The results are in for UNU Baseball’s picks for August 23rd and a funny thing happened. I’m not talking about the fact that the Swarm predicted the struggling White Sox would beat the powerful Phillies “by a lot” in advance of Chicago’s 9-1 drubbing. The crazy thing is that the Swarm generated 20-25% ROI  in its toughest week yet predicting baseball games.

We started UNU Baseball because we know that predicting baseball games is hard, and so the fact that Swarm Intelligence has allowed the group to correctly identify the winner in 2 out of every 3 games is kind of amazing.  But, of course the difficulty in picking winners means that not every week can be as amazing as last week’s big payday. So, this week, the Swarm went just .500 in picking winners, 7-7-1.

How did this happen? It’s relatively simple. To make a profit, you need to do more than pick winners. You need to know when not to bet.

Part of the motivation behind the UNU Baseball experiment is to find out how important the Swarm’s confidence and conviction are to its ability to make good predictions. We measure this conviction using a metric called Brainpower. Naturally we are still collecting data on all of this, but, as a rule of thumb, if the Swarm’s Brainpower is below 75%, we consider the recommendation to be a very weak one. And, so we’ve advised that bettors stay away from these games, even though the Swarm has identified a winner. 

So, while the Swarm registered 14 predictions, only 8 of those games had high enough Brainpower to recommend a bet. And, just as you might expect, those high Brainpower ( 5-3 ) games dramatically outperformed low Brainpower (2-4) games.

ATS: 2-1

High Brainpower ( above 75%) 5-3

Low Brainpower ( below 75% ) 2-4

1 game too close to call ( game ended up 2-1 )

Overall 7-7-1 record picking winners. But here’s where knowing when NOT to bet is important.

Cubs = $10 to win 6.67 = $16.67

White Sox = $10 to win 11 = $21.00

Mets $10 goes to $0

Bet $ 10.00 to win $ 7.14  = paid $ 17.14

Bet $ 10.00 to win $ 10.90  = paid $ 20.90)

Royal = $ 10.00 to win $ 10.20 = paid $ 20.20)

16.67 + 21.00 + 17.14 + 20.90 + 20.20 = 95.91
$95.91 back on $80 bets = 19.89% ROI ( 20% )

Here’s how the games turned out:

Blue Jays ( high bp )

Reds ( high bp )

Royals ( high bp )

LOSSES:

Indians high ( bp )

Rockies ( high bp )

You can see the results of my own wagering below. My 8 bets reflect the recommendations we made yesterday.

---