Thanks for the feedback! Determining the right weights is tricky business, so it's good to hear when things are going wrong.
There are some limitations to the complexity of what we can model, since the optimization is done by solving a linear programming problem. However, it is easy to reduce the importance of CHA/sound/shards/nether/nexus/chaos, and increase the importance of permanent light. We do have the ability to use different weights based on variant, race, and/or class; I just haven't filled in all the combinations yet.
I'll have to think if it's possible to handle things like giving 0 importance to infravision if you have see invisible, or 0 importance to resistances once you have the corresponding immunities..
@camlost: Using the randart code to compute the goodness of a set of equipment is a good idea, but there are two issues with it. (1) it makes the optimization problem *very* hard--basically you just have to try every combination of equipment by brute force, and (2) it's hard to maintain/extend across variants (or even different versions of vanilla).
There are some limitations to the complexity of what we can model, since the optimization is done by solving a linear programming problem. However, it is easy to reduce the importance of CHA/sound/shards/nether/nexus/chaos, and increase the importance of permanent light. We do have the ability to use different weights based on variant, race, and/or class; I just haven't filled in all the combinations yet.
I'll have to think if it's possible to handle things like giving 0 importance to infravision if you have see invisible, or 0 importance to resistances once you have the corresponding immunities..
@camlost: Using the randart code to compute the goodness of a set of equipment is a good idea, but there are two issues with it. (1) it makes the optimization problem *very* hard--basically you just have to try every combination of equipment by brute force, and (2) it's hard to maintain/extend across variants (or even different versions of vanilla).
Comment