Sunteți pe pagina 1din 39

|| A thought on an Algorithm for chess ||

By
Shishirbindu Charudatta Buddhisagar

|| Index ||
1) Opening 2) Chessasamathematicalfunction 3) Fundamentalideology 4) basicstructure 5) Piecevaluecalculation 6) Theminimummovetheory 7) Pieceineveryhouse(subpieces) 8) Subpiecethreat 9) Themobilityissue 10) Whatisattack 11) zeroposition 12) TheKingproblem 13) Threatdelayduetocheck 14) Delaycalculations 15) TheBishopproblem 16) Considerationoflayeredthreats

17) Comparisonofperpieceattackvs.organizedattack. 18) Factorofexponentialthreat 19) Whatissupport 20) Floatingvalueconsiderations 21) Predictingtheblockage 22) Predictingthecheck 23) TheLockingproblem 24) Forcedmoves 25) Predictingthecheckmate 26) KingpiecenearKing 27) TheCastlingproblem 28) ThePawnpieceactivationproblem 29) ThePawnthreatawayfromhomeproblem 30) PawninfrontofPawn 31) Thenpassproblem 32) Thebattlefield 33) MartyrMoves 34) TheThreerepetitionproblem 35) TheStealmateproblem

36) Modelgame 37) Ending

..!!SSSSSSSSSSSSSS!!..

|| Opening ||
Iamdelightedandsatisfiedtobeabletowritedownathoughtofanalgorithmforchess. Manysimpleandsmallthingscompositeacomplexstructure. Anycomplexstructuredisintegratedintosmallerparts,becomeseasyforunderstanding. Notarrogance,butatruthisbuildingasearchtreeisnotanintelligentway. Thisalgorithmmaynotwork,ormaybeitistoopremature. Manyoftheaspectsareneedtobetestedandverified. Manyoftheaspectsofthisarelikearevelationinit,forexamplesupportforapieceisnothing butaproposedattackontheopponentspieceorminimummovelogicorbeitdistributionof influenceetc.andsoIexpectagoodthoughtfromthereaderonthese. Thisalgorithmismainlybasedon5pillars 1) Gainlosscalculationsforafinaltotalandcomparisonofnextmovetotals. 2) Piecevalueandinfluence 3) Minimummovetheory 4) Subpieces 5) Supportisnothingbutaproposedattack. Thesearefollowedbymanyproblemsandchallengeswhicharediscussedinvarioustopics.

Thisalgorithmorfewofitscomponentssuchasfloatingvalueconsiderations,layered threatsoranyothercouldbeusedinconjunctionwiththeconventionalwaysofsearchtree, pruning,Alfabeta,heuristicsandstaticevaluationfunctionsetc. Everyword,everysinglelineofthishasbeenwrittenafteragoodthoughtandthencarefully composed. SoIrequestreadertoreadallthislinebyline,wordbyword. Asyouwillgothroughthissequentiallythechapterswillbecomemoreandmoredifficultand challengingintermsofsolvingtheissues,understandingandvisualization. SubpiecethreatisthemostdifficultintermsofunderstandingwhereasStealmateisin termsofsolving. Althoughwewillbeusingtotalinfluenceofapiecethroughsubpiecesandnotthebasevalue ofapiece,thewholeprocessofcalculatingbasevalueisexplainedjusttotakeyouinthe thoughtflow. Theparagraphsingreenexplainthemissinglinksorrelatedaspectsorsometimesevensimple comments. Irequestreadertoreadthoseaswellwithoutomitting. ThankYou!

||

Chess as a mathematical function

||

Chessisagameofdefiniteruleswhereeverypiecehasadefinitewayofnavigation,scope, influenceandlimitations. Itisplayedwithawelldefinedsetofrules. Thetargetoraimofthegameisveryspecificanddefinite.

Inagivensituationthenumberofpieces,theirpositionontheboardandtheirinfluencesall aretheoreticallycalculableentities. Surelyanopponentsmovedependsonnumberoffactorssuchashisorherown understandingofthesituation,intelligenceandplanningwhichleadstomanyprobabilities. Calculationofallthoseprobabilitiesbybruteforcewiththeavailablecalculationcapabilitiesor RestrictinganddirectingthesearchwiththehelpofHeuristicsandotherartificialintelligence functionsismajorchallengefacedtodaywhiledesigninganalgorithmforchess. Howeverstillwecansaythatforagivensituationthebestpossiblesolutiondoesexists. Reachingforitisthechallenge. {BruteForce:Incomputerscience,bruteforcesearchorexhaustivesearch,alsoknownas generateandtest,isatrivialbutverygeneralproblemsolvingtechniquethatconsistsof systematicallyenumeratingallpossiblecandidatesforthesolutionandcheckingwhethereach candidatesatisfiestheproblem'sstatement. Heuristic:referstoexperiencebasedtechniquesforproblemsolving,learning,anddiscovery. Whereanexhaustivesearchisimpractical,heuristicmethodsareusedtospeeduptheprocess offindingasatisfactorysolution.Examplesofthismethodincludeusingaruleofthumb,an educatedguess,anintuitivejudgment,orcommonsense.}

|| Fundamental Ideology ||

Thefundamentalideologybehindthisalgorithmistocreateaconceptofasysteminwhichwe caninputthedataofpiecenavigation,piecescope,andpieceinfluencewhichwillreflectthe bestpossiblereactionforeveryopponentsactiontotrymaintainingthebalance. ItissimilartoMinimaxtheoremorzerosumgamebutatthesametimeitgetstrickyand complicated. Maybesomelossisinevitable. Itisanefforttobuildanevaluationfunctionofaparticularstatebutwhichincludesthemany variableswhichrepresentthefutureinfluenceandthepriorityisaccuracy,notspeed. {Minimaxtheorem:Minimax(sometimesminmax)isadecisionruleusedindecisiontheory, gametheory,statisticsandphilosophyforminimizingthepossiblelossforaworstcase (maximumloss)scenario.Alternatively,itcanbethoughtofasmaximizingtheminimumgain (maximin).Originallyformulatedfortwoplayerzerosumgametheory,coveringboththecases whereplayerstakealternatemovesandthosewheretheymakesimultaneousmoves,ithas alsobeenextendedtomorecomplexgamesandtogeneraldecisionmakinginthepresenceof uncertainty. Minimaxtheorem Theminimaxtheoremstates[1] Foreverytwoperson,zerosumgamewithfinitelymanystrategies,thereexistsavalueVanda mixedstrategyforeachplayer,suchthat (a)Givenplayer2'sstrategy,thebestpayoffpossibleforplayer1isV,and (b)Givenplayer1'sstrategy,thebestpayoffpossibleforplayer2isV. Equivalently,Player1'sstrategyguaranteeshimapayoffofVregardlessofPlayer2'sstrategy, andsimilarlyPlayer2canguaranteehimselfapayoffofV.Thenameminimaxarisesbecause eachplayerminimizesthemaximumpayoffpossiblefortheothersincethegameiszerosum, healsominimizeshisownmaximumloss(i.e.maximizehisminimumpayoff).

Zerosumgame:Ingametheoryandeconomictheory,azerosumgameisamathematical representationofasituationinwhichaparticipant'sgain(orloss)ofutilityisexactlybalanced bythelosses(orgains)oftheutilityoftheotherparticipant(s).Ifthetotalgainsofthe participantsareaddedup,andthetotallossesaresubtracted,theywillsumtozero.Thus cuttingacake,wheretakingalargerpiecereducestheamountofcakeavailableforothers,isa zerosumgameifallparticipantsvalueeachunitofcakeequally(seemarginalutility).In contrast,nonzerosumdescribesasituationinwhichtheinteractingparties'aggregategains andlossesareeitherlessthanormorethanzero.Azerosumgameisalsocalledastrictly competitivegamewhilenonzerosumgamescanbeeithercompetitiveornoncompetitive. Staticevaluationfunction:Anevaluationfunction,alsoknownasaheuristicevaluation functionorstaticevaluationfunctionisafunctionusedbygameplayingprogramstoestimate thevalueorgoodnessofapositionintheminimaxandrelatedalgorithms.Theevaluation functionistypicallydesignedtoprioritizespeedoveraccuracy;thefunctionlooksonlyatthe currentpositionanddoesnotexplorepossiblemoves(thereforestatic).}

||Basic structure ||
Thebasicstructureisverysimpleandsomewhatsimilartozerosumgamebutitgetsintricateas weconsiderthebehaviorofindividualpieces. {TheseproblemshavebeendiscussedinfollowingchapterslikeKingProblem,Bishop Problem,PawnProblems&LockingProblemetc.} Onemoremajorchallengeiswearenottracingaprobabilitytreeherebutitisnotastatic evaluationfunctioneitheraswewillbecalculatingthevaluesforthosevariableswhoreflectthe futureinfluencesofthepieces.

