WENATCHEE — With the governor's election trial ending today, Judge John Bridges has answered one key question but left another a mystery:

He said he'll rule from the bench on Monday. He didn't say how much math he'll be doing over the weekend.

The session in Chelan County Superior Court yesterday was dominated by testimony about alleged felon voters and dueling statistical analyses. The bottom line: how illegal votes should be subtracted from the totals for Democratic Gov. Christine Gregoire and Republican Dino Rossi.

Bridges made self-deprecating remarks about his math skills and asked questions directly of experts for both sides. He worked to understand the Republican theory that illegal votes cast in any given precinct should be deducted from each candidate's total by the same proportion as the overall vote for each candidate in that precinct.

The judge has not ruled that he accepts that method of divvying up illegal votes, referred to as proportional deduction. But if he does, he then will have two more important steps to take.

The first step would be to settle on a list of votes he considers illegal, selecting from among the thousands alleged by Republicans and Democrats.

He then would have to apply the proportional-deduction model used by the experts.

The prospect has left more questions than answers. Will Bridges do that himself over the weekend, or does he have an expert standing by to assist? Will he arrive in court today with a list of illegal votes and ask the experts to do it? Might he show up Monday, issue a ruling on the number of illegal votes and then ask for the analysis to be done?

"I've really got no idea," said Republican attorney Mark Braden, who has handled questioning the expert witnesses. Attorneys for the Democrats and for Secretary of State Sam Reed had the same response.

Republicans have claims in addition to the alleged illegal votes. And some questions remain about how Bridges will consider, for example, errors by election officials.

The lawsuit was filed in January by Rossi, GOP Chairman Chris Vance and other Rossi supporters. They allege that errors and fraud robbed Rossi of a victory over Gregoire. Rossi won the initial count and a machine recount. Gregoire won a hand recount by 129 votes and was declared governor.

Felons and other illegal voters have been a central piece of the Republican lawsuit, and it remains one of their best hopes for overturning the election.

Felons in Washington state are prohibited from voting until they have served their prison terms, completed community supervision and paid all fines and restitution. After months of wrangling over numbers of alleged felon voters discovered by both parties, Democrats and Republicans have largely accepted the bulk of each other's lists.

Each party's alleged felons fall mostly into precincts around the state that voted heavily for the opponent; by arguing to subtract those votes, each party hopes to hurt the other's candidate.

There are remaining disputes over illegal votes, most notably hundreds Republicans claim, based on King County's inability to reconcile ballot tallies.

Much of yesterday was taken up with discussions about illegal votes. Noel Frame, who led the Democrats' research efforts, testified about the process her team used for finding felons in 34 counties.

Both sides also called expert witnesses to weigh the pros and cons of the Republicans' proposed statistical analysis. And Bridges, who said at one point the technical testimony was "getting very close to the top of my abilities," had questions for both.

The Democrats' expert was Mark Handcock, a professor of statistics and sociology at the University of Washington. He testified that the Republicans' theory was not scientifically sound because they took illegal votes only from Gregoire strongholds, which created a bias in the sample.

Republican attorney Mark Braden asked him repeatedly whether he could apply the analysis to an overall list of illegal votes if Bridges settled on one.

Handcock resisted answering but seemed to say he still thought that was not a sound scientific principle.

"Did I understand correctly that the court's number is not better than anybody else's?" Bridges asked Handcock.

Answered Handcock, "That's close to correct."

The professor said he would still have concerns that the list was arbitrary and he would still be uncomfortable using a list determined by the court.

"Unless of course I was instructed by the court to do that," he said. "I would always try to do the best I can from a scientific perspective."

Republican expert witness Anthony Gill, a University of Washington associate professor of political science, got a rough cross-examination from Democratic attorney David Burman.

Gill was presenting his latest proportional-deduction analysis using felon voters from both parties, but he included hundreds of other alleged illegal votes from Republicans while excluding similar votes on the Democrats' list.

Burman said Gill took numbers from the Republicans' data analyst, Clark Bensen, and did no independent analysis to see if it was a fair sampling.

"You took what Mr. Bensen and the attorneys fed you ... and you were more than happy to go along with that, weren't you?" Burman asked Gill.

Gill said he had time constraints and was given the task of analyzing the list the attorneys gave him.

Burman bored in on Gill, reading from the professor's Web site, where he cautions his students to be critical and skeptical of statistics.

Burman: "You never considered any alternative to what Mr. Bensen and Mr. Braden fed you, correct?"

Gill: "That is correct. I was given a data set to make a simple estimation here. ... I saw that as a basic, straightforward calculation."

It was some of the harshest questioning of the trial, and Bridges told Gill not to take it personally. He said the professor was just unfortunate to be the last witness of a long day and one of the last in the two-week trial.

But Bridges had questions of his own. He was confused by the wildly different conclusions Gill and Handcock reached when doing proportional reduction.

"I thought Dr. Handcock attempted to do what you did and his results are totally different, or I should say your results are totally different," Bridges said.

"Maybe I'm wrong," he said and joked about his difficulty in understanding statistics.

"I would run from your class like the wind," Bridges said.

Handcock and Gill both used figures that included illegal votes from both parties, but Gill's report did not include as many votes as Handcock's.

After subtracting the illegal votes precinct by precinct, Gill's report found Rossi to be the winner and Handcock's favored Gregoire.

"Am I wrong in thinking that if we have similar numbers we should have similar results?" Bridges asked Gill. "I'm not able to reconcile that. Is there anything you can do to help me?"

Gill told the judge that depending where in the state the votes come from, subtracting them hurts one candidate more than the other.

"Ahh, OK," Bridges said.

David Postman: 360-943-9882 or dpostman@seattletimes.com