the MMA-AI-probability-cube shows fighters in specific bouts from the last 2 years. find subspaces, so called neighborhoods,Â  in order to find regions with a large proportion of winners. whenever you have found such a secret Area you can use it to predict future bouts. estimate probabilities and place a fighter in the cube. if he lands in one of your secret subspaces you might know the winner already. THE probabilites are not implied odds from bookmakers but the probabilities learned by mma-ai. thus finding good subspaces is a human extension of the mma-ai capabilities and should lead to EVEN FURTHER improvements for the prediction of winners. the extend of improvement is not evaluated so far. SEE a detailed explanation below. If YOU found some promising areas, please let me know via email or @MMAAIcom1.

Neighborhood X (0-1)
Neighborhood Y (0-1)
Neighborhood Z (0-1)
Neighborhood MAX Size (0-1)

THE CUBE CONTAINS ONLY DATA, THAT MMA-AI HAS NEVER SEEN BEFORE.
x, y, z: Estimated Fighter Win Method Probabilities (as def. above) by MMA-AI
Fighter Wins as Favorite | Fighter Loses as Favorite | Fighter Wins as Underdog | Fighter Loses as Underdog |

## Example

The cube shows fighters in a specific bout from the last two years. Red means the fighter lost,Â  green he won. With this tool you can try to find neighborhoods with a high density of winners, i.e. green points, and use that neighborhood to predict future winners. In order to do so:

1. Define 3 probability dimensions that span the cube.
2. Define 3 center-coordinates of your neighborhood (subcube).
3. Define the max size of your neighborhood.

For example, I pick the dimensions “DEC_A” (the probability that the fighter wins by decision), “FIN_A” (the probability that the fighter wins inside the distance), and “ITD” (the overall probability that the find ends inside the distance). I set my neighborhood center at DEC_A = 0.35, FIN_A = 0.30 and ITD = 0.30 and the neighborhood size to 0.12. I find a neighborhood where almost all bouts are winners (green circles).

Whenever you find subspaces with a large proportion of green circles, you may have found a secret. You can use this subspace to predict winners in the future.