{ThisisdiscussedindetailinfollowingchapterslikeThemobilityissue,Predictingthe blockage,Considerationoflayeredthreats,Subpiecethreat,Threatdelayduetocheck, Predictingthecheck,Predictingthecheckmate&Predictingthelocksetc.} Nowcomingbacktothebasicstructureitisroughlyasfollows. Considerthateverypiecehasavalue Piecevalueistheinfluenceofthatpieceonotherpieces. Thevalueofapieceisdirectlydependsonitsmobilityforthenextmove. {Now,howtocalculateapiecevalue? Ifitisdirectlydependingonthemobilityforthenextmovethenshoulditchangewiththe situationorremainconstant? ThesetopicsarediscussedinPiecevaluechapter. TheKingposesauniquechallengeinthiscaseasitsmobilityislesscomparedtosomepieces butstillweneedtoprotectit.Wecannotsimplyassignagreatervaluetoit.Itmight unnecessarilyoverprotectthekingbysideliningotherthreats.Alltheseproblemsarediscussed inTheKingproblemchapter.}

Everypiecesthreattotheopponentiscalculatedandstoredasa+veentityforus Everyopponentspiecesthreattoourpiecesiscalculatedandstoredasaveentity

{SupposetherearetwopiecessayB1forblackpieceandW1forwhitepiece. NowwhenwesayB1hassomeamountofthreatfromW1meanshowquicklyW1canmoveto B1splace. NowasapartofstrategyW1mighttakeadifferentorlongerrout.Sohowdowecalculatea valueforthreat? ThisisdiscussedinchapterTheminimummovetheory,Whatisattack,Whatissupport, Pieceineveryhouse(subpieces),Themobilityissue,Considerationoflayeredthreatsand Subpiecethreatetc.}

Whenplayerhastomakeamove,allpossibilitiesforthenextmovearecalculatedandforeach moveatotaliscalculatedwiththeadditionofall+veandveentitiesforthatparticular situationonboard. Intheendthestrongestsumisconsideredasthebestmoveandisplayed. {Stillthisisnotsimplyaconventionalstaticevaluationfunctionaswewillbeconsideringthe futureinfluencesalsoandaccuracyispriority,notspeed.}

||Piece value calculation ||


Everypiecehasavalue. Piecevalueisitsinfluence. Pieceinfluenceisdirectlydependsonitsmobilityforthenextmove. Morethemobilitymoretheinfluence. AQueencanmoveuptomaximum28placesforthenextmovewhereasaRookcanmoveupto maximum14placesforthenextmove.SotheinfluenceofQueenismorethanRook. AndthisissameforallotherpiecesexceptKing. TheKingposesauniqueproblemforthislogicasitsmobilityforthenextmoveislessthan Queen,Rook&Bishopandthusalesserinfluenceandalesservalueeventually. InahypotheticalsituationwherethereisanequalattackonKingandQueenthecalculations wouldcertainlyinclinefortheQueenssafety. Butinreality,givenaperfectopportunityonecouldevenscarifysomepiecesifadefinite captureoftheopponentsKingispossible. Thusasituationispossiblewhereaplayerhassignificantlylessernumberofpiecesbuthasa definitepositionaladvantagetocaptureopponentsKing. {Theabovesentenceisactuallytrickyandsomewhatoutofcontexttoobutpurposefullywritten toenvelopthethoughtprocessforsimilarcircumstances. i.e.Herewearenotcomparingthesumofthepiecevaluesoftwosides.

Wearecomparingallthepossiblesums(totalforgainandloss)ofthenextpossiblemovesofa player. e.g.TherearetwoplayersA&BandsupposeAhas10pieceslessthanB,andsupposeitsA playersturnandhehas14possiblenextmoves,thenwewillbecomparingsumsofthose14 moves. Ahasnotreached10pieceslessinasinglemoveandifalossishappeningataparticularstage isprimarilybecauseofwrongselectionofthemovewhichshouldbesolvedintheevaluation function.} Comingtothepointofbringingimportancetothekinginanequalthreatscenariowithother copieces,itissurethatweneedtoincreaseitsinfluence. Butwecannotsimplygiveanyrandomhighervaluetokingasitmightmisleadthecalculations suchasunnecessarilyovercalculatingathreatforKingoveragenuineimmediatethreatforthe otherpiece. Thisproblemcanbesolvedwiththehelpoftwomethodswithoutunnecessarilygivingahigher valuetoKing. 1) ConsideringathreatforKingonestepearlier. WegivechecktotheKingwhichsaysthatapieceofopponentcannowcometoKingsplace inthenextmove. ThusoverallincreasingthreatstoKingbyopponentsallpiecesonestepearlierand eventuallyincreaseitsimportancealittle. {ThisKingproblemhasbeendiscussedlaterinchapters,TheKingproblem,Considerationof layeredthreats,Threatdelayduetochecketc.} 2) WhenaKingdiesitdieswiththewholegameandallpieces. Thuswecanthinkofafactorwhichcalculatesoverallthreattogameasoverallcombined vevalueofallpiecesagainstthreattoKingbyallopponentspieces. WhereasforothercalculationswecanassignKingitsnormalvaluedependingonitsmobility soasnottoaffectcalculationbyanunnecessarythreatindicationtoit. ThatmeansinfactwearecalculatingthethreatforKingandusingitintwoways. a) CalculatingthreatforKingbyeveryopponentswhilebykeepingitsvaluetruetoits influence. b) Calculatingafactorwhosevaluewillbethesumofvaluesofallpiecesandwhichwill usethethreatdataoftheKingasitsownandwillkeepatrackonmobilityand strategicallysqueezingoftheKingincalculations.

{Thisisdiscussedindetailinchapter,PredictingtheChekmate,considerationof layeredthreats.} ThusasaKingwillbesqueezedbyreducingmobilityandattack,theoverallfactorwillshow signsofdefeatwhereasinotherboardcalculationsKingisplayingwiththetrueinfluencevalue bynotaffectingtheimmediatethreatforotherpieces. Anothertwomajorproblemsare 1)Whetherweshouldkeepafixedvalueforpiecesorshoulditchangeasmobilityofa piecechangesonboarddependingonitspositionandpositionofotherpieces. Mythoughtoverthisisthatapiecevalueisitscapacitywhichisfixedandwhenithasmovedon thecornersoratedgesautomaticallyitwilltakelongerroutstoreachanopponentthusby reducingtheirthreatwhichhasnowbeenalreadyregisteredasaplusfortheopponentin calculationsandnowthereisnoneedtoagainfurtherreduceitsvalueforthesamething. 2) Nowthemobilityofapiececanbeobstructedbyitsownpiecesorbyopponents pieces.Soshouldweconsiderthesetwoassameordifferent? Againsamething. Ifapiecehasbeenobstructedbyitsownpiecesitwilltakelongertimetoreachopponents piecesandthusreducingtheirthreatwhichhasnowalreadybeenregisteredincalculations sononeedtofurtheragainreduceitsvalue. Andifitisbeensurroundedbyopponentspiecesthenagaintwoconsiderations a) Itisabletoposeadirectthreattoobstructingpieces. Ifitisabletoposeathreattosurroundingopponentspiecesthenaplusvalue isalreadybeenincreasedinthecalculationsaccordingly.Noneedtoincreaseor

