In the unwieldy Republican presidential field, where attention is as important as money, there was supposed to be a formula for an underappreciated candidate to break out.

It went like this: Excel in the second-tier, undercard Fox debate early in August, get a bump in the polls to break into the top 10 candidates, and arrive on the main stage for the next debate, hosted by CNN in September.

Carly Fiorina, the former Hewlett-Packard CEO, unquestionably has done the first two, but the last may be beyond her power. It requires overcoming CNN debate criteria that couldn’t be more harmful to her if she had shot CNN honcho Jeff Zucker’s dog.

CNN tried its best back in May to come up with fair, transparent standards for who will occupy the 10 slots in its prime-time debate. It’s just that in the real world they make no sense.

Consider the perversity of the CNN criteria. They will almost certainly exclude Fiorina, even though she is seventh in the current RealClearPolitics national polling average, ahead of John Kasich, Mike Huckabee, Rand Paul and Chris Christie, among others; even though she tied for seventh in CNN’s own national poll in mid-August; and even though she has been surging in the early states, popping up to third place in the latest Iowa and New Hampshire polls, ahead of both Jeb Bush and Scott Walker.

At this rate, Carly Fiorina will hold the dubious distinction of being the strongest primary candidate excluded from a presidential-nomination debate in recent memory, although she will get the consolation prize of the CNN undercard event.

What happened? CNN decided to use polls going all the way back to July 16, weeks before the first Fox debate on Aug. 6 scrambled the field. This reaches back to a period when Fiorina was routinely polling at 1 or 0.

On top of this, CNN is only considering polls from select organizations. Some of these polling outfits or news organizations aren’t doing national tracking polls of the Republican race; one of them — McClatchy-Marist — hasn’t done a poll since it said it doesn’t want its surveys used to determine debate eligibility; and others seem likely to wait until after the CNN debate to do their next survey.

All of this means, perversely, that there will probably be more polls from before the first debate included in the CNN formula than after the debate. So, in effect, Fiorina’s performance in the first debate is wiped out.

She won the undercard event by acclamation with a sharp, forceful performance, but she could have done somersaults on stage, recited Daniel Webster’s “Second Reply to Hayne” from memory, beaten Rick Perry in an arm-wrestling contest and played Mozart’s Violin Concerto No. 3 in G Major — and still be left out in the cold by CNN.

There is an easy fix for this, which is for CNN to acknowledge its criteria have played out differently than it expected and to put more emphasis on recent polling. Other candidates on the bubble will cry foul, but who can object to a debate that features the current top candidates now, rather than the top candidates from five weeks ago?

There is no good way to handle a field of 17. If it were up to me, I’d have the candidates draw straws and randomly split them up into two debate groups appearing back-to-back in prime time.

Admittedly, even this arrangement would have a downside in throwing no-hope candidates together with the candidates who have built serious campaigns and gained a following.

What is manifestly unfair is to watch a candidate rise from the very low single digits into more serious contention as she begins to catch on with voters, and then leave her out of the main event regardless. That is what will happen to Carly Fiorina, unless CNN relents.

(Tuesday evening, CNN announced it changed the criteria for the Sept. 16 debate, which is expected to give Fiorina a better chance of appearing.)

Rich Lowry is the editor of the National Review magazine of conservative commentary.