In Section 5.2, the following note was carried forward mistakenly from the erroneous previous version of the tutorial. The MC player implementation was fixed, making this is unnecessary.
Note: the Monte-Carlo player doesn't seem to be doing much better than the random player... This is because training of a good MC player is VERY compute intensive and we have not done extensive training here. In Bonus 2 below, you can play with a Monte-Carlo Tree Search (MCTS) player that has been trained well and you'll see that it performs much better!
In Section 5.2, the following note was carried forward mistakenly from the erroneous previous version of the tutorial. The MC player implementation was fixed, making this is unnecessary.
Note: the Monte-Carlo player doesn't seem to be doing much better than the random player... This is because training of a good MC player is VERY compute intensive and we have not done extensive training here. In Bonus 2 below, you can play with a Monte-Carlo Tree Search (MCTS) player that has been trained well and you'll see that it performs much better!