decreasethepiecevalue,itspositionshasaddedplusorminusincalculations accordingly. b) Itisnotabletoputadirectthreattoobstructingpiecesandispurposefully blockedandtrapped. Ifitisnotposingadirectthreattosurroundingthenthatmeansitsthreattotheopponent pieceshasalreadybeenfurtherreducedinthecalculationssononeedtoagainreducethepiece valueorthesamething. Inshortapiecevalueremainsconstantwhereasitspositiveornegativeinfluenceonthepieces changesaccordinglytoitsownandotherpiecespositiononboardwhichaddspositivesor negativestothecalculationsaccordingly. Wecansaythatthebasicvalueofapieceisconstantbutitsinfluenceonotherpiecesin differentsituationscanbecalledasitstruevaluewhichwillautomaticallyberepresentedinthe gainlosscalculations. Thusaconclusionisthatthebasevalueofapieceisconstant. Nowwhenitissurroundedbyitsownpiecesorwithopponentspiecesyoucannotsaywhenthe surroundingpieceswillmakeawayforitremovingtheblockage,esp.foropponentpieces.But thiswillbeconsideredincalculatingandpredictingthreatandsupportinfluencevaluesfor thosesituations.Basevalueofapieceremainsconstant. {Thiswillbediscussedfurtherinchapters,Whatisthreat,Whatissupport,Mobilityissue, Thelockingproblem,Predictingthelocks,Threatdelayduetochecketc.} Furthermore,inmanysituationsapiecemayhavebeensurroundedbyotherpiecesandhas verylowmobilitybutstillitholdastrategicallyimportantposition(intermsofthreatto opponentandsupport),sotojustincreaseitsvalue(ifitissupposedtobechangingwith mobility)byincreasingitsmobilityweshouldnotoverlookitsstrategicimportance. Nowletusthinkaboutgivinganexactvaluetoapiece. Asdiscussed,valueofapieceisdirectlyproportionaltoitscapacitytomoveinmaximum numberofplacesforthenextmove. AKingcanmoveupto8placesforthenextmovewhereasaQueencanmoveupto28places forthenextmove.

Thisisabigdifference. Butthisdoesntindicatethemaximumtotalpossibleinfluenceofapieceoverthewholeboard. Nowhereweshouldbecomparingthetotalinfluenceofpiecesandnotonlynumberofplacesa piececanmoveforthenextmove. Thiswillbethetotalmaximuminfluenceofthatpieceontheboardwithoutobstructions. Thiswillbethebasevalueofapiece. Nowwhenitwillfaceobstructionssuchasblockingbyotherpiecesorbymovingtocornersof theboard,thesewillbereflectedwhilecalculatingtheattackinfluencewhichwillinturnadd positivesandnegativestototalcalculations(butbasevaluewhichistotalachievableinfluence ofapieceonwholeboardremainsconstant). Totalmaximumachievableinfluenceofapieceisthesumofinfluencesofthatpieceonall influencedblocks. {Notethatherewearecalculatingthebaseinfluenceofapiece.Theactualinfluenceofapiece istheinfluenceithasexertedonotherpieces.} Nowletusconsiderthetotalinfluenceofapieceoveroneblockwhenitcapturesthatblockis 1ifthatpieceismovinginthespecifiedblockinthenextmove. Ifitistaking2movesforapiecetocapturethatblockthentheinfluenceofthatpieceonthat blockwillbe. Sowecansaythattheinfluenceofapieceoverablockis1dividedbythenumberofminimum possiblemovesitwilltaketoreachthatblock. Foreverypieceexceptapawnapieceactuallymovesinthatplace. Incaseofpawnitspathisstraightbutitinfluencesthediagonallyplacedblocksfromthenext row.Soweshouldconsiderthoseblocksasinfluencedandnottheactualoneinwhichthepawn willmove. Morethanthatwealsoneedtothinkonnpassrule. {Thisisdiscussedinchaptersnpassproblem&PawninfrontofPawnindetail.} TheQueenBasevalue Nowtocalculatethebasevalueforthequeenletuskeepitinoneofthecentralblocksandsee howitcannavigatethroughtheboardinminimummoves.

Letusseetheplaceswhereitcanmoveinthenextmove.

Nowletsseetheplaceswhereitcanmoveintwomoves.

Sothetotalinfluenceofthequeenonboardwillbesomewhatlikethis.

ThePawnbasevalue ThereisadifferencebetweenthethreatposedbyanyotherpieceandthreatposedbyPawn. Otherpiecesusuallyreachouttoopponentpiecestoposeathreatandkill,whereasPawnkills thosepieceswhichcomeunderitsdiagonalhouseareaintheadjacentcolumnonly. Pawnhasrestrictionsandcannotgoinallthehousesontheboard.

|| The Minimum move theory ||

Thisisoneofthemostcentralconceptsofthisalgorithm. Chessisoftencalledagameofprobabilityasformanymovesaplayerhasmanypossible optionstotryoutandthosedependonthatplayersunderstandingofthesituation,experience andstrategy. WhenwesaythatablackspiecesayBhasagrowingthreatfromawhitespiecesayW meansWhasagrowingcapacitytoreachatBsplaceinlessermoves.ItmeansthatWis approachingB. Sobasicallyathreatofapiecetoanotherisitscapacitytoreachthatpieceinminimummoves. Therearebasicallyfourthingstothinkabout 1) Therecouldbemorethanonewaystoreachapieceinsamenumberofmoves

2) Asastrategyanopponentmaychoosealongerrout. 3) Whattodowhenanobstructionofanotherpiececomesintheminimumway? A) Whatifthatpieceisofitsown B) Whatifthatpieceisofopponents?Youcannotsaywhenitwillmove. 4) Whatiftheobstructionhasevenfurtherobstructionsinturnwhichmakesitsmobilityfor thenextmovetodecrease? {ThisisbeenevenfurtherdiscussedinchapterslikeWhatisthreat,andWhatissupport.} Nowletusconsiderthesecasesonebyone Now,therecouldbemorethanonewaystoreachapiece. Primarily,whenwesaythatanopponentspiececankilloneofourpiecemeansitcannow reachtoitinfewmoves.Sothethreatisbasicallyinhowmanyminimummovesapiececan reachtowardsanotherpiece. Nowwhenwesaythatweneedtofocusonmainlytheshortestpossiblewayapiece reachestoanother,thenalsotherecouldbemorethanonewaysforapiecetoreachto opponentspieceinsamenumberofminimumpossiblemoves. Primarilyinthis,thepathdoesntmatterasfarasthethreatisgoingtoremainsame. Andfurthercomparisonsbetweenthetwopossibleminimummovepathswillbe automaticallydoneonthebasisofhowitaffectsotherpieceswhoareassociatedwiththese routes,andwillbereflectedinthecalculationofthosepiecesandthusinthefinal calculationtoo. Furthermorethesecomparisonswillbedoneeasywhenwewillbeputtingasubpiecefor everypieceineveryhousethatwillhelpcalculatetheinfluenceandchangesinitofallthe piecesonallplacesoftheboard. {} Nowsupposefromastrategicpointofviewanopponentchoosestoattackonapiecefrom alongerroutethenalsoeverymoveoftheopponentspiecehasbeenregularlymonitorsin thecalculationsofthreat,plusthroughthecalculationsandinfluencesofsubpieces.So oncetheopponentstartsbuildingonthepressureofattack,itautomaticallyagaingets reflectedandupdatedthroughinfluencesasitapproachesthepieceorstartsblockingit. Thishasbeenfurtherexplainedinthebelowexample. {} {Thisisfurtherdiscussedinchapterpieceineveryhouse(subpieces).} Nowwhatiftheshortestroutehasobstructionsduetootherpieces? Inthiscasesupposethereisanyobstructioninthepathduetoanyotherpieceofitsownor ofitsopponent,thepathintoconsiderationi.e.thenumberofmovesitistakingtoreach

