As Honolulu Marathon officials delve deeper into the implications of Sunday's timing mistakes, the work ahead of them grows exponentially.

After initial review of video taken from the 35th annual race, Honolulu Marathon Association president Dr. Jim Barahal now believes the entire field of runners may have been subject to timing errors.

"The data is a little off," he said. "Not dramatically, but from a few seconds to as much as 70 seconds."

"It looks like we'll have to manually go through all 24,000 runners."

Marathon officials and the company responsible for recording the split and finish times for runners believe that heavy rain and misunderstanding by runners about how to use a new timing device contributed to a situation in which thousands of results were inaccurately or incompletely recorded, or not recorded at all.

Pending video confirmation, Barahal said, about 24,300 runners likely completed the course. That figure is roughly 3,500 more than was originally reported by the official timing system.

The marathon association had been planning to spend as much as 100 hours looking through videotape to identify the missing runners. It was assumed that the finishing times that were recorded were accurate.

However, after watching a few minutes of the video, race officials realized even those finish times that were recorded were not accurate.

Barahal now estimates that it will take up to 800 man hours to closely scan digital video (originally captured for the race's MyMarathonDVD service) and determine each runner's finish time.

"We'll have to go second by second," Barahal said. "And it takes about a minute to review each second of video."

The timing system used this year was developed by Michigan-based SAI Timing and Tracking and relies on a paper strip implanted with a microchip given every runner. Electronic readers along the course are supposed to detect the chip as runners pass and record their times.

SAI co-founder David Simms, who oversaw the timing, said heavy rain short-circuited generators, which caused the readers to reset errant times. Simms also said that runners who did not properly attach timing strips to their shoes might not have been recorded at all.

He initially contended that only a few hundred runners were likely affected. However, a day after the race, Barahal expressed concern about the huge disparity between the number of registered entrants and the number of runners detected at the start.

Based on the traditional percentage of entrants who don't actually participate — roughly 15 percent to 30 percent — Barahal said as many as 3,500 runners might not have been recorded.

Barahal later found evidence of his "worst-case scenario" when he audited the number of finisher shirts and medals that had been distributed (roughly 24,300).

Once the review is completed, runners will only receive certification of their "clock" time, not their "chip" time. Split times will not be reported.

Clock time is measured from the official start of the race at 5 a.m. to the time when the runner crosses the finish line.

Chip time — more accurate because it accounts for the time it takes runners to reach the start line from wherever they are in the pack — is supposed to begin as soon as the runner crosses the start line and end when the runner crosses the finish line.

"Unfortunately, because of the failure of SAI to provide the proper timing services, there's only so much we can do," Barahal said.

OTHERS USED SAI, TOO

The certification of clock time has caused anxiety among those hoping to use their time in Honolulu to qualify for the prestigious Boston Marathon. Barahal said the Honolulu Marathon Association will work with the Boston Athletic Association to clear runners who have a legitimate time claim.

From 2000 to 2006, the Honolulu Marathon used the ChampionChip timing service, which also relies on microchip-based timing. Simms, who has worked with the Honolulu Marathon on timing for 21 years, broke off from ChampionChip to form SAI and helped convince the marathon association to adopt the new timing system, which had been tested in 40 races and used officially in the Philadelphia and Las Vegas marathons.

Simms told The Advertiser on Monday that there had been "no glitches" previously. However, official results posted on the Las Vegas Marathon Web site show missing split times for hundreds of runners and there are online discussion group threads devoted to finishers' complaints about timing problems at both the Las Vegas and Philadelphia marathons.

In a letter to The Advertiser, Las Vegas Marathon finisher David Masuda said he received split times for the 5-kilometer and half-marathon marks, but no finishing time. He said he tried to contact Simms twice but did not receive a response.

Barahal said the association is focused on providing accurate clock times to all finishers but may consider legal remedies against SAI after that. Barahal said he has not yet decided on a timing service for next year, but it's a safe bet it won't be SAI.

"I'm definitely leaning toward not using SAI just like I'm not leaning toward falling off a cliff," he said. "They represented to me that the technology was fully tested and ready to go, but it wasn't. So, to me, their level of credibility is zero."

Organizers of the Chevron Houston Marathon apparently agree. Houston was to have been the fourth marathon to use the service, but the Houston Chronicle reports that organizers changed their mind after learning what had happened in Honolulu.

The same article quoted Terry Collier, executive director of the Las Vegas and Los Angeles marathons, as saying the Los Angeles Marathon will still use SAI services as planned next year.

Dan Goltz, 36, of Volcano, completed the Honolulu Marathon in 2:56:44, according to the official record. However, his clock and chip times were the same, even though he knows it took him a few seconds to reach the start line after the starting gun went off.

'IMPORTANT QUESTIONS'

But that discrepancy doesn't bother Goltz as much as the marathon association's immediate response to the problems.

"I feel badly that they went through all of that because a lot of it was unavoidable," Goltz said. "But they should have addressed it right away because there were a lot of worried people wondering why their loved ones weren't showing up on the results. They should have posted a message on the Web site and made an announcement at the race."

Goltz said he also felt that marathon officials should have better prepared runners to use the new timing chips. Goltz said that when he picked up his packet, an attendant told him it would be all right to tuck the tag into his laces, even though doing so might cause the chip not to be read, according to the directions.

Honolulu Marathon co-founder Jack Scaff has been fielding calls from around the country as news of the mishaps spread.

Scaff is loath to fault Barahal but said that things should have been handled differently.

"There are a lot of things that need to be done to make sure new technology works properly," Scaff said. "Did they test the paper strip to see if it was waterproof? Were people properly instructed on how to use those new chips? Were the generators water-proofed? I don't want to step on any toes, but these are important questions."

In addition to the timing errors, the Honolulu Marathon Association yesterday said that 80-year-old BJ McHugh's finishing time was incorrectly identified as a world record due to incomplete information on a racing statistics Web site.

Also, the association confirmed that Rani Tanimoto was the winner of the female kama'aina finisher award. Rachel Ross (who was born in Florida) was originally credited by a race official with winning the special division for Hawai'i-born runners.

Reach Michael Tsai at mtsai@honoluluadvertiser.com.

Make a difference. Donate to The Advertiser Christmas Fund.