// you win the fifth battle. Your score is the total cheese you gained.

  // This module tracks the state over the course of a game:

  // stats, changes as you train them, cheese accumulated, and so on.

  // These are made available in properties so that scripts can use them

  // without having to parse the response text for themselves. The properties

  // reset at the beginning of a game and are valid while a game is underway.

  //

  // This module records the results of games in other properties, which will

  // persist until rollover, in other scripts wish to analyze them.

  //

  // If the user opts in (via property), the results of all battles and all

  // cheese acquisitions are recorded in files in the "data" directory for

  // later analysis by ASH programs.

  // *** Ongoing research:

  //

  // Stats:

  //

  // - We know stat bonuses offered by style sets as displayed by

  //   "needles". Is it possible to determine actual stat values?

  //   Maybe: 12 cheese encounters scale (positively or negatively) according

  //   to one of the 6 stats. We collect data based on stat bonuses - and the

  //   yield is linear, with randomizing fuzz. Look at the x-intercepts?

  // - Is there a randomizing factor in player stats per game? One hopes not.

  // - How do the potions affect your stats during battles? They do not

  //   register on the "needles" as stat bonuses, and they do not affect

  //   cheese yields that scale by stat. But you can definitely win against

  //   tougher castles if you have them in effect than if you do not.

  //

  // Castles:

  //

  // - What are initial stats for the six castles?

  //   (Conjecture is that each is "better" at one of the six stats.)

  // - Is there a randomizing factor per game or battle?

  // - How do they scale as fight # increases?

  // - When comparing castle and player stats, one is always "higher" or

  //   "lower" than the other. Really? Perhaps equal stats have a 50% chance of

  //   winning or losing the toss?

  //

  // Battles:

  //

  // - It "feels" like you sometimes just can't win against what are normally

  //   easy opponents. Can we use statistics to confirm or deny the "feeling"?

  //   If there is a randomizing factor, which has fuzz? Player stats or castle

  //   stats? KoL "monsters" have such for attack/defense - not the player.

  //

  // Cheese:

  //

  // - What are the linear formulae for the 12 stat-scaling cheese encounters?

  // *** Solved research:

  //

  // Cheese:

  //

  // - The yields of the 3 non-scaling encounters (20, 50, 100)

  // - The yields from defeated castles (45 * castle level)

  // - The wishing well succeeds 1/3 of the time with a yield of 300.

  // - Potions that affect stats do not affect stat-scaling yields

  //

  // Stances:

  //

  // - offensive is 80% aggressor, 20% defender

  // - waiting is 50% aggressor, 50% defender

  // - defensive is 20% aggressor, 80% defender

  // We don't know what your (internal to KoL) stats start at,

  // Each of the four castle Upgrades provides bonuses to one or more stats.

  // (won or lost) of the day which affect your stats (for combats only).

  // The image of the rig has six indicators ("needles") at the bottom which

  // show your upgrade-granted boosts to your six stats. The potions do not

  // affect that display.

  // other, but I am sure the stats are not internally measured in pixels.)

  //     Generic - higher military defense?

  // Each upgrade provides a reward at the end of your first game of the day,

  // depending on the style you selected for the upgrade.

  // Each upgrade/style also provides a boost to specific attack/defense stats

  // It turns out that those values are not independent; the same upgrade swap

  // might grant +1 or +2 Castle Attack, say, depending on which other upgrades

  // are in place.

          // Should be impossible.

        // Ignore the style names; they are for humans to read

  // Write data file: for testing or for generating the first time.

  // *** Cached state. Resets when you visit the Bastille Battalion control rig

  // When you are in choice 1314 - Bastille Battalion (Master of None) - you

  // can focus on offense or defense, or choose to seek cheese.

  // with 3 different options out of a pool of 16 possibilities. You can take

  // each option only once per game. Since you have 2 rounds of preparation and

  // up to 5 castles per game, if you do nothing except look for cheese, your

  // first prep round will offer 3 out of 16, the second, 3 out of 15, until

  // the 10th, which will offer 3 out of 7 options.

  // will not have the 10 cheese required to activate it. If you skip it, like

  // all untaken options, it may be offered again later in the same game.

  // The Wishing Well is not affected by a stat, but either gives you no cheese

  // (2/3 chance) or about 300 cheese (1/3 chance).

  // One of the reasons I started this project was to collect data that could

  // be analyzed to understand how to do well at this game. The improved

  // logging makes the game play much more enjoyable, but the collected data

  // makes automated data analysis possible without manual data entry.

  // Observations so far:

  // Each kind of castle has particular strengths and weaknesses. There are six

  // kinds of castle. Each appears to be stronger in one of the six stats.

  // A "game" has 5 rounds. Your foes increase in power depending on which

  // round you encounter them.  For example, if I attack castle type A on round

  // one, my attack vs. his defense may be 3:0, but on rounds 2 - 5, attack

  // vs. defense may decrease to 2:1, 1:2, and eventually 0:3. Your cheese

  // reward for beating a foe go up correspondingly to the difficulty.

  // Depending on your stat configuration, you will have to depend on offense

  // or defense (i.e., which stance you select) to even have a chance against

  // higher difficulty castles. For a given stat configuration, some castles

  // will be unbeatable at higher levels - and to get the highest scores, you

  // have to be lucky enough to get one of the castles you CAN beat.

  //

  // Observations collected from 4308 battles as of 2022/06/20:

  //

  // "offensive" stance (901): 80% aggressor/20% defender

  // "waiting" stance (1586): 50% aggressor/50% defender

  // "defensive" stance (1821): 20% aggressor/80% defender

  //

  // With an "offensive" stance (80% of the time):

  // You charge toward your enemy.

  //

  // With a "defensive" stance (20% of the time):

  // You squat and wait for the attack, but it never comes. You sigh, uproot yourself, and attack

  // them.

  // Your Stance indicates your desire to attack vs. defend, but it's not

  // entirely up to you.  Even if you charge in, your foe may attack

  // first. Even if you try to defend, you may end up attacking first.

      case 1319: // Cheese Seeking Behavior

  // This is the format.

  //     string enemy;       // {frenchcastle,masterofnone,bigcastle,

  //                         // berserker,shieldmaster,barracks}