theotherobjectshouldbeaddedwiththenumberofminimummovesrequiredtoremove thatblockorpiece,andthenitshouldbecomparedwiththesecondbestpossibleminimum route.Ifstillitholdsthepositionofaminimumorshortestroutethenitshouldbecounted so,elsethesecondbestshouldbetheminimummoveway. {Infactlateronwewillbeconsideringasubpieceforeverypieceineveryhousewhichwill behavingavaryingvaluedependinguponthepositionofitsparentpiece,andthesesub piecesareboundtocreatesomeobstacleswithvaryingfloatinginfluence. Soinfactaminimummovepathoranypathisalwaysfloodedwithobstacleswithfloating valuesandisdelayedaccordingly. Thisisfurtherdiscussedinthefollowingchaptersasitisoneofthecorepartsofthis algorithmwhichtriestodifferentiatethisfromcontinuousiterations,orprobabilitytree creation.Asofnowitisbettertorestrictourthoughtsonlyonoccurrenceornon occurrence(yes/no)ofobstacleinatraditionalway.} {} Nowtheblockingpiececouldbeourownoroftheopponentandwhethertomoveitnow orlaterdependsonthestrategyandpositionaladvantages.Butthenthesebenefitsand lossesarebeingcalculatedinattackanddefensecalculations,anddependingupontheloss orgainwemighttakeadecisionbutasofnowwearecalculatingthebestminimumway existingonboard.Itisjustlikeamagneticfieldofonepieceontheother,whetheritwillbe enforcedordiffusedwilldependoncomparisonoffinaltotalsoflossandgain. Nowsupposethereisanobstructionofapiecewhichcouldbemovedforthenextmove, butifremoveditmightcausesevereloss,thenthislossisalsobeingalreadyconsideredinto thecalculationsandthatshouldntpreventusfromcalculatingtheinfluenceofpiecesif supposeitsremovedfromthenextshortestpossiblemove. {}

|| Piece in every house (Sub Pieces)

||

Thisisoneofthecentralconceptsofthisalgorithmtocalculatethedistributionofinfluence ofapieceineveryhouseoftheboard. Whilecalculatingthebasevalueforapiecewehadputallthepiecesinallthehousesand calculatedthevaluesforthatpieceinthathousedependinguponthenumberofminimum movesrequiredtoreachthere.

Nowletusconsiderapieceasasubpieceineveryhouseforallblackandwhitepieces. Valueofasubpiecewillbeequaltoitsparentormainpiecewhentheyareinthesame houseanditwilldecreaseaccordingtothenumberofminimummovesparentpiecetakes toreachthehouseofthatsubpiece. Thusasubpieceisarepresentativeofthemainpieceinthatparticularhouseanditsvalue changesaccordingtothepositionofthemainpiece. Asubpiecehelpsustocalculatethedistributionoftheinfluenceofapieceovertheboard. Similarlyopponentspiecesalsohaveinfluenceonsubpiecesandviceversa. Sowhenapiecediesallthesubpiecesandtheirrespectiveinfluencesontheopponents piecesalsovanishes. Thustheactualinfluenceandvalueofapieceisitscollectiveinfluenceofallsubpieces. Thisfurtherhelpsstabilizetheinfluencesandreducetheinstantfluctuationsincalculations. {ThisisexplainedinchapterTheBishopproblem} Thiswillhelpuscalculatemanythingssuchas, 1) Comparisonbetweenminimummovepathsinpiecethreats. 2) Calculatingthemobilityofapiece 3) PredictingtheLocks 4) Predictingtheblockage 5) Predictingthecheck 6) Predictingthethreatdelayduetocheck 7) Predictingthecheckmate 8) Piecetruevalueandoverallinfluencecalculations 9) CalculatingPawnpieceactivations 10) SolvingNPassproblem 11) CalculatingPawnthreatawayfromhome 12) SolvingCastlingProblem 13) PredictingMartyrmoves Inthischapterwewilldiscussonlyaboutcomparisonbetweenminimummovepathsin piecethreatsandpredictingtheblockage,allotherthingswillbediscussedintheir respectivechaptersdedicatedforthem. Soeventuallyasubpieceexistsinallhousesanditdoesntactuallychangesitspositionbut changesitsvalue,andthevalueofthesubpieceisequaltothevalueofitsparentpiece whentheyareinasamehouse.Thismeansthatapieceactuallydoesntexists,whatexists areonlysubpieceswithvaryingvalues(andthatswhythenameofthealgorithmas

Diffuse).Butforthesakeofunderstandingwewillreferthesubpiecesaspiecesthathave valueequaltotheirparent. Piecesinfluence(opponents)pieces. Subpiecesinfluence(opponents)subpieces. Everysubpiececreatesanobstacleatitsplace. Thestrengthofthisobstacledependsuponthevalueofthesubpiecewhichgradually increasesordecreasesinsituations. Sobasicallyitislikeafluidofinfluencefloodedeverywhereonboardwithvaryingdensity. Nowwheneverwehavetocalculateaminimummovepathbetweenanytwosubpieces,it isalreadyfloodedwithothersubpieceswithvaryingvaluesasobstacleswhichcombined delaythatpathaccordingly. Soanyobstacleinanywayisnotasuddenappearancebutitgainsitsorlossesitsvalueina floatingway.Andthepathisdelayedaccordingly.Inthiswaywecanpredictaconsiderable buildingofablockageataparticularhouse. {ThisisalsodiscussedinFloatingvalueconsiderations.}

|| Sub-piece threat ||
NowsupposeatplaceA,whitesXhasapiecein2movesandatplaceBblacksYhasapiece inthreemoves.AtaplaceCanobstacleisbeingbuiltaftersayfourmoves.Butallhavebeen alreadyrepresentedbytherespectivesubobjectsinthoseplaces. Nowwhilecalculatingtheinfluencesbetweenthesubpieces,againwehaveminimumways betweentwosubpieceswecanconsidersubsubpiecesforasubpieceandsoon.And againthesemighthavesubsubpiecesasobstacles.Thusagainitmaygooniteratingand willeventuallycreateaprobabilitytree.Soathoughtcomestomindthatmaybeevenafter calculatingvaluesandpositionswithminimummovetheorywearenotcalculatingany threatlogicandeventuallyprobabilitytreeisinevitable. But,asubpieceexistsineveryhouse. Itcreatesanobstacleatthatplace. Thefloatingvalueofthesubpiecesisconstantlyindicatingthefloatinginfluenceatthat place. Itsvaluecanbemaximum1whenitsparentpieceisinthesamehouse. Inanyminimumwaywetakeintoconsiderationallobstructingsubpiecesanditdelaysthe pathaccordingly.

SoevenfortheaboveexampletheobstacleatCisbeingbuiltmaycounterandanobstacle ataplacesayD,isalreadybeingreflectedinthecalculationswiththefloatingvalueofsub pieceswhicharealreadypresentatD. Andthusitrepresentstheinfluenceforthefuturemovesjustlikeaprobabilitytree. {}

|| The Mobility issue

||

Insomesituationaconflictarisesbetweenthemobilityofapieceagainstthreattoanother piece. Inshort,youcouldfaceasituationwhereatonehandyouhaveapiecebeingblockedfrom allsidesandimmobilizedwhereasanotherpieceisinthreat. Nowthepriorityentirelydependsontheoverallboardsituation. Andtheboardsituationcanbebetterrepresentedwiththehelpofallsubpiecesandtheir influences. Butthemainchallengeishowtocalculatethetruechangeininfluenceofapiecedueto changeinitsmobilityaccurately. Butthisproblemhasalsobeenresolvedwiththehelpofsubpiecesandtheirinfluences. Whenthemobilityofapieceisdecreased,itcausesasignificantdepressioninthevalueof itssubpiecesandtheirinfluencesandthecollectiveeffectislarge.

