UNU Baseball Knocks it out of the Park: 60% ROI Against the Spread

UNU BASEBALL KNOCKS IT OUT OF THE PARK: 60% ROI Against the Spread 

Four weeks into the UNU Baseball experiment, the Swarm Intelligence is already hitting playoff form

What is the UNU Baseball Swarm? This is an anonymous group of baseball fans from all across the internet who meet regularly online to harness the power of Swarm Intelligence to predict major league baseball games. Swarm Intelligence is an innovative artificial intelligence technique modeled after flocks of birds, schools of fish, and swarms of bees, and has been proven to amplify the wisdom of groups by pooling the knowledge and intuition of all the participants.  So, UNU Baseball is all about using the power of swarming to make good predictions and, in turn, make some money.

Want to join the group? Just drop me a line – joe@unanimousai.com – and I’d be thrilled to add you.

The Swarm was asked to make a wagering recommendation for all 16 Major League Baseball games being played on Tuesday, August 16th. The process for each question takes just a few seconds, and allows the Swarm to intuitively express a collective opinion on the game. If the Swarm finds consensus around one team winning “by a lot,” that is a recommendation to bet on that team Against the Spread (ATS) which pays better but is more risky, while “by a little” suggests a Moneyline (to win) bet and “too close to call” means that the Swarm makes no recommendation on the game. All of the Swarm’s picks are sent to the UNU Baseball mailing list before the games, and posted to the UNU blog for anyone to see.

Mets to Win

 

Week 4 was another excellent week for the Swarm. How good? I’ve included a full rundown of all the predictions at the bottom of this post, but let’s take a look at the numbers.

  • 11 correct predictions in 15 predictions. 1 game deemed too close to call.
  • 73.3% accuracy in picking winners this week, 67% overall accurate through four weeks
  • 38.% ROI if you bet all 15 games according to the Swarm’s recommendations
  • 75% accuracy Against the Spread  — 3-1 where an ATS bet recommendation was made
  • 60% ROI if you bet only the four ATS games

There’s a lot of information to unpack there, but the numbers are worth digging into just a little bit more. One of the biggest reasons for making these predictions is to see how the Swarm compares to the average individual. After all, the big idea behind Swarm Intelligence is that thinking together allows the group to improve on the decision-making of the individual. And, yet again, the Swarm did exactly that. The average individual made 6 incorrect predictions out of 16, versus only 4 for the Swarm. Since one of the keys to being a good gambler is not only having a feel for who will win, but also when not to bet, those 2 extra losses have a disastrous effect on the average individual’s ROI.

CIN by a lot

The Swarm was especially accurate this week when they were feeling confident. Not only did they only decline to pick just one game, but the Swarm’s conviction in each recommendation was almost directly correlated to the accuracy of the pick. While this is a bit of an oversimplification, the Swarm’s conviction is best expressed in the Brainpower stat displayed after each replay.

The incredible thing is that the Swarm made a weak recommendation only 4 times and yet those lower than 70% Brainpower outcomes accounted for three of the Swarm’s four incorrect predictions. ( The fourth low Brainpower game was the Toronto game, which was a correct prediction, and the low Brainpower was the result of the Swarm debating between betting Toronto ATS or on the Moneyline. The Blue Jays won by 6 runs, so either bet would have won. ) That’s why looking at replays like the ones included here gives insight into not just what the Swarm said, but how it arrived at its conclusion.

In short, a smart bettor might be inclined to throw out those weaker recommendations as unworthy of wageringIf you do that, the Swarm could have finished an incredible 11-1 overall!  In fact, that’s essentially what one wise UNU user chose to do. As you can see in the screenshot below, this user chose to bet on the games where the Swarm expressed the highest confidence. His personal cutoff was 75% Brainpower, which included the seven games below. As you can see, six out of seven of those picks were accurate, turning $70 into $118.71 — an incredible 70% ROI!

4fd429f5-40d6-4f20-ba67-4dc695e5af6b

One of the reasons the Swarm continues to do so well is because we have a great group of people who meet up every week. If you love baseball, we want you in the Swarm!

Just drop me a line at joe@unanimousai.com and I’d be thrilled to add you to the group.


