Peter Hart, a top Democratic pollster, assesses the 1988 presidential campaign this way: ``Anybody who carries California and Texas wins the election.''
Using Mr. Hart's own polling data, that assessment could be discouraging for Democrat Michael Dukakis.
In Texas, Hart's newest surveys show George Bush building a 10-point lead, 51 percent to 41 percent. Earlier this summer, the candidates were neck and neck there.
In California, where Governor Dukakis was sporting a double-digit margin just a month ago, it has become a tight race, with Mr. Dukakis barely ahead, 47 to 44.
Nationwide, the latest Time magazine poll gives Mr. Bush a 47 to 42 advantage over Dukakis.
At a Tuesday breakfast, Hart conceded to reporters that Bush clearly won the month of August.
Within the next week, Dukakis must move quickly to seize control of the campaign, Hart says. Specifically, Dukakis must turn the day-to-day focus of the campaign toward family issues, toward the economy, and toward the future of the country.
What concerns Democrats like Hart, who was a top adviser to Walter Mondale in 1984, was that Bush made what they consider a major blunder in August - the selection of Dan Quayle - and still gained ground.
If Bush had selected a stronger vice-presidential nominee, he might have opened a lot of daylight between himself and Dukakis, Hart suggests.
Some Democratic analysts, including Hart, now see possible openings for Dukakis in coming days. One involves the economy. The other involves leadership.
Although he has moved up in the polls, Bush so far has failed to convince many voters that he can handle the Oval Office. ``Neither candidate has really gained the confidence of the voters of the United States that he would be a good president,'' Hart says.
A survey of five major states - Texas, California, New York, Illinois, and Florida - finds that 52 percent say they believe Bush would be a good president, while 47 percent believe Dukakis would do a good job. Both numbers are well below the 60 percent considered ``normal'' at this point in the campaign.
If Dukakis seizes the agenda of this race, he can move his numbers past Bush, and potentially break into the lead, Hart says.
Dukakis may also get some help from the economy if the jobless rate goes up again. A trend that shows rising unemployment, higher interest rates, a big trade deficit, and a lackluster stock market could feed unease about the economy and play into Dukakis's hands.
But Hart - who reads election trends through his numbers - admits that the tide is still running in Bush's direction. Dukakis has very little room for error.
In recent days, for example, Bush has opened a huge edge over Dukakis among Southern whites. The vice-president now leads among that group by 31 points, 62 to 31, apparently because there is a growing perception that the economy is doing well, and that the country is heading in the right direction.
At the same time, President Reagan's popularity, already high in the South, is again on the rise in all parts of the country. Bush benefits from that.
Southern whites give Bush a major Electoral College advantage by delivering most of the Southern states to his column. They wipe out any impact from Dukakis's strong edge with Southern blacks.
If Bush carries the entire South, plus the Rocky Mountain states where Republicans always do well, it will force Dukakis into a do-or-die situation where he must win all the remaining ``mega-states'' - California, New York, Pennsylvania, New Jersey, Ohio, Michigan, and Illinois.
Without the South, even a single loss in one of those large states could defeat Dukakis, Hart observes.
That's why Hart argues that the choice of Lloyd Bentsen as the Democratic vice-presidential nominee was wise, for it gives Dukakis at least a fighting chance in Texas, plus Arkansas, Tennessee, and Kentucky.