AI recruitment_save_gold Aspect Performance

This page lists the scenarios that have been tested against the new recruitment_save_gold AI aspect, enabled by default by Wesnoth 1.11.7. Included are the results for each scenario and, where applicable, the action taken to correct any issues that cropped up during testing.

Only the items in bold have been tested.

E1

• E1S1
• E1S2
• E1S3
• E1S4 - Changed gold and AI recruitment patterns
  • Commit a345e44feeb9fb955b87446148fb7288f201373a
  • Commit 53819614970c8b949e751ea61c7192e7cd540c03
• E1S5 - No changes needed, difficulty unchanged
• E1S6.1, E1S6.2 - N/A
• E1S7 - Made it so recruitment_save_gold is disabled the undead guards on their first turn
  • Commit 700b88bea6956e959c470a94bdc8daa153041675
• E1S7x - N/A
• E1S8 - No changes needed, difficulty slightly increased
• E1S9.1, E1S9.2, E1S9.3 - N/A
• E1S10 - No changes needed, difficulty unchanged
• E1S11.1 - No changes needed, difficulty probably slightly increased, AI seems more competent
• E1S11.2 - No changes needed, difficulty unchanged
• E1S12 - No changes needed, difficulty unchanged
• E1S13 - N/A

E2

• E2S0 - N/A
• E2S1 - Disabled aspect for side 2 as it makes it dumber, kept for side 3 as it makes it more competent
  • Commit 7d97a4f23a7722d3994ad06d8cb2cf6a292327c6
• E2S2 - Disabled aspect for the bandits as they seem overcompetent with it
  • Commit 954b781761baf1cd8e5a00a36fb99d6e5b048fc4
• E2S3.1 - N/A
• E2S3.2 - N/A
• E2S4 - Kept aspect for all AI players as it seems to make them play more according to the intended design
• E2S5 - Difficulty appears to have decreased with the aspect enabled, decision pending
• E2S5x - N/A
• E2S6 - No changes needed, difficulty appears to have marginally increased
• E2S7 - No changes needed, difficulty unchanged
• E2S8 - Difficulty seems inconsistent between two runs (increased/decreased) with a similar set of units for the human player, disabled aspect for some AI sides
• E2S9 - No changes needed, difficulty unchanged
• E2S10 - No changes needed, difficulty changed over time during a single playthrough, scenario is slated for a major reorganization anyway
• E2S11 - No changes needed, difficulty unchanged
• E2S12 - N/A

E3

• E3S0 - N/A
• E3S1 - The scenario is very difficult by design, not sure if the difficulty actually changed
• E3S2 - The scenario is very difficult by design, not sure if the difficulty actually changed; increased initial gold for side 2 in any case, decision pending
• E3S3 - Difficulty seems unchanged.
• E3S4.1 - The scenario is very difficult by design, seems to have become marginally easier
• E3S4.2 - N/A
• E3S5 - An AI side behaved as though grouping=defense was set for it (it wasn't!); scenario seems generally easier (a good thing in this case)
• E3S6 - N/A
• E3S7A.1F1 - Difficulty seems unchanged
• E3S7A.1F2 - Recruiting AIs are completely dysfunctional, scenario became easier because AIs won't recruit unless the user reloads a save (???). Otherwise they seem to start recruiting only when directly confronted by the player (i.e. the player reveals the enemy leader through shroud) (???).
• E3S7A.2 - N/A
• E3S7B - SE and E yellow sides are completely dysfunctional and pretty much sit idle doing nothing, reloading doesn't help. Difficulty goes down significantly after turn 10, but this might be for the best. Considering reducing difficulty in a more deterministic fashion without using the recruitment_save_gold aspect.
• E3S8A - N/A
• E3S8B - Difficulty seems unchanged
• E3S8C - The recruiting AI became incompetent, disabled aspect
• E3S8D - Difficulty seems unchanged, and the AI doesn't seem able to save much gold anyway
• E3S9 - N/A

•  - N/A
  

• E3S10 - Difficulty seems unchanged
• E3S11 - N/A
• E3S12 - N/A
• E3S13 - N/A