Mercurial > pokemon-types
view org/types.org @ 8:4f9ef752e2f0
more cleanup and better docstrings for clojure functions
author | Robert McIntyre <rlm@mit.edu> |
---|---|
date | Wed, 02 Nov 2011 07:01:49 -0700 |
parents | d6b8dab05d9d |
children | fd38763de457 |
line wrap: on
line source
1 #+TITLE: Best-First Search for Effective Pokemon Types2 #+AUTHOR: Robert McIntyre & Dylan Holmes3 #+EMAIL: rlm@mit.edu4 #+description: Finding interesting pokemon type combinations through Best-First search in clojure.5 #+SETUPFILE: ../../aurellem/org/setup.org6 #+INCLUDE: ../../aurellem/org/level-0.org8 * The Pok\eacute{}mon Type System10 The Pok\eacute{}mon type system consists of seventeen different11 \ldquo{}types\rdquo{} (Rock, Grass, Ice, Psychic, Ground, Bug, Flying,12 Fire, Fighting, Dark, Dragon, Poison, Water, Ghost, Normal, Electric,13 and Steel) that interact like an extended version of14 Rock-Paper-Scissors: for example, the Fire type is strong against the15 Grass type but weak against the Water type. In the table below, we've16 recorded the relative strengths of each of the types in the17 Pok\eacute{}mon type system; the number in each cell indicates how18 effective an attack of the type in the row is against a19 Pok\eacute{}mon of the type in the column. We call these numbers20 /susceptibilities/.22 In the Pok\eacute{}mon games, only four susceptibility values (two,23 one, one-half, and zero) occur. These numbers indicate particularly24 high susceptibility, average susceptibility, particularly low25 susceptibility, and no susceptibility (immunity).27 - The suceptability of Flying types /against/ Ground is 0, because Ground28 attacks cannot hurt Flying pok\eacute{}mon at all. The damage that29 a Ground type attack normally does is /multiplied/ by 0 when it is30 uesd against a Flying type pok\eacute{}mon.32 - The susceptability of Fire types against Water attacks33 is 2, because Water type attacks are strong against Fire type34 Pok\eacute{}mon. The damage that a Water type attack normally does35 is doubled when it is used against a Fire type pok\eacute{}mon.37 - The susceptability of Water types against Water attacks is38 $\frac{1}{2}$, because Water type attacks are strong against Water39 type Pok\eacute{}mon. The damage that a Water type attack normally40 does is halved when it is used against a Water type41 pok\eacute{}mon.43 There are two pok\eacute{}mon type systems in use. The first is the44 classic system which was used for the very first pok\eacute{}mon45 games, Red, Yellow, and Blue. This old system was used from 1998 to46 2000 in America, and is known as the /Generation I Type System/. The47 modern pok\eacute{}mon type system was introduced in 2000 with the48 introduction of pok\eacute{}mon Gold and Silver, and has been in use49 ever since. It is called the /Generation II Type System/.51 Here are the the definitions of the two type systems.53 * Generation I and II Type System Data55 ** Generation II Type System56 #+label: pokemon-matchups57 #+tblname: pokemon-table-gen-two58 | | normal | fire | water | electric | grass | ice | fighting | poison | ground | flying | psychic | bug | rock | ghost | dragon | dark | steel |59 |----------+--------+------+-------+----------+-------+-----+----------+--------+--------+--------+---------+-----+------+-------+--------+------+-------|60 | normal | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 0 | 1 | 1 | .5 |61 | fire | 1 | .5 | .5 | 1 | 2 | 2 | 1 | 1 | 1 | 1 | 1 | 2 | .5 | 1 | .5 | 1 | 2 |62 | water | 1 | 2 | .5 | 1 | .5 | 1 | 1 | 1 | 2 | 1 | 1 | 1 | 2 | 1 | .5 | 1 | 1 |63 | electric | 1 | 1 | 2 | .5 | .5 | 1 | 1 | 1 | 0 | 2 | 1 | 1 | 1 | 1 | .5 | 1 | 1 |64 | grass | 1 | .5 | 2 | 1 | .5 | 1 | 1 | .5 | 2 | .5 | 1 | .5 | 2 | 1 | .5 | 1 | .5 |65 | ice | 1 | .5 | .5 | 1 | 2 | .5 | 1 | 1 | 2 | 2 | 1 | 1 | 1 | 1 | 2 | 1 | .5 |66 | fighting | 2 | 1 | 1 | 1 | 1 | 2 | 1 | .5 | 1 | .5 | .5 | .5 | 2 | 0 | 1 | 2 | 2 |67 | poison | 1 | 1 | 1 | 1 | 2 | 1 | 1 | .5 | .5 | 1 | 1 | 1 | .5 | .5 | 1 | 1 | 0 |68 | ground | 1 | 2 | 1 | 2 | .5 | 1 | 1 | 2 | 1 | 0 | 1 | .5 | 2 | 1 | 1 | 1 | 2 |69 | flying | 1 | 1 | 1 | .5 | 2 | 1 | 2 | 1 | 1 | 1 | 1 | 2 | .5 | 1 | 1 | 1 | .5 |70 | psychic | 1 | 1 | 1 | 1 | 1 | 1 | 2 | 2 | 1 | 1 | .5 | 1 | 1 | 1 | 1 | 0 | .5 |71 | bug | 1 | .5 | 1 | 1 | 2 | 1 | .5 | .5 | 1 | .5 | 2 | 1 | 1 | .5 | 1 | 2 | .5 |72 | rock | 1 | 2 | 1 | 1 | 1 | 2 | .5 | 1 | .5 | 2 | 1 | 2 | 1 | 1 | 1 | 1 | .5 |73 | ghost | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 2 | 1 | 1 | 2 | 1 | .5 | .5 |74 | dragon | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 2 | 1 | .5 |75 | dark | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 1 | 1 | 1 | 2 | 1 | 1 | 2 | 1 | .5 | .5 |76 | steel | 1 | .5 | .5 | .5 | 1 | 2 | 1 | 1 | 1 | 1 | 1 | 1 | 2 | 1 | 1 | 1 | .5 |78 The rows are attack types, while the columns are defense types. To79 see the multiplier for a pok\eacute{}mon attack against a certain type, follow80 the row for the attack type to the column of the defending type.82 ** Generation I Type System83 #+label: pokemon-matchups-gen-184 #+tblname: pokemon-table-gen-one85 | | normal | fire | water | electric | grass | ice | fighting | poison | ground | flying | psychic | bug | rock | ghost | dragon |86 |----------+--------+------+-------+----------+-------+-----+----------+--------+--------+--------+---------+-----+------+-------+--------|87 | normal | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 0 | 1 |88 | fire | 1 | .5 | .5 | 1 | 2 | 2 | 1 | 1 | 1 | 1 | 1 | 2 | .5 | 1 | .5 |89 | water | 1 | 2 | .5 | 1 | .5 | 1 | 1 | 1 | 2 | 1 | 1 | 1 | 2 | 1 | .5 |90 | electric | 1 | 1 | 2 | .5 | .5 | 1 | 1 | 1 | 0 | 2 | 1 | 1 | 1 | 1 | .5 |91 | grass | 1 | .5 | 2 | 1 | .5 | 1 | 1 | .5 | 2 | .5 | 1 | .5 | 2 | 1 | .5 |92 | ice | 1 | 1 | .5 | 1 | 2 | .5 | 1 | 1 | 2 | 2 | 1 | 1 | 1 | 1 | 2 |93 | fighting | 2 | 1 | 1 | 1 | 1 | 2 | 1 | .5 | 1 | .5 | .5 | .5 | 2 | 0 | 1 |94 | poison | 1 | 1 | 1 | 1 | 2 | 1 | 1 | .5 | .5 | 1 | 1 | 2 | .5 | .5 | 1 |95 | ground | 1 | 2 | 1 | 2 | .5 | 1 | 1 | 2 | 1 | 0 | 1 | .5 | 2 | 1 | 1 |96 | flying | 1 | 1 | 1 | .5 | 2 | 1 | 2 | 1 | 1 | 1 | 1 | 2 | .5 | 1 | 1 |97 | psychic | 1 | 1 | 1 | 1 | 1 | 1 | 2 | 2 | 1 | 1 | .5 | 1 | 1 | 1 | 1 |98 | bug | 1 | .5 | 1 | 1 | 2 | 1 | .5 | 2 | 1 | .5 | 2 | 1 | 1 | 0 | 1 |99 | rock | 1 | 2 | 1 | 1 | 1 | 2 | .5 | 1 | .5 | 2 | 1 | 2 | 1 | 1 | 1 |100 | ghost | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 0 | 1 | 1 | 2 | 1 |101 | dragon | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 2 |104 This is the old table from Generation I. The differences