OneofthebestexamplesisonewhichisgivenbelowwheretheQueenisunderimmediate attackandtheKingisatthecornerafterCastlingwiththreePawnsblockingitinthelastrow andtheopponentsRookissettoCheckMateinnextmove. ThisistheultimatetestofKingsinfluencecomparedwithQueen. {} ButstillinanycasepureinfluenceofQueenwillbealwaysmorecomparedtoKingasitalso hassubpiecesandtheirinfluences.AndforthistheKingsvalueandinfluencecanbe increasedwiththehelpofconsideringfactorssuchasconsideringthreatforKingonestep earlierandThreatdelayduetocheck. Nowinthiscase,thesubpieceofRookinthehighlightedhouseshouldaddasignificant threatinthecalculationsaswellasthreatdelaysonopponenttopredictacheck. NowifwemoveanyPawninfrontoftheKingormoveKingonemoverightsoastomakean escapeway,itsignificantlyincreasestheinfluenceofitonopponentspiecesbutstilleven thenatcheckmovingKingshouldremainthepriorityanditwillbeifitisfortheabove mentionedsituationtoo.

|| What is Attack ||

Thebasevalueofapiecehelpsuscalculatethedistributionofitsinfluenceonboard. Thetotalinfluenceofapieceisthecollectiveinfluenceofallitssubpieceswhichisinfactits truevalueanditsattackontheopponent. Totalcollectiveinfluencesofallpiecesisthetotalcollectiveattack. Everyhouseontheboardinfactbecomesabattlefieldwithdifferentsubpiecesofblack andwhiteputtingtheirinfluenceinthere. Astheinfluenceofpiecesfromonesideincrease,itreducestheinfluenceofopponents pieceinthathouseaccordinglyandthusincreasinginfluenceofpiecesofonesidestart creatingablockageandattackdelayfortheopponentinahouse. Afterknowingallthisstillitsnotclearwhenonepieceattackstheotherandhowshouldwe putitinthecalculations? Itstimetodecidethevalueofasubpieceonceandall. Therearefewoptionsforconsideration.

1) Basevalueofitsparentpiecedividedbythenumberofminimummovesrequiredfor parentpiecetoreachinthathouse. 2) Influenceofaparentpieceinhouseismaximum1forthatparticularsubpieceofthe parentandvariesaccordingtothenumberofminimummovesittakestoreachthere. Thesecondexampleissimilartowhatwehavedonewhilecalculatingthebasevalueof apieceandlooksmorerelevant. Thebasevalueofapiecehasalreadygivestheideaofthetotaldistributionand comparativedifferenceoftotalinfluenceofapiece.Andineveryhousebysettingits maximumas1foranypiece,isitQueenoraPawnwehavedonethis. Thefirstexampleisfaultyasweareconsideringthebasevalueofapieceincalculationof everysubpiecethuscountingthesamethingrepeatedlyandmultiplyingtheeffect. Sonowwewillconcentrateonthesecondexamplebutwehaveonemoreproblemathand. Thatis,whatvalueforthesubpieceweshouldconsiderwhichishostingtheparentpiecein itshouse? Onethingfromtheconsiderationofthesecondexampleisclearthatthisvalueshouldbe sameforallthepiecesonboardwhetheritisQueenorPawn. {ThisdiscussionisnotdoneinPiecevaluechapterastheconceptofSubpieceswasnot introducedatthattimeandatruepiecevalueisinfactitsattack.} {ComingbacktothepointofcalculatingthevalueofaSubpiecewhichishostingtheparent pieceinitshouse,thistopiciselaboratedinchapterZeroposition.} Nowagaincomingbacktothepointofputtingallthisincalculationsweneedtoanalyze everyhouseseparatelyasasmallbattlefieldintherebetweeninfluencesofallpieces. SupposethereisapiececalledWofwhiteandtwopiecesfromblack,sayB1andB2 respectively. InaparticularhousetheinfluenceofWi.e.thevalueoftheSubpieceofWinthathouseis say1asitisabletocomeinthathouseintheverynextmove. ThevaluesforSubpiecesforB1&B2inthathouseissay0.5&0.3respectively. ThebattlebetweenWandB1inthathouseis(10.5)andthebattlebetweenWandB2in thathouseis(10.3). NowkeepinginmindthelogicofMinimummovetheorywhereweconsideronlythe maximumpossiblethreatfromanopponentspiece,similarlyherewecansaythatthe threatfromB1isgreaterthanthethreatfromB2anditalreadyshowshowmuchquicklyW cangetkilledandsoweshouldbeconcerningonlyaboutthatthreatandaddingother

threatsforthesamepiece(Althoughithasbeenalreadydefinedhowquicklyitcanbekilled) wewillbeunnecessarilymultiplyingit.Butthethingisitinfactdoesntmultipliesthethreat whenweconsiderthreattoapiecefromallopponentspiecesastheyarefromdifferent piecesandweshouldbecountingthemallthetimeandforthesamepurposewehave implementedthelogicofSubpiecesforeveryPiecesothatwecantrackthechangein influenceofallthePiecesonboard.

|| Zero position ||

Herewehaveoneunsolvedprobleminourhandthatis,whatshouldbethevalueofaSub piecewhichhoststheparentPieceinitshouse? Wehavealreadysaidthatandhaveconsideredwhilecalculatingthebasevalueofapiece thatthemaximumvalueofaSubpieceinahouseis1ifitcanreachtoitintheverynext move.Andforotherhousesitis1dividedbythenumberofminimummovesrequiredfor thatpiecetoreachthathousei.e.ifitreachesahouseinsay2movesthenthevalueofthe Subpieceshouldbei.e.0.5forthathouse. OnemoreassumptionwehavemadethatthevalueofaSubpieceisequaltoitsparent pieceiftheparentpieceisinthesamehouseorinotherwordsifthatSubpiecehoststhe parentpiece.Andinthelastchapterwehavealsosaidthatthisvalueshouldbesameforall thepieceswhetheritisaQueenoraPawnandthedifferencebetweentheirinfluenceswill bereflectedthroughthevalueandnumberoftheirSubpieces. Keepingthesamelogicinmind,wecansaythatthevalueofaSubpiecewhichhoststhe parentpieceisinfinityasthenumberofminimummovesrequiredforittoreachthereis zeroasitisalreadyinthesamehouse. Soagainthequestionis,ifnotinfinitywhatshouldbethevalueofaSubpieceinthat house?Shoulditbebig?Andhowmuchifso? Toavoidthiswecandoasimpletrick. ConsiderthevalueoftheSubpiecewhichhoststheparentpieceas1forallpiecesasifthey havearrivedtherefroma0thimaginarypositioninonemove.Andforthenextmove(which isinfacttheimmediatefirstnextmove.)wecanconsideritasi.e.1dividedbythenumber ofminimummovesrequiredforthatpiecetobeinthathouse. EvenafterconsideringthiswhenwecalculatethevaluesforallSubpiecesforapiece,the basicstructureremainssameandcomparativevaluesofpiecesalsoremainsameavoiding theinfinityproblem.

{Maybetheproblemneverexisted,wecreatedandsolvedit.}

|| The King problem || || Threat delay due to check ||

ThisisoneoftheimportantaspectstogiveKingitsdeservedimportancewithout unnaturallyincreasingitsvalue. WhenevercheckisgivenweneedtoremovethethreattotheKingeitherbymovingitorby blockingthecheckwiththehelpofotherpieces. Thisinturndelaysthethreattoopponentspiecescausedbyourswhichindirectlyincreases theimportanceofKing. AlthoughthereisnochecktotheKingataparticularsituation,manysubpiecesof opponentspiecesarealwaysinapositiontogivechecktotheKing. Buttheirvaluehasnotreachedtothevalueoftheirparentpiece. InfactthethreathasnotreachedtoitsmaximumforKing. Butitisalwaysavaryinganditdelaysthethreatofourpiecesforopponentaccordinglyina floatingvariation. Sothismeansthatwheneverthereisanydelayitisnotinintegerslikeonemoveortwo movesbutitisinfactafloatingvalue. Similarlywheneverthereisanyblockagebetweentheminimummovespathbyanother piece,itisalsonotsuddenbutisbuiltgraduallybyincreasingvalueofsubpiecesandis floatingandbecomespredictable. ThisisthemaintopicofdiscussioninFloatingvalueconsiderationschapter.

|| Delay calculations ||
Differentpiecesreactdifferentlytoblockagesandgetsadelayontheirpathaccordingly.