UNU’s Picks for 8/16:

WINNING BETS

MIL: (Anderson 7-10 4.93 ERA) at CHC: (Hammel 12-5 2.90 ERA)
UNU says: Cubs by a lot — ATS bet
Brainpower 95%
http://go.unu.ai/r/157585
Bet $10 ATS to win 9.09 = $19.09

SD: (Jackson 3-2 4.19 ERA) at TB: (Snell 3-5 3.18 ERA)
UNU says: Tampa Bay by a little
Brainpower 86%
http://go.unu.ai/r/157580
Bet $10 to win $5 = $15.00
 
STL: (Garcia 9-8 3.93 ERA) at HOU: (Keuchel 7-11 4.56 ERA)
UNU says: St. Louis by a little
Brainpower 81%
http://go.unu.ai/r/157589
Bet $10 to win $15.50 = $25.50
LAD: (Maeda 11-7 3.31 ERA) at PHI: (Velasquez 8-4 3.94 ERA)
UNU says: Dodgers by a little
Brainpower: 77%
http://go.unu.ai/r/157568
Bet $10 to win $6.21  = $16.21
NYM: (Syndergaard 9-7 2.75 ERA) at ARI: (Shipley 2-1 2.96 ERA)
UNU says: Mets by a little
Brainpower 76%
http://go.unu.ai/r/157593
Bet $10 to win $6.21 = $16.21 
 
MIA: (Urena 1-3 6.69 ERA) at CIN (DeSclafani 6-1 3.11 ERA) 
UNU says: Reds by a lot — ATS bet
Bet $10 ATS to win $16.70 = $26.70
MIN: (Santana 5-9 3.62 ERA) at ATL (De La Cruz 0-5 4.09 ERA)
UNU says: Minnesota by a little
Brainpower 73%
http://go.unu.ai/r/157575
Bet $10.00 to win $7.58 = $17.58 
CHW: (Quintana 9-8 2.85 ERA) at CLE: (Kluber 12-8 3.21 ERA)
UNU says: Cleveland by a little
Brainpower: 72%
http://go.unu.ai/r/157569
Bet 10.00 to win $5.26 = $15.26
OAK: (Triggs 0-1 5.50 ERA) at TEX: (Harrell 3-2 4.20 ERA)
UNU says: Texas by a little
Brainpower: 71%
http://go.unu.ai/r/157582
Bet $10.00 to win $8.26 = 18.26 
MIL: (Garza 4-4 4.83 ERA) at CHC: (Cahill 1-3 3.07 ERA)
UNU says: Cubs by a lot — ATS bet
Brainpower: 70%
http://go.unu.ai/r/157560
Bet $10 ATS to win $8.33 =  $18.33
TOR: (Estrada 7-5 2.95 ERA) at NYY: (Pineda 6-10 5.07 ERA)
UNU says: Toronto by a little
Brainpower 60% ( note: BP was low due to debate between “a lot” and “a little”
http://go.unu.ai/r/157565
Bet $10.00 to win $9.09 = $19.09 
LOSING BETS
SEA: (Paxton 4-5 3.53 ERA) at LAA: (Chacin 2-6 6.00 ERA)
UNU says: Seattle by a lot — ATS bet
Brainpower 80%
http://go.unu.ai/r/157596
Bet $10 ATS = $0 
BOS: (Rodriguez 2-5 5.43 ERA) at BAL: (Gallardo 4-4 5.17 ERA)
UNU says: Baltimore by a little
Brainpower 67%
http://go.unu.ai/r/157563
Bet $10 = $0 
WSH: (Gonzalez 8-9 4.24 ERA) at COL: (Bettis 10-6 5.27 ERA)
UNU says: Washington by a little
Brainpower 66%
http://go.unu.ai/r/157591

Bet $10 = $0 

KC: (Duffy 9-1 2.82 ERA) at DET: (Verlander 12-6 3.42 ERA)
UNU says: Detroit by a little
Brainpower 65%
http://go.unu.ai/r/157572
Bet $10 = $0 

TOO CLOSE TO CALL
PIT: (Taillon 3-2 2.85 ERA) at SF: (Samardzija 10-8 4.23 ERA)
UNU says: too close to call
Brainpower: 42%
http://go.unu.ai/r/157599

SWARM on BLUE 3