from105 Generation II are:106 - Dark and Steel types are missing (these were introduced in107 Generation II).108 - Bug is super-effective against Poison (not-very-effective in109 Generation II).110 - Poison is super-effective against Bug (normal in Generation II).111 - Bug is regularly effective against Ghost (super-effective in112 Generation II).113 - Ice is normally effective against Fire, (not-very-effective in114 Generation II).115 - Ghost is completely ineffective against Psychic. This is considered116 to be a programning glitch.120 * Representing the Data122 After creating the Pok\eacute{}mon types namespace, we store the123 tables of susceptibilities above in =pokemon-table-gen-one= and124 =pokemon-table-gen-two=, each of which is a simple vector of125 vectors. Because a vector of vectors can be cumbersome, we do not126 access the tables directly; instead, we use the derivative structures127 =attack-strengths= and =defense-strengths=, which are functions which128 return hash-maps associating each row (respectively column) of the129 table with its corresponding Pok\eacute{}mon type.132 #+srcname: header133 #+begin_src clojure :results silent134 (ns pokemon.types135 (:use clojure.set)136 (:use clojure.contrib.combinatorics)137 (:use clojure.contrib.math)138 (:use clojure.contrib.def)139 (:use rlm.rlm-commands))140 #+end_src142 #+srcname: data143 #+begin_src clojure :results silent144 (in-ns 'pokemon.types)145 ;; record type strengths as a vector of vectors146 ;; the variables pokemon-table-gen-one and pokemon-table-gen-two147 ;; are replaced with the tables above when this file is tangled.148 (def pokemon-gen-one pokemon-table-gen-one)149 (def pokemon-gen-two pokemon-table-gen-two)151 (defn type-names [] (vec (doall (map (comp keyword first) pokemon-gen-two))))153 (defn attack-strengths []154 (zipmap155 (type-names)156 (map (comp vec rest) pokemon-gen-two)))158 (defn defense-strengths []159 (zipmap (type-names)160 (map161 (apply juxt (map (attack-strengths) (type-names)))162 (range (count (type-names))))))163 #+end_src165 The two statements167 #+begin_src clojure :exports code168 (def pokemon-gen-one pokemon-table-gen-one)169 (def pokemon-gen-two pokemon-table-gen-two)170 #+end_src172 probably look weird. When the actual source file is created, those173 variables are replaced with the data from the tables above.175 See [[../src/pokemon/types.clj][types.clj]] to look at the final tangled output.178 #+begin_src clojure :results output :exports both179 (clojure.pprint/pprint pokemon.types/pokemon-gen-two)180 #+end_src182 #+results:183 #+begin_example184 (("normal" 1 1 1 1 1 1 1 1 1 1 1 1 0.5 0 1 1 0.5)185 ("fire" 1 0.5 0.5 1 2 2 1 1 1 1 1 2 0.5 1 0.5 1 2)186 ("water" 1 2 0.5 1 0.5 1 1 1 2 1 1 1 2 1 0.5 1 1)187 ("electric" 1 1 2 0.5 0.5 1 1 1 0 2 1 1 1 1 0.5 1 1)188 ("grass" 1 0.5 2 1 0.5 1 1 0.5 2 0.5 1 0.5 2 1 0.5 1 0.5)189 ("ice" 1 0.5 0.5 1 2 0.5 1 1 2 2 1 1 1 1 2 1 0.5)190 ("fighting" 2 1 1 1 1 2 1 0.5 1 0.5 0.5 0.5 2 0 1 2 2)191 ("poison" 1 1 1 1 2 1 1 0.5 0.5 1 1 1 0.5 0.5 1 1 0)192 ("ground" 1 2 1 2 0.5 1 1 2 1 0 1 0.5 2 1 1 1 2)193 ("flying" 1 1 1 0.5 2 1 2 1 1 1 1 2 0.5 1 1 1 0.5)194 ("psychic" 1 1 1 1 1 1 2 2 1 1 0.5 1 1 1 1 0 0.5)195 ("bug" 1 0.5 1 1 2 1 0.5 0.5 1 0.5 2 1 1 0.5 1 2 0.5)196 ("rock" 1 2 1 1 1 2 0.5 