Xeon said:
First, get the crappy actions out of the way. These are usually Station skills and have mod Progress, low Quality. Only use the Progress actions for these steps.
That sounds good, depending on how it affects a low level crafter. Could you have the bot check the crafting level of the Crafter, the gear availible, and decide from there? I do not know how easy/possible that is. Just a question for your consideration.
Xeon said:
Next, pick the "best" Quality action and try to get Quality up to the Min Quality slider setting.
* What do people think are the "best" quality actions?
* Or should I just add a checkbox letting people pick Utilities vs Tools to weight Progress actions?
Checkbox, because at low levels tools are not the way to go, where in high levels they are.
Xeon said:
After Quality is above Min Slider value, now choose Actions with Progress+Quality while trying to stay under the Prog Ratio slider value
* Prog Ratio is: totalStepsLeft / APLeft * 10000
(The * 10000 is just to get an easy int number to work with)
* This is slider is supposed to Make sure you always have enough AP left to finish the recipe. I have found it very hard to calculate the actual number of AP required to finish a stage/step because of the sometimes HUGE differences in progress vs AP cost + Complications, etc.
This sounds like a great idea... "min quality achieved, work up progress, if AP>Cost2Complete by 300 pump up quality, else progress and finish. Instead of a slider, how about an input window for minimum AP left before progressing to finish (ignoring complications).
Xeon said:
If Quality gets above Max Slider value, then lower the priority of Quality. Drop priority of Quality to zero once we hit 1000 Quality.
Again, I'm thinking an input box for the level of quality the person wants.
Xeon said:
If Prog Ratio is exceeded, just do "best" Progress steps.
* Right now best steps are computed as the words (High, Moderate, Low) vs the AP Cost of that action. I use values of Very High = 16, High = 8, Moderate = 4, Low = 2 and Very Low = 1 to weight the selection in favor of the highest Progress actions.
---
Does that sound like a good Strategy?
Right now the AI is saving the Utility Quality step for last and trying to pump out as much progress as possible before using it. I like that because it means I can get a more accurate idea of how many AP are left before trying to push Quality higher. But the drawback is the it often chooses bad actions because the Quality is below the Min Quality Slider level.
Ideas, comments and suggestions?
As you can see, I favor check boxes and input boxes. I'm not so hot on sliders, they are difficult for many to understand and use properly. A properly named entry box, with well written instructions would aleviate many issues. Or so I believe.
One other thing, is it possible to make it look at the options for fixing a complication, and see if it has the tool or supply item to fix it, and then use the highest value fix to start, and then determine if the lowest, or medium (if there are that many choices) will finish out fixing it without costing too many AP. For example, 3 choices 1 is 100AP High progress, 2 is 75 AP Medium progress, and 3 is 50 AP Low progress. First use the 100AP, then based on the amount of the complication still left it would choose the Low progress because it is very close to being fixed, or the medium one because it's at about 75%.
Make it look at the required items for the item(s) to be crafted are in sufficient quantity and refuse to do the work if it cannot get a full 25 stack out of what is availible. This also means that it should, between each round of crafting, take the short stacks of the same kind and combining them to ensure the maximum number of full stacks is always availible.
Those are my ideas, and my takes on this issue. I hope they are helpful.