IfaRookorQueenencountersanopponentspieceonitslinearwayalongacolumnofrow orevenonadiagonalway,ithastofirstkillthatpieceandmoveforwardproducingadelay ofonemove. WhereasaKingmightsimplykeeponmovingbykillingthatpiecewithoutcreatingadelay foranymoveinthesamesituationifthatpiecedoesntposesathreattoKing.Andifitpose athreattoKing,thenagainthesituationisdifferentanditcreatesadelayaccordingly.

||

The Bishop problem

||

TheBishopProblemorproblemwiththeBishopisthatBishopfromthewhitehousecantgo toblackhouseandviceversa. Thus,asituationofconflictmayarisewhereifwemovetheQueentotheblackhouseto avoidthethreatfromtheopponentsBishopinwhitehouse(whichisnotanimmediate threat)raisesthepositiveinthetotalcalculations(asnowthereisnothreatwhatsoeverto QueenfromBishop)morethanthatofanimmediatethreatforaPawn. Nowwhetheritshouldbeornotentirelydependsontheboardsituations. ButthisalsohasbeenmadefairerbysubpiecesofQueen,BishopandPawn. AssubpiecesofQueenarestillbeingaffectedbyBishopandhissubpiecesithasreduced theintensityofsuddenON/OFFofthethreattoQueenbyBishop.

|| Consideration of layered threats ||


Layeredthreatisnothingbutcreatingandreferringtodifferentthreatcharts. Mainlywewillbeusingthisin 1) 2) 3) 4) PredictingtheCheckmate. DecidingpieceactivationinthePawnslasthouse. Predicting(andforcing)Stealmate Predicting(andforcing)martyrmoves.

|| Comparison between per piece attacks vs. organized attack ||

Organizedattackisbasicallyatraponopponentspieceorpiecesbygraduallydecreasing theirmobilityandincreasingtheattackthreat,whichifdisintegrated,isnothingbuta collectiveperpieceattackofallinvolvedpieceswhichisalreadybeingreflectedinthefinal gainlosscalculationsoftotalfortherespectivepiecesfromthemomentwestartthetrap. Sothebestresponsebeingcalculatedthroughcomparisonoftotalsisnothingbutatrap beingset. Thecalculationshouldinvariablyshowustothebestpossibletrap.

|| Factor of exponential threat ||


Thisisoneoftheoldestconceptsofthisalgorithm.Itwasdesignedwellbeforetheconcept ofsubpieceandmayhavelostitssignificanceaftertheinclusionofsubpiecesormaybestill wecanuseitinconjunction. Thiswasdesignedtosettheprioritiesforthreats. Nowsupposetherearetwopieces,oneisaPawnandtheotherisQueen(Thedifference betweentheirvaluesifmorethanbetweenanyothertwopieces). NowsupposethereisadirectthreattothePawnandthereisthreattoqueenalsoatnext move. NowasQueensvalueismuchmorethanthePawn,itmayhappenthatinthissituationalso thecalculationswilltendtoprotecttheQueen(whichcanbemovedonthenextmovealso )atthecostofPawn.

{Thisproblemhasbeenremovedwiththehelpofsubpiecesaswecanmoreaccurately calculatethevaluesofpiecesandwhetherapieceshouldbesavedorleftfordying dependingupontheoverallsituationatthattimeonboard. Also,wewillbeconsideringsubpiecesforthepieceswhichwillbeactivatedwhenaPawn reachestothelasthouse.ThesesubpiecesgaintheirvaluestepbystepasaPawn approachesthelasthouseandthushaveanincreasinginfluence.SolosingaPawnalsocosts lossforthesesubpieces. ThishasbeenfurtherdiscussedinchaptersrelatedtoPawnproblemslike,Pawnpiece activationproblem,Pawnthreatawayfromhomeetc.} Nowcomingbacktothepointofconsideringanexponentialthreatfactor,itisanumber whichwewillbesubtractingfromtheminimummoveswhilecalculatingtheinfluenceor threatofapieceataparticularhouse. Aswehaveseen,maximuminfluenceofapieceinahouseis1andisreducedasitmoves awayfromit.(Thiswehaveseenatthetimeofcalculatingpiecevalues) Soapiecereachingtoahouseintwomoveswillhavehalfoftheinfluence.i.e. Andsoforapiecewhichwillbereachinginthreemoveswillbe1/3 Soitis1dividedbyminimummoves,letssayM. Soathreatofapieceinahouseisalways1/Mofthatpieceforthathouse. NowletusconsideranexponentialthreatfactorasXwhichwewillbesubtractingfrom minimummoveM. Sonowthethreatwillbe1/(MX). NowletusfornowconsiderthevalueforthreatexponentialfactorXas0.3.(Nowthisisa topicofresearch.) Sonowsupposeifapieceismovingtowardahousestepbystepstartingfromminimum5 moves,theincreasinginfluencewithandwithouttheexponentialthreatfactorwillbeas following. Withoutexponentialthreatfactor:1/5,,1/3,,1/1.Whichisalineargraph. Withexponentialthreatfactor:1/(50.3),1/(40.3),1/(30.3),1/(20.3),1/(10.3) Whichisultimately:1/(4.7),1/(3.7),1/(2.7),1/(1.7),1/(0.7)whichisanexponential graphanditincreasestheimportanceofthreatmorethanlineargrowth.

{AsIsaid,thiswasoriginallydesignedmuchmorebeforetheinclusionofsubpieces, whenIconsideredthevaluesofpiecesdependingupontheirmobilityforthenextmove withoutsubpieces.I.e.28forQueenand8forKing. Sothedifferencewasproportionatelymuchmore. WhenconsideredQueenandKingbothcomingunderattackundernextfewmovesand thatweconsiderthreatforKingonestepearlier,28/(20.3)islessthan8/(10.3).

|| What is support

||

Supportisnothingbutaproposedattackontheopponentspiece. Whenwesaythatapiecehasasupportoftheotherpiecemeansthatotherpiecehasposed tokillthekillerofthefirstpieceifitconquersthat. {Istartedwithaddingapositivewhenitwassupportedbyotherpieceandrealizedthatitis unnecessarilyincreasingtheimportanceofpiecesthataresupportedbymanyoverthe pieceswhornotimmediatelysupported. Furthermoreassupportisnothingbutaproposedattacktotheopponentspiecesthe positivesforthesupporthasalreadybeencalculatedinfinaltotalthroughthissononeedto furtheraddvaluesforsupport.}

||

Floating value considerations ||

Asubpiecerepresentsaparentpieceinthathouseaftercertainmoves. Thismakesitpredicttheapproachormovingawayoftheparentpieceatthatplaceinstead ofsuddenappearanceordisappearanceofapieceinahousewhichisveryimportantin tacticalcalculations. Thisisinawayfloatingvaluerepresentationofapieceinahouse. Nowifweapplythistoeventhepathsornumberofmoves,itgivesustremendousflexibility incalculationsanddoesntaffectthebasicTotalcomparisonidea.

||

Predicting the blockage ||

Subpiecesfilltheboard.Sotheyareboundtocreateanobstructionformanyofthe minimumpathsofmanypiecestoother.Theirvalue(influence)changesdependingupon thepositionoftheirparentpieceandwhenitisequaltoitweconsiderapiece(theparent one)atthatplacewhichaddsonestepdelaytoallthecrossingminimummovepathson whichitrests. Butbeforethatwhenitsvaluewasnotmaximumtherewasnoparentpieceatitsplaceand virtuallynoblockagetotheminimummovepathscrossingit. SothismakesblockageONwhenitsvalueisequaltoitsparentpiece,andOFFwhenitisless thanthat. Thisdoesntpredicttheblockageandmakesitsuddenlyappearataparticularmove. Soinsteadofthisweconsiderablockagefromthatsubpieceinafloatingmannerasits valuebeingincreasedbyapproachingparentpiece,wecandelaythecrossingminimum movepathsaccordinglyinafloatingwayinsteadofintegervalues. Nowthiswillreflectthegradualbuildingofablockageataparticularplaceandmakesit predictable.

||

Predicting the check

||

Asweconsiderasubpieceforeverypieceineveryhouseandthenthreatsalsofromthe subpieces,obviouslymanyofthesubpiecesareincheckpositionforKing.Andasthevalue ofasubpiececausingcheckincreasesdependinguponthepositionofitsparentpiece,we canpredictacheck(asweunderstand)inhappening.

