Okay, so I’ve been seeing this “Paulo Costa curse” thing floating around the MMA community, and I was like, “What the heck is that?” I got curious and did some digging, and decided I will try to make something of it.

My Deep Dive into the Paulo Costa Curse
First, I spent, like, a solid hour just scrolling through Twitter, Reddit, and some MMA forums. I wanted to get a feel for what people were saying. It seemed like it started as a joke, but then some people started taking it kinda seriously. The general idea is that whenever Paulo Costa posts something, especially predictions, the opposite happens.
So, I started to keep track. I literally made a note on my phone. Every time Costa tweeted about a fight, I wrote down his prediction. It wasn’t anything fancy, just a quick:
- Fighter A vs. Fighter B
- Costa’s Pick: Fighter A
Then, I watched the fights. I didn’t change my own predictions or anything; I just wanted to see if Costa’s picks were consistently wrong.
Initially, it was 50/50. Some of his picks were right, some were wrong. It was nothing remarkable, really. It seemed random, and I nearly gave up and was like, “This is dumb.”
But I had a lot of free time and I kept at it. I continued to jot down the picks and the results. I even added a little column to my note: “Result (Did the opposite happen?).”

After a few more weeks, I noticed that the win rate for his prediction is so bad. It was actually starting to lean towards the “curse” being real. More often than not, the fighter he didn’t pick was winning. I got more fighters and fights to put on my record.
Finally, I decided to tally up the results after a couple of months. I had a decent sample size – maybe 20 or 30 fights. And, I’m not gonna lie, the “curse” seemed pretty real. The percentage of times the opposite of Costa’s prediction happened was, without a doubt, greater than a random coin flip will give you.
I ended up concluding that there might be a case for this “curse”. I don’t really believe in curses, but the data I collected was pretty interesting, and the result of the curse is indeed real. It was a fun little experiment, anyway!