I got to say this - COTA has one of the best AIs I have seen
Posted: Tue May 30, 2006 11:42 pm
I am sorry, but I just had to say this.
I know you are thinking, "well what would a beta tester say about a game that they are closely involved with?".
Fair point.
You know COTA is not the only game I play. I will not enumerate the other games I play here and point out the flaws in the various AI implementations. But here is what I know:
Most reasonable quality games that I buy (which is the only type I buy, since I usually wait until the game has been out for quite a while), I am able to figure patterns in the AI behavior after no more than six months. Those patterns tend to repeat themselves from game to game such that a few standard formulas are sufficient to beat the AI provided that the scenario author hasn't buried some nasty surprises in the scenario or the force balance is not heavily weighted against the player.
I have been playing/testing the AA engine which powers COTA since 2001. So, I know the engine much better compared to these other games that I already own or purchase from time to time. Yet after five years with the AA engine I have yet to come up with a set of human-vs-AI formulas that guarantee one victory after another. Now, I do win these days the majority of scenarios which I play on the first time around. However, those victories come through careful analysis of the scenario, situation, and terrain --- no formula templates. The techniques which I employ are those very same ones that have shown up in my mini guide threads. For the most part, we are talking about high level plans composed out of building blocks of realistic military tactics. And in my case, there may also be an occassional ahistorical military practice here or there, since I am no grognard (just a game player).
But that is really something when you think about it. Five years and I have yet to arrive at a few basic rules for trouncing the AI every game I play. I have played video games since around 1992/1993 and done a number of gaming guides focused on MP or SP over the years. Those patterns of weakness in the AI found in other games, just don't seem to be there in COTA.
I am not saying the AI is perfect. I write scenario reviews when I play them and often do have criticisms, but those criticisms are more often than not about how the AI played a specific situation and not how it handles an entire class of situations.
Like I said, I am a beta tester and you can write this off as your usual "fan boy" ranting if you want. But those who have read my air combat gaming guides will know that I am very good at spotting patterns. If COTA's AI has patterns, then its going to take a lot more smarts than me to spot them.
COTA is as close you are going to get to being able to find "war in a box" and get a sense of what it means to command. And if you are not a grog and only a guy or gal who likes computer games, then it will be an intellectual challenge worthy of the investment of your time.
And that's all I have to say about that.
I know you are thinking, "well what would a beta tester say about a game that they are closely involved with?".
Fair point.
You know COTA is not the only game I play. I will not enumerate the other games I play here and point out the flaws in the various AI implementations. But here is what I know:
Most reasonable quality games that I buy (which is the only type I buy, since I usually wait until the game has been out for quite a while), I am able to figure patterns in the AI behavior after no more than six months. Those patterns tend to repeat themselves from game to game such that a few standard formulas are sufficient to beat the AI provided that the scenario author hasn't buried some nasty surprises in the scenario or the force balance is not heavily weighted against the player.
I have been playing/testing the AA engine which powers COTA since 2001. So, I know the engine much better compared to these other games that I already own or purchase from time to time. Yet after five years with the AA engine I have yet to come up with a set of human-vs-AI formulas that guarantee one victory after another. Now, I do win these days the majority of scenarios which I play on the first time around. However, those victories come through careful analysis of the scenario, situation, and terrain --- no formula templates. The techniques which I employ are those very same ones that have shown up in my mini guide threads. For the most part, we are talking about high level plans composed out of building blocks of realistic military tactics. And in my case, there may also be an occassional ahistorical military practice here or there, since I am no grognard (just a game player).
But that is really something when you think about it. Five years and I have yet to arrive at a few basic rules for trouncing the AI every game I play. I have played video games since around 1992/1993 and done a number of gaming guides focused on MP or SP over the years. Those patterns of weakness in the AI found in other games, just don't seem to be there in COTA.
I am not saying the AI is perfect. I write scenario reviews when I play them and often do have criticisms, but those criticisms are more often than not about how the AI played a specific situation and not how it handles an entire class of situations.
Like I said, I am a beta tester and you can write this off as your usual "fan boy" ranting if you want. But those who have read my air combat gaming guides will know that I am very good at spotting patterns. If COTA's AI has patterns, then its going to take a lot more smarts than me to spot them.
COTA is as close you are going to get to being able to find "war in a box" and get a sense of what it means to command. And if you are not a grog and only a guy or gal who likes computer games, then it will be an intellectual challenge worthy of the investment of your time.
And that's all I have to say about that.