1 0.5 2 1 2 1 1 1 1 0.5)197 ("ghost" 0 1 1 1 1 1 1 1 1 1 2 1 1 2 1 0.5 0.5)198 ("dragon" 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 0.5)199 ("dark" 1 1 1 1 1 1 0.5 1 1 1 2 1 1 2 1 0.5 0.5)200 ("steel" 1 0.5 0.5 0.5 1 2 1 1 1 1 1 1 2 1 1 1 0.5))201 #+end_example203 =pokemon-gen-two= is a simple list-of-lists data structure.205 #+begin_src clojure :results output :exports both206 (clojure.pprint/pprint (pokemon.types/defense-strengths))207 #+end_src209 #+results:210 #+begin_example211 {:water [1 0.5 0.5 2 2 0.5 1 1 1 1 1 1 1 1 1 1 0.5],212 :psychic [1 1 1 1 1 1 0.5 1 1 1 0.5 2 1 2 1 2 1],213 :dragon [1 0.5 0.5 0.5 0.5 2 1 1 1 1 1 1 1 1 2 1 1],214 :fire [1 0.5 2 1 0.5 0.5 1 1 2 1 1 0.5 2 1 1 1 0.5],215 :ice [1 2 1 1 1 0.5 2 1 1 1 1 1 2 1 1 1 2],216 :grass [1 2 0.5 0.5 0.5 2 1 2 0.5 2 1 2 1 1 1 1 1],217 :ghost [0 1 1 1 1 1 0 0.5 1 1 1 0.5 1 2 1 2 1],218 :poison [1 1 1 1 0.5 1 0.5 0.5 2 1 2 0.5 1 1 1 1 1],219 :flying [1 1 1 2 0.5 2 0.5 1 0 1 1 0.5 2 1 1 1 1],220 :normal [1 1 1 1 1 1 2 1 1 1 1 1 1 0 1 1 1],221 :rock [0.5 0.5 2 1 2 1 2 0.5 2 0.5 1 1 1 1 1 1 2],222 :electric [1 1 1 0.5 1 1 1 1 2 0.5 1 1 1 1 1 1 0.5],223 :ground [1 1 2 0 2 2 1 0.5 1 1 1 1 0.5 1 1 1 1],224 :fighting [1 1 1 1 1 1 1 1 1 2 2 0.5 0.5 1 1 0.5 1],225 :dark [1 1 1 1 1 1 2 1 1 1 0 2 1 0.5 1 0.5 1],226 :steel [0.5 2 1 1 0.5 0.5 2 0 2 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5],227 :bug [1 2 1 1 0.5 1 0.5 1 0.5 2 1 1 2 1 1 1 1]}228 #+end_example230 =defense-strengths= is a more convenient form of =pokemon-gen-two=,231 with key/value pair access.233 * Interfacing with the Data235 In the pok\eacute{}mon games, a pok\eacute{}mon can have up to two236 types at the same time. For example, Zapdos, the fearsome legendary237 that can control lightning, has both the Electric and Flying types. A238 pok\eacute{}mon with more than one type gains the advantages and239 disadvanteags of both types. The suceptibilitys of each type are240 multiplied together to produce the hybrid type's susceptibilities. For241 example, Electric is weak to Ground (susceptibility of 2), but Flying242 is immune to Ground (suceptibility of 0). Zapdos' type,243 Electrig/Flying, is immune to Ground because $2 \times 0 = 0$.245 #+srcname: types246 #+begin_src clojure :results silent247 (in-ns 'pokemon.types)249 (defn multitypes "All combinations of up to n types" [n]250 (vec251 (map vec252 (reduce concat253 (map (partial combinations (type-names))254 (range 1 (inc n)))))))256 (defn susceptibility257 "Hash-map of the susceptibilities of the given type combination258 to each type of attack"259 [pkmn-types]260 (rlm.map-utils/map-vals261 clojure.core/rationalize262 (apply hash-map263 (interleave (type-names)264 (apply (partial map *)265 (map (defense-strengths) pkmn-types))))))267 (defn susceptance268 "The cumulative susceptibility of the given type combination"269 [types]270 (reduce + (map #(expt % 2) (vals (susceptibility types)))))271 #+end_src273 Now we can work out the suceptability of Zapdos automatically.275 Electric is weak to Ground.276 #+begin_src clojure :exports both277 (:ground (pokemon.types/susceptibility [:electric]))278 #+end_src280 #+results:281 : 2283 Flying is immune to Ground.284 #+begin_src clojure :exports both285 (:ground (pokemon.types/susceptibility [:flying]))286 #+end_src288 #+results:289 : 0291 Together, they are immune to Ground.292 #+begin_src clojure :exports both293 (:ground (pokemon.types/susceptibility [:electric :flying]))294 #+end_src296 #+results:297 : 0302 * Best-First Search304 I'd like to find type combinations that are interesting, but the total305 number of combinations gets huge as we begin to consider more306 types. For example, the total possible number of type combinations307 given just 8 possible types is: 17^{8} = 6,975,757,441 combinations.308 Therefore, it's prudent to use search.310 These functions are a simple implementation of best-first search in311 clojure. The idea to start off with a collection of nodes and some way312 of finding the best node, and to always expand the best node at every313 step.315 #+srcname: search316 #+begin_src clojure :results silent317 (in-ns 'pokemon.types)319 (defn comparatize320 "Define a comparator which uses the numerical outputs of fn as its criterion.321 Objects are sorted in increasing numerical order. Objects with the same fn-value322 are further compared by clojure.core/compare."323 [fun]324 (fn [a b]325 (let [val-a (fun a)326 val-b (fun b)]327 (cond328 ;; if the function cannot differentiate the two values329 ;; then compare the two values using clojure.core/compare330 (= val-a val-b) (compare a b)331 true332 ;; LOWER values of the function are preferred333 (compare (- val-a val-b) 0)))))335 (defn-memo best-first-step [successors [visited unvisited]]336 (cond (empty? unvisited) nil337 true338 (let [best-node (first unvisited)339 visited* (conj visited best-node)340 unvisited*341 (difference342 (union unvisited (set (successors best-node)))343 visited*)]344 (println best-node)345 [visited* unvisited*])))347 ;; memoize partial from core so that for example348 ;; (= (partial + 1) (partial + 1))349 ;; this way, best first search can take advantage of the memoization350 ;; of best-first step351 (undef partial)352 (def partial (memoize clojure.core/partial))354 (defn best-first-search355 "Searches through a network of alternatives, pursuing356 initially-promising positions first. Comparator defines which357 positions are more promising, successors produces a list of improved358 positions from the given position (if any exist), and initial-nodes is359 a list of starting positions. Returns a lazy sequence of search results360 [visited-nodes unvisited-nodes], which terminates when361 there are no remaining unvisited positions."362 [comparator successors initial-nodes]363 (let [initial-nodes364 (apply (partial sorted-set-by comparator) initial-nodes)365 initial-visited-nodes (sorted-set-by comparator)366 step (partial best-first-step successors)]367 (take-while368 (comp not nil?)369 (iterate step [initial-visited-nodes initial-nodes]))))371 #+end_src374 Now that we have a basic best-first-search, it's convenient to write a375 few pok\eacute{}mon-type specific convenience functions.377 #+srcname: pokemon-search378 #+begin_src clojure :results silent379 (in-ns 'pokemon.types)380 (defvar type-compare (comparatize susceptance)381 "compare two type combinations wrt their susceptibilities")383 (defn type-successors384 "Return the set of types that can be made by appending a single type385 to the given combination."386 [type]387 (if (nil? type) '()388 (set (map (comp vec sort (partial into type)) (multitypes 1)))))390 (defn immortal?391 "A type combo is immortal if it is resistant or invulnerable to392 every pokemon type. This is because that set of types can just be393 repeated to achieve as low a susceptance as desired"394 [type]395 (every? (partial > 1) (vals (susceptibility type))))397 (defn type-successors*398 "Stop expanding a type if it's immortal, or if it is longer than or399 equal to limit-size. Also, only return type additions that are400 strictly better than the initial type."401 [limit-size type]402 (if (or (<= limit-size (count type)) (immortal? type)) '()403 (set (filter #(< 0 (type-compare type %)) (type-successors type)))))405 (defn pokemon-type-search406 "Search among type-combos no greater than length n, limited by limit407 steps of best-first-search."408 ([n] (pokemon-type-search n Integer/MAX_VALUE))409 ([n limit]410 (first (last411 (take412 limit413 (best-first-search414 type-compare415 (partial type-successors* n)416 (multitypes 1)))))))418 (defvar immortals419 (comp (partial filter immortal?) pokemon-type-search)420 "find all the immortal pokemon types ")422 #+end_src424 Because there are so many type combinations, it's important to narrow425 down the results as much as possible. That is why =type-successors*=426 only returns types that are actually better than the type it is given.428 Best-first search can get caught optimizing a single type forever, so429 it's also important to limit the search space to be finite by setting430 an upper bound on the length of a type combo.432 * Results433 ** The best dual-type combo435 #+begin_src clojure :results cache verbatim :exports both436 (first (pokemon.types/pokemon-type-search 2))437 #+end_src439 #+results:440 : [:dark :ghost]442 Dark and Ghost, which additionally has the property of having no443 weaknesses to any other type, is the best type combo in terms of444 susceptance.446 The Dark and Steel types were introduced many years after447 pok\eacute{}mon started. In addition to the additional types, the448 pok\eacute{}mon games gained a few new rules concerning some of the449 matchups of the original types. Therefore, it's also interesting to see what450 type combination was most powerful before those types and new rules were introduced.452 The easiest way to do this with my setup is to just rebind the453 =pokemon-gen-two= table to the =pokemon-gen-one= table. Since454 everything that references this variable is a function and we're not455 doing anything too crazy with lazy-sequences and late-binding, this456 simple macro will do the job.458 #+srcname: old-school459 #+begin_src clojure :results silent460 (in-ns 'pokemon.types)462 (defmacro old-school463 [& forms]464 `(binding [pokemon-gen-two pokemon-gen-one] ~@forms))465 #+end_src467 Using the =old-school= macro, it's easy to find answers for the468 original 15 pokemon types as well as the expanded pokemon types469 introduced later.471 #+begin_src clojure :results verbatim :exports both :cache yes472 (pokemon.types/old-school (first (pokemon.types/pokemon-type-search 2)))473 #+end_src475 #+results[f43470fdf460ed546e9c57879abc9eda56da129f]:476 : [:ghost :psychic]478 Ghost and Psychic also manages to have no weaknesses to any of the original479 