Back in April, Mary Pat Clarke said her poll figures were so good, they were worth bragging about.
A poll paid for by the Clarke campaign showed Clarke trailing incumbent Kurt Schmoke by a scant 6 percentage points with a 4.4 percentage point margin of error.
"The incumbent holds a slight advantage that is statistically within the poll's margin of error," Clarke's poll report said. "For the race to be this close . . . indicates a weakness in Schmoke's support that makes this a winnable race for Mary Pat Clarke."
Clarke was delighted by her own poll.
"It definitely looks like a tossup at this point," she crowed to The Sun in late April.
But it is now July. The mayoral primary is less than eight weeks away and The Sun ran a Mason-Dixon poll on Sunday indicating that Clarke trails Schmoke by 15 percentage points.
One might say, therefore, that after months of campaigning by Clarke and the expenditure of hundreds of thousands of dollars, the gulf between her and Schmoke has more than doubled.
So is this the end of Mary Pat Clarke?
Naw. But that's because the results of her own poll were baloney to begin with.
Her poll, presented in a report stamped "Confidential" and then handed out to reporters, did not represent a measure of the true gap between Clarke and Schmoke.
As a Clarke campaign source admitted to me at the time: "This is not a 6-point race. Schmoke is ahead by more than that."
The real margin, I was told, showed Schmoke ahead in the low double digits or roughly where the Mason-Dixon poll shows him now.
So why did the Clarke campaign release the 6-point poll result, but keep secret the more realistic poll result?
Clarke needed money. And most people won't give you money if they feel you have no chance of victory.
Clarke needed a poll that showed her in the ballpark. And so she went out and got one.
I am not suggesting she made the figures up. But what she did do shows you something about modern campaigning and the use of polls:
In the case of the Mason-Dixon poll, pollsters asked 409 likely Democratic voters: "If you were voting today for the Democratic nominee for mayor, which of the following candidates would get your vote?"
That question yielded 47 percent for Schmoke and 32 percent for Clarke. But the poll that Clarke released to the press in April, the one that showed her trailing by just 6 points, was conducted in a much different way.
Her pollsters read statements to 509 likely Democratic voters and then asked them who they would vote for.
The Clarke campaign would not tell me what those statements were. But obviously the statements could go a long way toward shaping the result.
For instance, you could say: "Crime under Kurt Schmoke has increased by 41 percent. Mary Pat Clarke has battled hard to keep your property taxes down. So who would you vote for if the election were held today?"
Clarke got the result she wanted: A single-digit gulf between her and the incumbent. And, perhaps, that result did help Clarke raise money.
But it was a gamble. Clarke knew that eventually the media would get around to doing an independent poll.
She hoped, however, that by then her campaign efforts really would elevate her to a single-digit gap.
They have not.
But that may not be the worst news the Mason-Dixon results now present to Clarke.
Back in May, Clarke campaign manager Cheryl Benton said that Clarke expected to get 32 percent of the black vote and 68 percent of the white vote in order to win the Sept. 12 Democratic primary.
"That's doable," Benton told me.
Perhaps. But so far Clarke is not doing it.
According to the Mason-Dixon results, Clarke is getting just 13 percent of the black vote and 57 percent of the white vote.
So where does this leave Clarke?
It leaves her saying what politicians usually say:
Polls are accurate when they show you close but hooey when they don't.
"One of the things we know is that polls don't vote," Clarke said Sunday.
Now she tells us.