Tenn. players picked up on new odds before lottery revealed glitch
Posted: 10/28/2007 7:43:05 AM
Updated! Includes illustrative charts below.
Somebody knew. Or somebody caught on.
An analysis of 55 million tickets played in the Cash 3 and Cash 4 lottery games shows that in August, more people started playing the numbers most likely to come up.
That is the time period when a "glitch" in a new computerized system of picking numbers prevented duplicating combinations such as 1-1-2 or 3-3-3-3 from being drawn.
A Tennessean computer analysis indicates that players either knew or figured out that duplicate numbers weren't coming up, because many stopped playing them. An outside auditor hired by the lottery to examine the glitch isn't investigating whether fraud was involved in such purchases, however.
Todd Northrop, creator of the Web site Lotterypost.com, where lottery enthusiasts trade their betting strategies, said it was essential that the lottery look back at what happened to ensure fraud was not a factor.
"In order to run a lottery, I think you have to be so far above board," Northrop said. "You have to demonstrate that there is no way something is happening behind the scenes."
Lottery CEO Rebecca Paul Hargrove said lottery officials did look into the questions of fraud. They looked at the 15 numbers that paid out the most money during the time the computers were drawing only singles, and at other time periods, and the payouts were similar.
"If you are going to do fraud, you would put a lot of money on specific numbers," she said. "That just didn't occur."
Coding error ran 3 weeks
On July 28, the lottery switched from little white balls to computers to pick random numbers for Cash 3 and Cash 4.
Unbeknownst to the lottery, the computers were blocked from picking repeat numbers because of a computer coding error.
The lottery didn't discover the problem until Aug 20, and announced it on Aug. 21.
"In order to run a lottery, I think you have to be so far above board. You have to demonstrate that there is no way something is happening behind the scenes."-Todd Northrop, CEO, Lottery Post
The Tennessean requested a database of tickets played throughout 2007, for the purpose of scrutinizing betting practices during the malfunction. The analysis was conducted by The Tennessean and two members of Vanderbilt University's Biostatistics Department, Rafe Donahue and Jeffrey Horner.
The analysis shows — within a week or so — a rise in the percentage of non-repeating numbers, such as 3-4-7, compared to repeating numbers such as 4-4-2.
Among the 5.6 million numbers played during the time when the computers weren't able to pick all numbers, roughly 190,000 more single numbers were played than would have been expected based on prior betting patterns.
More than 2 million numbers were played that had no chance of winning because of the glitch.
"Even without the lottery announcing there had been an issue, people were changing their strategy based upon what they saw," said Donahue, a research associate professor.
"It's very obvious that the people playing the games were detecting something was amiss," Donahue said. "At least they believed it was amiss, because they changed their betting habits."
Maybe, as Donahue said, players just figured it out over time.
Or maybe someone actually knew about the glitch and took advantage of it to buy more non-repeating number combinations while the odds were more favorable for those who knew.
Donahue said he discounts fraud based on the gradual rise in single numbers being played after the computer system took over Aug. 28.
"It wasn't like there was a spike that occurred right after they switched," he said.
"I generally don't attribute to malice what can be adequately explained by incompetence," he said. "I just think a bunch of administrative bureaucrats didn't really know what they were doing. But lottery players started adjusting their strategy. That's pretty obvious."
Hargrove disagreed the analysis showed any such trend. She said the increase in single numbers being played could have been the result of other factors, not any observation by players that only singles could win.
"The Cash 3 player and Cash 4 players tend to be very, very superstitious," Hargrove said. "They tend to play numbers for a whole wide variety of reasons."
A number of players did realize something was amiss, however.
They began contacting the lottery about the lack of duplicating numbers being drawn as early as a weekafter the computers took over picking the numbers. They said they received boilerplate replies with assurances that the computers had been checked and were running properly.
Others simply quit playing. A number of them told The Tennessean they stopped or cut back severely because they weren't winning anymore, didn't trust the system and didn't like the change to computers picking the numbers.
"I have definitely stopped buying them," said Bertha Pinkerton of Nashville. "The way they have got the drawing now, you just don't know what is coming out of there. When we had the balls, you could see them coming out. I felt safe, even though that could be rigged, too."
Reports due Monday
The lottery has hired an outside auditor, KPMG, to make sure the number-generating computers are working properly. But KPMG's $90,000 contract does not ask for a review of possible fraud.
The lottery board's internal auditor has considered fraud, at the request of the lottery board's audit chairman.
But after The Tennessean reported that fraud would not be investigated by KPMG, Sen. Bill Ketron, R-Murfreesboro, co-chairman of the legislature's lottery oversight committee, asked the state Comptroller's Office to do so.
Lottery and KPMG auditors are expected to report their results at a lottery board meeting Monday.
On Nov. 26, legislators will also discuss the issue.
Legislators fear the glitch incident will undermine player confidence and lead to lower sales. That would affect how much money the lottery raises for scholarships and other educational programs.
"Numerous people have come up and told me they are not buying them anymore," Ketron said.
Lottery sales did drop off in August, the analysis shows. However, sales figures for prior years show similar August dropoffs, and both the Cash 3 and Cash 4 games pulled in more money during and after the glitch than in either of the prior two years, lottery figures show.
"That's very encouraging," Ketron said.
If anything, news stories about the glitch improved sales, Hargrove said. She added that the $67.7 million returned to the state for the quarter ending Sept. 30 was almost 9 percent higher than in the same quarter a year ago.
Source: Tennessean