Infact,checkisnotsomethingwhichisgivenataparticularmovebutisalwaystherefrom subpiecesandhasafloatingvalueinfluencedependinguponthefloatingvaluesubpieces andbecomespredictablethroughthem. {Everyfloatinginfluencecheckbyasubpieceinfluencesadelayinthreatforopponentby allourpiecesaccordinglyandthusgivesKingitsdeservedimportanceincalculations.}

||

The Locking problem

||

WhenweobstructachecktoourKingwiththehelpofourownpiece,wecansaythatthe pieceislockedasitcantbemovednow. Similarlyopponentalsotriestolockourpiecesasastrategicgain. Againthislockalsocanbepredictedthesamewaywecanpredictacheckwiththehelpof subpiecewhichhasalreadyappliedachecktoKing. Alockaddsadelayinthethreattoopponentsallpiecesfromthelockedpiece. Thusthegainsandlossesofthisareautomaticallybeingcalculatedandthereisnoneedto addanyextracomponenttocalculationstoavoidit.

||

Forced moves

||

Themoveswhichareinevitableforaplayerareforcedmoves. Anyothermoveissimplyillegal. Wecanverywellrecognizeamoveasforcedmove. MainlytheKingisinvolvedinthis.

||

Predicting the checkmate

||

||

King piece near King

||

Infact,wecantputKingPieceinadjacenthouseofanotherKing. Tillnowwehavenotmadeanyprovisionincalculationstostopthisfromhappening. Butweneednottoworryaswhenthisisdoneforcalculatingatotal,automaticallythe threatfactorforKingwillbesohighthatthestepwillbeavoided. {Thereisonechancethatakingisbeinggivencheckandisblockedinallsurrounding houses,oneofthoseisadjacenttoopponentsKing,andthenitmightchoosetojumpin thatasitalsoincreasesopponentsthreattoo.ButthishavebeendiscussedinThesteal mateproblem,andPredictingthecheckmate.}

||

The Castling problem ||

AfterCastlingKingandRooktakeanewpositionotherthantheonewhichtheywouldhave takennormally. IntheseplacesalsowecanconsideraspecialsubpieceforKingandRookandcalculatetheir floating(asCastlingtoodoeshaveobstaclesanddelays)influenceonotherpiecesonboard. SimilarlysubpiecesofRookandKingalsohavetheirinfluencesonothers. NowherewehavetwokindsofobstaclesforCastling. 1) Obstacleduetopieces(ourownoropponents) 2) AttackonKingorRookoranyotherhouseinbetweenthem. Dependinguponthese,theCastlinggetsdelayed. NowshouldweconsidertheCastlingrelatedsubpiecesofKingandRooklittledifferentthan othersastheyshouldnothaveanyinfluenceandshouldsimplyvanishiftheCastlingisnot doneorcancelled?Ifyes,thenthischangewillbelittlesuddenandjerkyandnotsmoothfor thecalculations

Nowwehaveatrickysituationinhand. Keepinginmindthelogicofsubpieces,wehavealreadyputtheminalltheplacesforRook andKingalsoandtheirvaluedependsuponthenumberofminimummovesrequiredforthe respectiveparentpiecetoreachthosehouses. Wewillthinkontwooptions. 1) ConsideringtwoseparatesetsofsubpiecesforRookandKing,oneforwithCastlingand theotherforwithoutCastling. 2) WhiledecidingthevalueofasubpieceofRookorKing,consideringtheminimummove waybetweenthepieceandthesubpiecebycomparingtheshortestpossibleonewith CastlingandonewithoutCastling.Andhoweveristheshortestinbetweenthesetwo shouldbethefinalminimummoveway. Nowconsiderthefirstexample. IfweconsidertwoseparatesubpiecesetsforRooksandKingwithCastlingandwithout, theneitheronecompletesetofsubpiecesshoulddieifCastlingisdoneorcancelled respectively.Thischangecantbesmoothandwillaffecttheboardevenmore. Nowconsiderthesecondexample HerewewillbekeepingonlyonesetofsubpiecesforKingandRookandwewillbe calculatingthevaluesforthesamebycomparingshortestbetweenminimumwithCastling andwithoutandhoweveristheshortestwillbethefinalminimummoveway.Thissounds morelogicalandalsointunewiththeminimummovetheoryanddoesntaffecttheboard calculationsinajerkywaysuddenly. EventheCastlingiscancelledthenalsothesubpieceshavevaluescalculatedinnormalway byminimummovetheoryandthepiecedoesntgetvanishsuddenly. ThisisalsoapplicablefortheotherRookwhichwasnotinvolvedintheCastling.

||

The Pawn piece activation problem

||

WhenaPawnreachestothelasthouseitcanactivateapiece. Nowwehavethreemajorproblemsinhand. 1) Whichpiecewillbeortobeactivated 2) Howtopredictandcalculatetheinfluenceofthepieces. 3) WhatifthePawnkillsapieceandchangesthecolumn

WecanconsiderasubpieceinthelasthousewhichgainsitsstrengthstepbystepasPawn approachesit. WecanalsoconsidersubpiecesforthePawninallthelasthouses. Asweknow,aPawncanchangeitscolumnifitgetsanopportunitytokillapiece.Thusthe pieceinlasthouseitmayactivatemayalsobeinanothercolumnandthevalueofitmay changedependingonthenumberofmovesthatPawnmaytaketoreachthere. NowthesestepsvarydependinguponotherpiecesalsowhichmaycomeinwaytoPawns threatzonesoasitcankillthemandallthesepossibilitiescanbecalculatedineverymove bysubobjectsofotherpieceswhicharealreadyinPawnsattackzone.

||

The Pawn threat away from home problem

||

WhilecalculatingthevalueofthePawnwehaveseenthatthethreatfromaPawnisnot whereitismovinginastraightcolumnbutonthediagonalsidehousestothem. APawncanchangemultiplecolumnsbykillingmorethanonepiecediagonallyandthuswe needtocalculateitsinfluencetotheseplacestoo. ButamajordifferencebetweenthreatfromsubpiecesofotherpiecesandPawnisthatfor aPawntokilldiagonallyithastowaitforotherpiecestocomeinitswaysoastobekilled. Soitsthreatissomethingwhichcouldbecompletelyavoidedbytheopponent. Thisissimilarproblemofdecisionmakingaswasincalculatingtheblockagefromourpieces andopponentsaswecantsurelytellwhenopponentwillremoveitspieceandthe blockage. Butagainsimilarlyhere,weneedtocalculatethepossibleinfluencewhichislikeadanger zonearounditorlikeamagneticfield,whoevercomesinbetweenwillfeeltheinfluence, whetherotherpiecewillcomeunderitsinfluencebyselforistobeforcedontoitby movingourpieceisanothermatter. SimilartoBishopPawnalsocantgoinallhousesbutstillmanymovesofitcanbe predictableasitaffectsthesubpiecesofotherpiecesinthosehouseswhereitcouldgo. ThisdoesntlettheinfluenceofPawnonapiecesuddenswitchonoroff.

