Repository navigation
Conversation
…ith the 2026-11-16 rows (new); --level sweep
…two-turn move's charge turn is no KO risk The port took D from the live planning_risk(), which counted Venusaur's Solar Beam (168-198 at Lv54, 171 HP) on every turn, so D >= max HP and the policy only attacked: the Champion was 0%. The program reads frlg-20.01's foe_threats rows; the port now does too (by party slot, by species when the party is cut down), poison as the only residual. Solar Beam without sun (and Dive) threatens nothing until the foe is charging; then the row applies. frlg_league_chain: --fights and --only-foe for a subset; each fight records the turns a foe crit landed on us.
…er-fight spend tables, policy-line replay frlg_league_chain --budget takes per-fight budgets in place of the planner split. frlg_league_spend prints won/reached and item spend (mean/p75/p90) per fight and proposes a split. frlg_league_policycheck rebuilds each 'Battle: policy:' state from the console log, asks the autostory_new port for its action, and compares it, the foe's damage and our damage with the engine's previews. Results: ~/BDSPLab/notes/e4-sim/real-bag/.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Depends on #109
Ports Arduino-Source #613's budgeted-setup League policy into the chain sim as autostory_old (as run on 2026-11-16 15:35) and autostory_new (sleep cure, Full Restore for a status at KO risk, bigger heal when a heal doesn't outpace the foe, escalate after two item turns, Strength finisher, Hyper Potion and Full Heal budgets). The port now reads the plan's foe_threats rows like the program does, and a two-turn move (Solar Beam without sun) is no KO risk until the foe is charging; before that the Champion was 0% because Venusaur's Solar Beam counted every turn. Adds --level, --fights and --only-foe, and records foe crits per turn.
Current bag (FR 7, X Sp 10, X Spe 5, X Acc 3, GS 2, Max Elixir 1, Full Heal 2), 200 chains each, still 0 full wins at Lv51/54/57. The Champion alone goes from 0/200 to 7-10/200: a landed Solar Beam does 168-198 to our 171 HP. Lance is lost mostly with Surf out of PP (one Max Elixir for the whole League). Tested with the new test_autostory suite (red before the fix); result files are in the lab notes (e4-sim/escalate/fix1).
Real bag read at 17:47:49 (FR 8, X Sp 15, Elixir + Max Elixir), Lv54, 500 chains: still 0 full wins. The program's split gets 34% of chains to the Champion; a p75-based split (FR 1/2/1/3/1, X Sp 4/2/3/3/3, X Spe 0/3/1/1/0, X Acc 0/0/0/2/1) gets 48%, mostly from Bruno and Agatha. Adds --budget, frlg_league_spend (per-fight spend tables) and frlg_league_policycheck, which replays the console's policy lines through the port: 35/47 turns agree. Results in e4-sim/real-bag.