What would Schoolhouse Rock! have to say about the reconciliation process? The old animated educational short was a useful introduction to “regular order” in Congress. Add a little additional knowledge on committees, filibusters, hearings, and lobbying, and you’d have a working basic understanding on how laws are passed in the United States.

That’s regular order. What’s happening on the Senate floor now is nothing close to that.

As Senate Republicans try to push through a law replacing or repealing Obamacare, they are relying on a byzantine set of procedures and tactics that are often indecipherable for the senators themselves. Majority Leader Mitch McConnell’s strategy to avoid a sure Democratic filibuster involves the reconciliation process, which itself necessitates tricky things like Congressional Budget Office scores, parliamentarian rulings, and the Byrd rule. Despite Arizona Senator John McCain’s rousing speech Tuesday urging a “return to regular order,” two procedural votes later, it’s clear that regular order isn’t coming back soon.

One of the chief points of confusion in this process is the issue of what can and can’t pass this way. The reconciliation process can only be used for legislation that has a direct and substantive effect on the budget. The legislation also has to save at least $1 billion for each committee involved in its funding authority. That means in order to reach the floor by reconciliation, and thus face only a 50-vote majority vote, the provisions of a bill must be scored by the Congressional Budget Office and then those provisions reviewed by the Senate parliamentarian, who makes advisory decisions on what can pass this way.

But, as the former parliamentarian Alan Frumin told NPR, “[Reconciliation] was not designed for major policy changes.” And it certainly wasn’t made for the current legislative procedure, where McConnell is mixing and matching multiple different frameworks and amendments via consensus-building alchemy. So exactly where the CBO and parliamentarian fit within that process and its multiple iterations of legislation-building is a bit murky.

The process started with a budget resolution, which in this case dictated that the Senate Committee on Finance and the Senate Committee on Health, Education, Labor, and Pensions had to come up with a reconciliation bill that saved each committee $1 billion over the following decade. The bill that triggers reconciliation has to be scored by the CBO, and the parliamentarian then rules which provisions under that score can be passed by a simple majority.

In practice, that process began in full Tuesday afternoon, when McCain’s dramatic speech prefaced the majority-vote motion-to-proceed on formally starting reconciliation with the adoption of the House’s American Health Care Act. That vote kicked off 20 hours of debate, during which McConnell hopes to use amendments to find a bill that makes 50 Republicans happy.

According to the former CBO director Douglas Holtz-Eakin, however, it’s that very reliance on amendments that makes McConnell’s path so precarious. “Because the amendment would become part of the underlying bill, the new bill would have to satisfy those same reconciliation instructions,” Holtz-Eakin told me. “So the CBO should provide an estimate on each amendment so that they know that they're in compliance.” Major new amendments that haven’t been scored by CBO need 60-vote points of order to suspend the Budget Act requirements that dictate reconciliation rules, and then a second simple-majority reconciliation vote. In practice, that means most amendments introduced during discussion that haven’t received CBO scores would need 60 votes to pass—and they’re unlikely to be able to get them.

That wrinkle was on display Tuesday night, when McConnell used secondary amendments to essentially bring the Better Care Reconciliation Act—the Senate’s version of the Obamacare replacement—to a vote. Critically, McConnell’s version included two additional amendments. The first, from Texas Senator Ted Cruz, would allow skimpy plans to be eligible for premium tax credits alongside beefier Obamacare plans. The second, from Ohio Senator Rob Portman, would use Medicaid “wraparound” funds to provide assistance for low-income people. Although there had been at least some discussion of using alternative assessments of the Cruz amendment in lieu of the CBO score, that plan appears to have been scrapped.

Since neither of those amendments had been scored by the CBO, the BCRA vote needed to clear a 60-vote point of order, which served as a proxy for the final vote. It did not pass.

The scoring and parliamentary processes were both at the forefront on Wednesday as well. About eight hours into the debate period, McConnell moved on to stage two of his plan: a straight “repeal and delay” of Obamacare. Although the overall “repeal and delay” plan has been scored by the CBO and could pass reconciliation via a simple majority, Republican leaders tacked on an amendment banning the federal funding of abortion that has already been rejected by the Senate parliamentarian, which meant that Democrats had the chance to call a 60-vote point of order. They chose not to.

That means that Wednesday’s vote, which failed, was purely theatrical. McConnell knew straight repeal didn’t have the votes to pass, even with a simple majority, which is why he took the risk of attaching the amendment. Democrats didn’t call a point of order, which means they also knew the repeal-and-delay plan would not pass.

As the debate clock winds down, it is possible that new amendments introduced on the floor could receive scores in time to make it into a final bill. According to Holtz-Eakin, the CBO is on high alert during showdowns like these, and can score some provisions quickly. “During Medicare Part D,” Holtz-Eakin says, “at about 5:30 every morning, then-Senator Max Baucus would call my cell phone and tell me what I had to have ready. There was a regular workload of being ready to be able to give the Senate scores on amendments, and we did.”

Still, it seems that the real trajectory—and McConnell’s plan all along—is to steer debate to the vote-a-rama period and then adopt a final bill resembling a “skinny repeal” that rolls back Obamacare’s mandates and some of its taxes, but doesn’t touch Medicaid. According to Holtz-Eakin, during the vote-a-rama, smaller amendments that don’t change the overall shape of the bill can bypass CBO oversight and be accepted by Budget Chairman Mike Enzi.

The main game, though, is the skinny repeal. Even with one or two holdout votes, and maybe aided by a few amendments along the way, McConnell could steer a skinny repeal to the House or to a conference committee. The conference committee could again add any provisions that have been scored—including the Medicaid rollbacks that would make it a “full” repeal.

Notably, this process means that full CBO scores of entire drafts of legislation might not be forthcoming until bills are already passed. While the piecemeal approach works for the most part for rough budgetary analysis—which is the purpose of CBO scores in reconciliation—it means that if a bill does pass the Senate, its effects on coverage, premiums, and larger macroeconomic effects won’t all be known. And given previous CBO scores, it seems likely that something approaching a skinny repeal could destabilize exchange markets and reduce coverage by at least 15 million people.

By ensuring maximum chaos, keeping amendments secret, and making the CBO and the parliamentarian play catch-up during the 20-hour reconciliation debate time, McConnell trades away transparency and clear-eyed predictability for the chance to pass his bill. By mixing and matching provisions, he may be able to stumble upon 50 votes. But at this point, nobody—not even McConnell—really knows what the emergent Obamacare repeal options will do.