The California governor’s race and all its key House races will feature a Democrat against a Republican in the fall campaign. What would normally be an unremarkable statement counts as breaking news for California.

On Tuesday, the state selected fall candidates for all its statewide offices, one of its two U.S. Senate seats, all its U.S. House seats and most state legislative seats. Seven other states voted in the primary season’s “Super Tuesday.” But California got special attention because its unusual “top-two” primary system threatened to shut Democrats out of the fall campaign in some hotly contested U.S. House races.

What is the top two? Here’s what you need to know.

California’s ‘top-two’ primary system, explained

The “top two” is a radically open primary system that turns the normal approach to primaries on its head. In a traditional primary, voters decide which party’s primary they will vote in. The top vote-getter in each party primary goes on to represent that party in the fall election.

California’s primary breaks with this approach in three significant ways.

First, it blows up all the boxes. Every voter gets the same ballot and can choose any candidate for each office regardless of party, as in general elections.

Second, only the top two vote-getters advance to the fall election, regardless of party. So a Democrat could face a Republican, a Republican could face a Republican, a Democrat could face a Libertarian, and so on.

Finally, the top-two system doesn’t treat these two candidates as officially representing their parties. On the ballot, candidates are said to have a “preference” for a party.

All told, it’s more like a first-stage general election than a traditional primary.

[Democrats are contesting more state legislative seats than they have in decades.]

Where did this unusual approach come from?

California has been experimenting with its primary system for more than a century. Generally, these experiments have been aimed at breaking up the parties’ control to elect a different kind of candidate — one more flexible, moderate and committed to solving problems.

The top two’s quirks react to California’s last open primary experiment, the “blanket,” which the U.S. Supreme Court struck down in 2000 in California Democratic Party v. Jones. Like the top two, the blanket gave the same ballot to all voters, letting them choose any candidate they wanted for each office. But like a traditional primary, the blanket just advanced the top vote-getter within each party. To the court, this created special problems. The justices believed that if a party was choosing its official representative, it had the right to exclude anyone it wanted to from the decision.

Many of the peculiarities of the top two — like designating candidates with a party “preference” and advancing the top two candidates, regardless of party — were designed to ensure that the fall candidates would not be viewed as official party standard-bearers. The top two has passed all court challenges thus far.

Does the top-two system work as intended?

For the top-two system to yield moderate candidates, a large number of voters should be willing to vote against their party for a different kind of candidate. But research shows that voters still largely vote along party lines. When there’s a same-party contest in the fall election, large numbers of voters from the excluded party simply skip that race.

Voters also have trouble distinguishing the moderates from the extremists in the primary stage, making it hard for voters who do cross party lines to pick the most moderate choice.

Nevertheless, the earliest evidence suggested that Democrats in the state legislature became somewhat more moderate after California started using the top-two system, a finding largely confirmed by more recent research.

But some of the same research finds no ideological change in the congressional delegation, among either Democrats or Republicans, or among Republicans in the state legislature. Washington, another state that adopted the top-two system, hasn’t seen much change, either. So while the top two has accomplished some of its goals, the effect is limited.

Voter loyalty to parties complicates the top-two system

The fact that top-two voters still choose among candidates of their own party can lead to results where the majority’s preference doesn’t win. In a district with about the same number of Democratic and Republican voters, the party with a larger number of credible candidates could split its vote more ways and potentially fail to make the top two — even if that party received more votes overall. Democrats were very worried about just this result in the U.S. House districts mentioned above. They avoided that result this time — but only barely in District 48, where the second-place Democrat beat the next-best Republican by about 1,200 votes out of more than 100,000 tallied so far.

The same party loyalty arguably prevented the more moderate Democratic candidate from advancing to the fall in the governor’s race. Lt. Gov. (and former San Francisco mayor) Gavin Newsom has always been the front-runner. His strongest Democratic opponent was Antonio Villaraigosa, who ran to Newsom’s right and even drew endorsements and money from Republicans. But Villaraigosa finished a distant third behind Newsom and John Cox, a Republican who has never held elected office and who spent a fraction of the money (but was endorsed by Trump). Villaraigosa could not draw enough Republican voters across the aisle.

How might California surmount the problem of a party’s candidates being shut out of the fall election? Write-in candidates are banned from the fall campaign; I have suggested changing that, so that if a potentially competitive party doesn’t make the ballot, its voters can still have their say. Others have recommended ranked choice voting toward the same end.

Whatever the answer, the top-two system is probably here to stay in California. In December, the PPIC Statewide Survey found 60 percent of likely voters thought the system was mostly a good thing and 26 percent considered it mostly a bad thing.

Of course, it also hasn’t had much effect. The system could have larger effects if candidates and voters adapt — but for now, Golden State politics continues much as before.

Eric McGhee is a political scientist and research fellow at the Public Policy Institute of California, where he focuses on elections, legislative behavior, political reform and surveys and polling.