Top Mud Sites Forum - View Single Post

KaVir · 04-11-2006, 04:34 AM

That assumes there is an optimum decision. What about a system in which there are numerous viable decisions at each point, each with their own pros and cons? With the random element, no decision is perfect, and some may well be more risky than others. The script would be able to avoid the bad choices, but there would be no optimal good one.

Actually this rather reminds me of the . Building on Greggen's RPS example, imagine a very simple combat system based on the following five options:

1. Thrust (advantage against slash and feint)
2. Slash (advantage against pummel and feint)
3. Pummel (advantage against thrust, equal against feint)
4. Riposte (advantage against slash, thrust and pummel)
5. Feint (advantage against riposte, equal against pummel)

There's no way to know what move your opponent is performing. Each move takes 5 seconds to execute, although the defence part (for the purposes of advantage vs disadvantage) is in place the moment the command is entered. Your base chance to hit is 15%, +5% for every second you wait before entering the command, up to a maximum of 40% after five seconds. Your chance to hit is doubled if you have the advantage and halved if you have the disadvantage.

Which is the optimal choice? What is the optimal time to wait before launching your attack? How would you set up your script to fight in such a situation?

Given unlimited processing power you could script anything - but obviously nobody has unlimited processing power. I think it's possible to create a reasonable combat system that cannot be feasible scripted to be better than a human player. People who use a mixture of scripting and regular play will most likely still have an advantage, but you can certainly force the scripting to take a secondary role.

The point I was making is that not everything can be feasibly handled through scripts.

Interestingly enough my 'war' combat system is based on poker, and does often result in battles between large numbers of players - and, indeed, I've yet to see anyone script it. Admittedly it's turn-based (3 actions every 60 second turn), but even so, people will sometimes miss actions - so obviously thinking speed is still a factor.

That's assuming the only way to fight is to inflict pure damage, though. What if you want to stun one of your opponents so that he can't heal his friends? What if you've got an attack which allows you to hurl one of your opponent's at another, allowing you to injure two people at once? What if you're fighting on the move, and only have to inflict enough damage to knock your opponents off their horses? What if you're able to inflict damage-over-time (poison, bleeding wounds, etc), and can therefore do a larger amount of overall damage by splitting your blows between your opponents (once one opponent is poisoned you move on to the next)?

Then there's also the question of defence. Maybe you want to focus on one opponent, but you're still going to have multiple people focusing on you. Will you go pure offense in the hope of evening the odds as quickly as possible (or at least taking someone down with you)? Go pure defence in the hope that an ally can reach you before you die? Assume some sort of balance between the two, so that you can survive long enough to inflict reasonable damage?

04-11-2006, 04:34 AM	#85
KaVir Legend Join Date: Apr 2002 Name: Richard Home MUD: God Wars II Posts: 2,052	That assumes there is an optimum decision. What about a system in which there are numerous viable decisions at each point, each with their own pros and cons? With the random element, no decision is perfect, and some may well be more risky than others. The script would be able to avoid the bad choices, but there would be no optimal good one. Actually this rather reminds me of the . Building on Greggen's RPS example, imagine a very simple combat system based on the following five options: 1. Thrust (advantage against slash and feint) 2. Slash (advantage against pummel and feint) 3. Pummel (advantage against thrust, equal against feint) 4. Riposte (advantage against slash, thrust and pummel) 5. Feint (advantage against riposte, equal against pummel) There's no way to know what move your opponent is performing. Each move takes 5 seconds to execute, although the defence part (for the purposes of advantage vs disadvantage) is in place the moment the command is entered. Your base chance to hit is 15%, +5% for every second you wait before entering the command, up to a maximum of 40% after five seconds. Your chance to hit is doubled if you have the advantage and halved if you have the disadvantage. Which is the optimal choice? What is the optimal time to wait before launching your attack? How would you set up your script to fight in such a situation? Given unlimited processing power you could script anything - but obviously nobody has unlimited processing power. I think it's possible to create a reasonable combat system that cannot be feasible scripted to be better than a human player. People who use a mixture of scripting and regular play will most likely still have an advantage, but you can certainly force the scripting to take a secondary role. The point I was making is that not everything can be feasibly handled through scripts. Interestingly enough my 'war' combat system is based on poker, and does often result in battles between large numbers of players - and, indeed, I've yet to see anyone script it. Admittedly it's turn-based (3 actions every 60 second turn), but even so, people will sometimes miss actions - so obviously thinking speed is still a factor. That's assuming the only way to fight is to inflict pure damage, though. What if you want to stun one of your opponents so that he can't heal his friends? What if you've got an attack which allows you to hurl one of your opponent's at another, allowing you to injure two people at once? What if you're fighting on the move, and only have to inflict enough damage to knock your opponents off their horses? What if you're able to inflict damage-over-time (poison, bleeding wounds, etc), and can therefore do a larger amount of overall damage by splitting your blows between your opponents (once one opponent is poisoned you move on to the next)? Then there's also the question of defence. Maybe you want to focus on one opponent, but you're still going to have multiple people focusing on you. Will you go pure offense in the hope of evening the odds as quickly as possible (or at least taking someone down with you)? Go pure defence in the hope that an ally can reach you before you die? Assume some sort of balance between the two, so that you can survive long enough to inflict reasonable damage?