types, using the old Generation I rules.481 #+begin_src clojure :results output :exports both482 (clojure.pprint/pprint483 (pokemon.types/old-school484 (pokemon.types/susceptibility [:ghost :psychic])))485 #+end_src487 #+results:488 #+begin_example489 {:water 1,490 :psychic 1/2,491 :dragon 1,492 :fire 1,493 :ice 1,494 :grass 1,495 :ghost 0,496 :poison 1/2,497 :flying 1,498 :normal 0,499 :rock 1,500 :electric 1,501 :ground 1,502 :fighting 0,503 :bug 0}504 #+end_example506 ** An Immortal Type507 It's possible to quickly find an immortal type by giving the search508 a long enough maximum type length. 50 rounds of search with a max509 type limit of 10 is enough to find an immortal type.511 #+begin_src clojure :results scalar :exports both512 (first (pokemon.types/pokemon-type-search 10 50))513 #+end_src515 #+results:516 : [:dragon :fire :flying :ghost :grass :ground :steel :steel :water :water]519 #+begin_src clojure :results output :exports both520 (clojure.pprint/pprint521 (pokemon.types/susceptibility522 [:dragon :fire :flying :ghost :grass :ground :steel :steel :water :water]))523 #+end_src525 #+results:526 #+begin_example527 {:water 1/4,528 :psychic 1/4,529 :dragon 1/2,530 :fire 1/2,531 :ice 1/2,532 :grass 1/8,533 :ghost 1/2,534 :poison 0,535 :flying 1/2,536 :normal 0,537 :rock 1/2,538 :electric 0,539 :ground 0,540 :fighting 0,541 :dark 1/2,542 :steel 1/32,543 :bug 1/16}544 #+end_example546 ** Explanations for Common Pok\eacute{}mon Strategies548 Many people start out a battle with either a Normal pok\eacute{}mon or an549 Electric pok\eacute{}mon. Here's some justification for that choice.551 #+srcname: weaknesses552 #+begin_src clojure :results silent553 (in-ns 'pokemon.types)554 (defn critical-weaknesses [type]555 (count (filter #(> % 1) (vals (susceptibility type)))))556 #+end_src558 #+begin_src clojure :exports both :results output559 (clojure.pprint/pprint560 (sort-by pokemon.types/critical-weaknesses (pokemon.types/multitypes 1)))561 #+end_src563 #+results:564 #+begin_example565 ([:normal]566 [:electric]567 [:water]568 [:fighting]569 [:poison]570 [:ghost]571 [:dragon]572 [:dark]573 [:fire]574 [:ground]575 [:flying]576 [:psychic]577 [:bug]578 [:steel]579 [:ice]580 [:grass]581 [:rock])582 #+end_example584 Electric and Normal are among the best types with which to start the585 game, since they have the fewest weaknesses among all the types.587 At the beginning of the pok\eacute{}mon games, players are given a choice588 between the Fire pok\eacute{}mon Charmander, the Water pok\eacute{}mon Squirtle, or589 the Grass/Poison pok\eacute{}mon Bulbasaur.591 #+begin_src clojure :exports both :results verbatim592 (sort-by pokemon.types/susceptance [[:fire] [:water] [:grass :poison]])593 #+end_src595 #+results:596 : ([:water] [:fire] [:grass :poison])598 As can be seen, the Water pok\eacute{}mon Squirtle is the most solid599 choice starting out, insofar as susceptance is concerned.601 ** The Worst Pok\eacute{}mon Types603 #+srcname: weak-types604 #+begin_src clojure :results silent605 (in-ns 'pokemon.types)607 (defn type-compare-weak608 "compare first by total number of critical-weaknesses,609 then by overall susceptance, favoring weaker types."610 [type-1 type-2]611 (let [measure (memoize (juxt critical-weaknesses susceptance))]612 (if (= (measure type-2) (measure type-1))613 (compare type-2 type-1)614 (compare (measure type-2) (measure type-1)))))616 (defn resistant?617 "might as well get rid of types that are resistant to any type"618 [type]619 (not (every? #(< 0 %) (vals (susceptibility type)))))621 (defn type-successors-weak622 "Generate ways to weaken the given type combination. Discard type623 combinations that either strengthen the given type combination or624 that make it stronger"625 [limit type]626 (set (if (<= limit (count type)) '()627 (filter #(< 0 (type-compare-weak type %))628 (remove resistant? (type-successors type))))))630 (defn pokemon-type-search-weak631 "Search among type-combos no greater than length n, limited by limit632 steps of best-first-search. Find the weakest type combination633 possible in terms of susceptance."634 ([n] (pokemon-type-search-weak n Integer/MAX_VALUE))635 ([n limit]636 (first (last637 (take638 limit639 (best-first-search640 type-compare-weak641 (partial type-successors-weak n)642 (multitypes 1)))))))643 #+end_src646 #+begin_src clojure :results scalar :exports both647 (first (pokemon.types/pokemon-type-search-weak 1))648 #+end_src650 #+results:651 : [:rock]653 Poor Rock. It's just not that good a type. Maybe this is why Brock654 (who has rock pok\eacute{}mon) is the first gym leader in the games.656 #+begin_src clojure :results scalar cache :exports both657 (first (pokemon.types/pokemon-type-search-weak 2))658 #+end_src660 #+results:661 : [:grass :ice]663 # ;;bonus convergently immortal type combo664 # (susceptance (vec (concat (repeat 150 :water) (repeat 50 :poison) (repeat 50 :steel) [:ghost :normal :flying :ground :dark])))666 #+begin_src clojure :results output :exports both667 (clojure.pprint/pprint668 (pokemon.types/susceptibility [:grass :ice]))669 #+end_src671 #+results:672 #+begin_example673 {:water 1/2,674 :psychic 1,675 :dragon 1,676 :fire 4,677 :ice 1,678 :grass 1/2,679 :ghost 1,680 :poison 2,681 :flying 2,682 :normal 1,683 :rock 2,684 :electric 1/2,685 :ground 1/2,686 :fighting 2,687 :dark 1,688 :steel 2,689 :bug 2}690 #+end_example692 This miserable combination is weak to 6 types and double-weak to693 Fire. No pok\eacute{}mon in the games actually has this type.695 * Conclusion697 Searching for a type that is weak to everything takes a very long time698 and fails to reveal any results. That's the problem with a search699 over this large problem space --- if there's an easy solution, the700 search will find it quickly, but it can be very hard to determine701 whether there is actually a solution.703 In the [[./lpsolve.org][next installment]], I'll use =lp_solve= to solve this problem in704 a different way.708 * COMMENT main program709 #+begin_src clojure :noweb yes :tangle ../src/pokemon/types.clj :exports none710 <<header>>711 #+end_src713 ## this is necessary to define pokemon-table inside the source code.715 #+begin_src clojure :noweb yes :tangle ../src/pokemon/types.clj :var pokemon-table-gen-one=pokemon-table-gen-one :var pokemon-table-gen-two=pokemon-table-gen-two :exports none716 <<data>>717 #+end_src719 #+begin_src clojure :noweb yes :tangle ../src/pokemon/types.clj :exports none720 <<types>>721 <<search>>722 <<pokemon-search>>723 <<old-school>>724 <<weaknesses>>725 <<weak-types>>726 #+end_src