This will be the final week’s post as I am going to explain how well the Minimax algorithm runs by battling it against the Negamax algorithm.

Before continuing, I am going to tell that there will be 3 results from each battle. The 1st one is the duration, which means the duration of the time taken by the AI during each decision-making. The 2nd one is the expanded node, which is self-explanatory. The 3rd one is the winning result.

In depth 4, it can be seen that both of the AI only took a fraction of a second to make a decision. Both expanded their nodes at around 500+. The result was the Negamax won the battle by 49 VS 21.

In depth 7, it can be seen that both of the AI still only took a fraction of a second to make a decision. But, the result was quite different at the expanded nodes, in which the Negamax expanded its nodes at 1104 only while the Minimax went above 1900. Nonetheless, The result was the Minimax won the battle by 31 VS 39.

In depth 12, it can be seen that most of the time, both of the AI took a fraction only in their decision-making, but there were some moments where they took 2-5 seconds for making a decision. The Minimax seemed to be the most unstable one compared to the Negamax where there was only 1 occurrence of >1 second decision-making. Both of the expanded nodes were also different by a great number. The Negamax only expanded at around 6800+ while the Minimax blew itself by going up through the 21000+ mark. The result also showed that the Negamax still won the fight by 39 VS 31.

Conclusion

By going through the result of 3 battles, it can be concluded that the winner is the Negamax algorithm. Not just the result, but the time and expanded nodes generated from the overall decision-making is lower than the Minimax.