| |

Everything you need to know for Yan vs. Dvalishvili

Last week, we broke down the Jones vs Gane championship fight and provided insight on how the fight was going to go down and what was the best price for each fighter.

BCDC Data Scientist, Brandon Bergstrom, is 33-14 with over 20.35U on these 50/50 or main event fight picks. We will be breaking down this main event in a similar fashion with good statistics, statistical modeling, machine learning models, and finally best price for the fighters.

Betting Stats

Petr Yan suffered a defeat last time in the octagon against Sean O’Malley but has won 6 of his last 9 fights. Merab Dvalishvili has won 8 fights in a row and has won 6 of his last 6 fights. Dvalishvili’s reach is 68 inches; Yan is 6-3 against Bantamweights with a reach of 68″ or more. Yan’s reach is 67 inches; Dvalishvili is 4-1 against Bantamweights with a reach of 67″ or more. Merab Dvalishvili is +220, which is cheaper than 100.0% of his previous UFC fights (10 fights total). Bantamweights priced between -240 to -280 are 9-1 in the last 10 fights.(Petr Yan)
Bantamweights priced between 240 to 200 are 2-8 in the last 10 fights.(Merab Dvalishvili) Yan in his last three fights averages 22 minutes in the octagon.
Dvalishvili has a record of 5-1 when facing an opponent with a smaller reach.

Our type of fight models project this to be a striking fight, which should favor Yan. Also, our models project a knockout in the second round… for anyone looking for pizza money action by machine learning models.

Decision Percentage will be 0.34304518

Knockout Percentage will be 0.43561882

Submission Percentage will be 0.22133593

Round 1 Percentage will be 0.030275304

Round 2 Percentage will be 0.19669165

Round 3 Percentage will be 0.118844725

Round 4 Percentage will be 0.012509789

Round 5 Percentage will be 0.13348752


Merab Dvalishvili is:
2-0 against fighters with 2 or more for Submission Total Attempts.
1-1 against fighters with 2 or more for Reversals.
2-0 against fighters with 27 or more for Control Time.
2-0 against fighters with 176 or more for Total Time in Octagon.
2-0 against fighters with 8 or more for Total Wins.

Petr Yan is:

3-3 against fighters with 4 or more for Sig. Str. Landed 1 min.
4-1 against fighters with 10 or more for Sig. Str. Attempts 1 min.
2-1 against fighters with 3 or more for Sig. Str. Head Landed 1 min.
3-1 against fighters with 8 or more for Sig. Str. Head Attempts 1 min.
3-3 against fighters with 3 or more for Submission Total Attempts.
3-2 against fighters with 144 or more for Total Time in Octagon.
3-3 against fighters with 8 or more for Total Wins.

This graph shows you stats to help you see an edge in the fighters.

Statistical Modeling

Statistical Simulation Prediction
Fights simulated: 1000
Wins for Petr Yan 11
Wins for Merab Dvalishvili 989

Using Bayesian Simulation and not machine learning models. beforethecagedoorcloses is 64% sure that Dvalishvili is the better fighter with a risk of being wrong at 12%.

Two weeks ago, our statistical simulation predicted Gane (7-7) and our Bayesian Simulation predicted Jones (9-5)

Added another statistical simulation model using Monte Carlo. It takes in a both fighters striking per minute, takedowns per minute, knockdown and submission per minute and number of rounds. Jones was the prediction last time around. (1-1)

We are projecting in the monte carlo simulation model that Merab Dvalishvili will win all five rounds if it goes to a decision.

Our Machine Learning Model currently has a prediction accuracy of 66.92%. Its prediction for this fight is that Petr Yan will win.

Best Price for the Fight

The safe bet is Yan with the models. However, I am not willing to risk 260 dollars to win 100 on Yan. Dvalishvili trains with Sterling and I believe will dominant in the main event rounds. Dvalishvili will win by decision and will tire out Yan. I don’t like going against the models because it is the best in the business but man. All Merab all day it is the safer bet and has a clear advantage with his cardio/grappling.

Share this…
Share on facebook
Share on twitter

Similar Posts


Leave a Reply

Your email address will not be published. Required fields are marked *