This file contains two AIs (Raik/Raik II) instead of just the one. While the author's intention of saving space on Heaven's servers is no doubt admirable, there doesn't seem to be any real difference in the gameplay mechanics of each AI, except a faster castle time in Raik II. They are slow, both in advancing and building an army. demurring the fact they build walls, the game would be over in the feudal age. I decided to test the AIs using a gigantic AI wars game. A team of two Raiks, a team of two Raik II's, a team of two Zycats, and a team of two Swans. Both of the Raik AIs was practically dead before the castle age, both Swans attacked them before they could get their walls up. They were soon defeated, closely followed by a Raik II, who was wiped out by Zycat's knights. The remaining Raik II managed to get to the castle age a second before its town center was razed by a swarm of enemy knights. It was a devastatingly bad game for both AIs. On the description page the author writes "They can beat the
normal AI". I beg to differ, as the standard AI won in a seperate game I conducted. It is plain that sloppy military management is what killed them. In Raik II, you wrote a rule telling the AI to train a spearmen when it had 500 food/wood or greater. No AI should wait until it has 500 food until it starts making military, especially a unit as useful and versatile as the spearman. This mistake is repeated throughout the file, only allowing Raik to build military once it has a required quota of resources. I wish I had something a bit more favourable to say, but I would not recommend this AI to anyone except a newbie. The author does show promise, however, and if he continues writing AIs I advise him to visit the Tonto Clan, or AI Scripters, if he has not already.
Posted on 10/12/05 @ 10:42 AM
I decided to test the AIs using a gigantic AI wars game.
To test an AI like this better use 1vs1 arabia. Although Mr.Robink doesn't mention it i think very much that he tested it most of the time with those settings ;).
Btw. if it is true what the author, Mr. Robink, says, that it can beat the standard AI (which i think is true) then you may either accept the guidelines of AokH which are saying you should give at least 3 points for an AI which can beat it or just comment what you do not like about the AI but not rate it.