On the BlizzCon 09 we saw that there are still Starcraft players without Internet playing with Bots and relating on AI.

Not only players without Internet but people who just want to improve themselves. So the question is:
How much did the Bot improved in Starcraft II? He had 11 years to practise and become “better”. Better does not mean harder and more difficult.

A perfect bot would be an opponent where you can’t tell whether it’s a real human or a bot. How does the Starcraft II bot behave himself?

First we have the chat management. For example a “GG” from the Bot in the All Chat if he is “sure” that you have no chance of winning or other phrases like “lucker” or “nice move”.

So actually the Bot will spam you during the Game to break your concentration.  The Bot may also answer on some of your answers or insults.

The next feature is the scouting. We all know how unfair it was in the original Starcraft. You play with “blacksheepwall” playing on a 4v4 map against one bot and you have revealed map and the Bot doesn’t need to scout.

He just attacks your base without any questions about your position. Well, not any more. In Starcraft II Bots have to scout first and find your base before attacking. That’s a temptation to kill the scout and delay the opponents aggressive move. Who knows? It could work and be a great exploit against Bots.

Another feature is the learning from his opponent. If you play 5 games with Zerg against the Bot and you always go on

rush the Bot will notice it and in the next game he will go on early defense to protect himself.

This learning feature offers a lot of undreamed-of possibilities to make the Bot the “korean nerd” of tomorrow and there we have the ancient

propaganda of  the “Machine built by human destroying their creators” (e.g. Terminator). We have for example a chess AI that can beat the world champion in chess. So we could see a bot pwning David Kim or Matt Cooper. It is possible.