It's the most powerful, questionably scientific tool the internet has ever developed. It can make or break a person's entire career on Weird Twitter and small areas of the internet only dorks care about.

They call it the Ratio, and it's coming for CNN political pundit Chris Cillizza.

The term, which became popular in early 2017, measures a tweet's success (or more accurately its failure) based on the number of replies it gets compared to the number of responses. The higher the Ratio, the worse the tweet is estimated to be. For a long time, Cillizza, a pundit who has mastered the art of the bad take, was accused of having one of the highest Ratios on the political internet and some of the least popular tweets.

Now we've got the data to prove it.

Cloudera Fast Forward Labs is an "artificial intelligence lab based in Brooklyn that tackles hard problems like natural language understanding, probabilistic programming and who is the worst on Twitter." Previously, Fast Forward discovered that Paul Ryan had the highest Twitter Ratio of the seven major politicians they studied—a fact that brought tears of joy to some tired #Resistance eyes.

We had the lab analyze Chris Cillizza's tweets and compare them to pundits with similarly sized followings across the political spectrum: Joy Reid, David Frum, Meghan McCain, Matthew Yglesias, Ana Navarro, Judge Jeanine Pirro, and Jonathan Chait. We also threw in Trump, who's more of a pundit than he is a President, and Cher, the people's president. Here's what they found.

Chris Cillizza and Ana Navarro had two of the most hated tweets. Taking a look at the text of the tweets, you can easily understand why.

Cillizza's Ratio here is 12.77. A bad Ratio is considered to be anything greater than 2:1, so...

It wasn't the email story. It was how she handled the email story.https://t.co/iKzFPTvP09

— Chris Cillizza (@CillizzaCNN) September 11, 2017

Navarro's miraculously managed to be somehow worse, with a Ratio of 25.27.

Paul Ryan is a decent, thoughtful man. Compassionate. He's playing w/a bad hand- being Speaker w/a "GOP" POTUS who is a nutcase. Not pretty. https://t.co/KC9U0Niyns

— Ana Navarro (@ananavarro) June 11, 2017

Still, the tool only tells part of the story because it only catches tweets that have 50 retweets or more. Most of these pundits are heavy tweeters, the lab explained to Mashable, and tweets with just a few retweets frequently don't have enough data to be statistically meaningful.

Yet many of Cillizza's tweets, unlike Navarro's, are so despised that they aren't even captured by the algorithm. Take this tweet, which wasn't included in the sample because it only has 23 retweets. It just happens to have 1.1k responses (Ratio: 47.8) and most of them are, um, well:

Why booing Ivanka is wronghttps://t.co/BjBwFlss0X

— Chris Cillizza (@CillizzaCNN) April 25, 2017

She praised an admitted sexual assaulter & sexual harasser, she's part of his administration, she's NOT just "1st & foremost his daughter."

— Mark Hughes (@markhughesfilms) April 25, 2017

Because it’s impossible to boo any more after losing your voice booing @CillizzaCNN. https://t.co/BjF5Fk8rJe

— Jesse Spector (@jessespector) April 25, 2017

Or this one, with a Ratio of 18.13.

I wrote about why Hillary Clinton should be blamed for her losshttps://t.co/cJ18NsYPX0

— Chris Cillizza (@CillizzaCNN) June 1, 2017

Ratio: 21.35.

The @TODAYshow did Ted Cruz porn video and Seattle mayor resignation in its first 10 minutes.

Banner day for politicians!

— Chris Cillizza (@CillizzaCNN) September 13, 2017

The algorithm also doesn't capture tweets and replies, where Cilllizza has produced some of his most despised material:

Ratio: 39.7.

I ask again though: Why can't Trump be praised for delivering a good speech full stop? https://t.co/vvV7eMavsJ

— Chris Cillizza (@CillizzaCNN) March 1, 2017

Because he read it terribly, it was written for 6th Graders, it mapped out a route to Despotism, and you guys all fell for it - as usual. https://t.co/rWdvp3klxd

— Keith Olbermann (@KeithOlbermann) March 1, 2017

Or this one, which wasn't included in the sample because it happened too long ago to be included.

.@DanaBashCNN 👠 game: A++++ pic.twitter.com/byPHt8fjZ3

— Chris Cillizza (@CillizzaCNN) April 13, 2017

Compared to other pundits of his stature sampled, Cillizza has the highest Ratio with the exception of Meghan McCain. Part of that Ratio can plausibly be explained by McCain's fanbase, who doesn't always retweet her stories but who frequently likes them or posts favorable comments like this one below:

Megan's ass is perfect. Her body is perfect! She's the most beautiful woman that I've ever seen. And her hair is banging! Love ya, Meghan!

— Barry Pardee (@NoClintonOrTrum) July 13, 2017

Here's everyone's average Ratio of their last 3,200 tweets, per the lab.

CillizzaCNN: 0.913552

JoyAnnReid: 0.259206

JudgeJeanine:0.471165

MeghanMcCain: 1.017996

ananavarro: 0.484394

cher: 0.468431

davidfrum: 0.268403

jonathanchait:0.214872

mattyglesias: 0.274737

realDonaldTrump:0.716319

You can better see these disparities on a line graph. Note: Cillizza's analyzed tweets, in green, only go back three months. Cillizza tweets more frequently than most included in this sample, and the algorithm only catalogues each account's most recent 3,200 tweets because of Twitter's API.

Cillizza v. Navarro

Cillizza v. Reid

Cillizza v. Trump

What does all this mean for Cillizza's career? Meh, probably nothing. Even though he's losing at Twitter, he's winning where it matters most: on site.

His Reddit AMA attracted more than 100,000 views and his most trafficked column, "Donald Trump Just Had the Weirdest Cabinet Meeting Ever," received over 3,000,000 unique visitors in June. His other most popular columns average somewhere between 1 and 2 million visitors.

That being said, one of my most critically acclaimed articles, "Feminist blogger makes bread with yeast from her own vagina," received over 107,000 shares. I'm fully confident that approximately .000000001 of visitors thought that was a good idea, yet it was never given a ratio.

There are so many awful takes out there just waiting to be properly scored. In the meantime, scrappy ol' Twitter must do what it does best: dunking on bad tweets, one Judge Judy GIF at a time.