||

Pawn in front of Pawn ||

||

The n pass problem ||

ThisissomewhatsimilartoCastlingproblembuthasawiderimpactonboardsituationas theattackingPawnhasalreadyreachedinthesecondhalfoftheboardthusbyincreasing theinfluenceofthehiddenSubpiecesofthosewhowill/maygetactivatedifthePawn reachestheendhouse. NowsupposethereisaPawnfromwhiteintheF5andaPawnofblackisinE7. NowtheimmediatethreatfromthewhitePawninthenextmoveisinE6andG6. IfthePawnofblackinE7jumpstoE5

||

The battle field ||

Inchesswhentwopiecesfromofoppositesidehavingmorethanonesupportposeadirect threattooneorotherwecansaythatabattlefieldisbeingcreatedataparticularhouse. Insuchsituationsmorethanonepiecesometimeseven4/5piecesfrombothsidesget involvedinthis. Inthisthesequencingoftheusageofpiecestobeusedisveryimportant. NowsupposeaBishopofwhitehassupportfromQueen,Rook&PawnandsupposeBishop ofblackisindirectattackwiththeBishopinwhiteandishavingsupportoftwoPawns,Rook andQueenrespectivelyfromitsside. {well,asmentionedearlierasupportisnothingbutaproposedattack. ThismeanstheQueen,Rook&Pawnofwhitehaveactuallycreatedapotentialthreattothe Bishop,twoPawns,Rook&queeninthehouseofwhiteBishop.} NowinthissituationuseofthequeenshouldbedoneatthelastwhereasuseofPawns shouldbedoneinthebeginningofbattle. Infactweactuallydontneedtoworryandmakeaspecialarrangementaboutthisasthis hasalreadybeentakencareofinthetotalcalculationswiththehelpofSubpiecesandtheir influencesfrombothsides.

Forexample,ifweuseQueenatthebeginningofbattleandlosethatpiecethen automaticallywewilllosetheinfluenceofitsSubpiecesandamajorlossisregisteredinthe finaltotalofgainlosscalculations. Thussuchmovewillobviouslynotreflectabiggestfinaltotalandthemovewillgetavoided. NaturallythecarefortherightsequencinghasbeentakenthroughSubpiecesandallthe influences.

||

Martyr moves ||

Martyrmovesaretheonewhereapiecesacrificesitselfforabiggergaineventually. Andobviouslythisgainismorethantheinfluenceofthatpiece. Remember,thatthebettersituationgainedaftersacrificingapiecehasnotbeenachieved suddenlybuthasalreadybeenindicatedbytheSubpieces.

||

The Three repetition problem ||

Nowwhenanopponentpurposefullygoesforthreerepetitionsinvariablywemayhaveto goforthesecondbestfinaltotalchoiceandwhichmaynotbebeneficial. Sowehaveachoicetoselectitorsticktothebestandallowtheopponentforadraw. Itisnotmuchchallengingtopredictasituationwhereopponentcanforceathreerepetition butwhethertoimplementitornotistherealissueasitmightoverloadthecalculations. Ifthesuggestedbestmovebythetotalcalculationsforthenextmoveisequaltothe positionofpreviousmovethenthereisadangerofthreerepetitions. Thiswasallaboutknowingthesituation,nowaboutpredictingitbefore.

Herestillwearenotcalculatingtheprobabilitytreeforcheckfuturethreerepetition situations. Subpieceshelpuspredicttheincreaseanddecreaseofinfluencesandthusblocking, Castling,Checketc.buttocalculatethetotalsoftwosituationswherethereisapossibilityof repetition,firsttheyhavetobeinthesameinfluencevalueasforthosesituations. Therecouldbeprobablytwowaystodothis. 1) Lookingonestepahead. 2) Indicationbytwomaximumthreathouses. Lookingonestepaheadisnotactuallyprobabilitytreecreation. Whenthebesttotaliscalculatedforus,wecanthencalculatethebesttotalforthe opponentalsobysimilarcalculationsbutforhim(kindofplayinghissidealso). Andifthatopponentsnextstepmovecalculatedbyusshowsthepositionoftheprevious stateofhisthenwecansaythatitcouldbeasituationofthreerepetitions. Butbythismethodwecannotpredictastrategicbuiltupbyopponentforthreerepetition. Herethesecondtechniquemayhelp.(i.e.Indicationbytwomaximumthreathouses.) Hereweareprimarilyconsideringthatwiththehelpofthisentirealgorithmanditsstructure weareabletotargettheweakestpointofopponentorwearetryingtoplaythebest possiblemove.Thatmeanswearetryingtoplaythemovewhichwillcreatethemaximum possiblelossorminimumpossiblebenefitfortheopponent. Thusrepeatingmovesissupposedtobeinthebestinterestoftheopponent. Andifitisnotandstillopponenttriestorepeatthemovesthenasitisnotinhisbest interest,naturallyourcalculationswillalsonotnecessarilygetindulgeintorepeatingwith himandtakeanadvantagebytheotherwayaround. Sobasicallywhenopponentrepeatsthemovesofapiecemeansthattwosubpiecesofthat piecehaveclosevalues(asitcanreachbothofthembyadifferenceofonemove)andtotal gainlossthreatsinthesetwohousesaremorethananyothertotalsfortheopponent. Hereitisnotnecessarytoapplythesamealgorithmforcalculatingfinalgainlosstotalsfor opponentseparatelyasdoneforthefirstexampleasourfirsttwomajortotalgainsare obviouslyopponentsmajortotallossesandhavebeenalreadycalculatedinourtotals. Andthiscanbepredictablethroughsubpieces. Thuswiththisevenwecanpredictagradualbuiltupforthreerepetition.

||

The Steal Mate problem ||

Stealmateposesanotheruniqueandveryinterestingandoneofthetoughestchallenges. TillnowwehavebeentryingsohardtonaildowntheKingbyreducingitsmobilityand attackingitfromallpossiblesidesandways.Andthiswearedoingwiththehelpofsub piecesaroundtheKingbyincreasingtheirvalue.Soinfactinastealmatesituation,theKing isperfectlyblockedbysurroundingsubpiecesandisattacked(justnotdirectly)atitssecond worst. Butastealmateisadrawandthisisdefinitelyabenefitandevencouldbealifesaverfor theonewhoislosing. Sothisissomethingwhichshouldbepredictedtoavoiditfromoursideandshouldbe predictedsothatwecanforceitagainstopponentwhenrequired.Andagain,allthis withoutgoingforprobabilitytree. Predictingisnotmuchdifficultbutforcingitattherequiredtimeisverydifficultand tremendouslyriskyasitmightcreateblundersbymakinghilariousmistakes. Toavoidsuchmistakesperfectandsurepossiblepredictionthroughforcedmovesonlyis necessary. {Forcedmovesaretheonewhichtheopponenthastoplaywithoutanyotheroption.Any othermoveissimplyillegal.} Wecanworkintwoways 1) Puttingsomedelayfromstealmatesituationsoastoeitheravoidfromoursideortoforceif possible. 2) Similartowhatwehavedonesofar,wecanalreadysomehowpredictthebuildingofsteal matesituationwiththehelpofchangesoftheinfluencesofsubpiecesasitisbuildbutone moreconcernistheoverloadingofthecalculationswiththiswhichmightleadtoother blundersasthisconceptofstealmateitselfisagainstwhatwehavebeentryingtodowith theopponent. Toavoidthiswecanseparatelycalculateconditionsforstealmateandkeepthemseparate withoutmixingandinfluencinginregularcalculationsandconsiderthemincalculationsonly onceitstobeavoidedfromoursideoritispossibleandbeneficialtobeforced. Thisissomewhatsimilartotheconceptoflayeredthreatswherewehavetwoseparate arraysofcalculationsandrefertotheotheronlyinsomeparticularconditions.

InstealmatenotonlyKingbutotherpiecesarealsoimmobilizedifthereareanyandtheir numberisvarying.

||

Model Game || || Ending ||

S-ar putea să vă placă și