marathon will resume using ChampionChip for next year's race. Barahal said the association has retained legal counsel to try and recoup damages from SAI.

Reach Michael Tsai at mtsai@honoluluadvertiser.com.