Söl dner - Secr et War s

     A Post Mortem
         Teut Weidemann
     Managing Director, Wings
     Direcor of Design, Jowood
Overview
• History
• Examining Issues
  –   Design
  –   Experience
  –   Code
  –   QA
  –   Marketing
• What we learned
• Q&A
  – Wings Resource count
  – Söldner Sales
History: 2000-2004

Gamestar Test Quote: “When you die you loose
        all your hard earned money”

 Wrong: You always get a minimum of 3500
when you die, if you had more you even keep it.
History: Year 2000
•   Acquisition of Wings in Mai
•   Plans: Base Tech - Söldner - MMO
•   Pre Production, Planning, Staffing
•   Meanwhile:
    –   Counseling: acquisition of Massive
    –   Counseling: acquisition of Products (Spellforce)
    –   Online Presence, Websites
    –   Web/Ftp Server setup & management
    –   Teamgrowth: +4 Artists (team 7 total)
History: 2001
• Team Growth:
  – Programmers: +4
  – Artists: +1
  – Admin: +3, team total: 15
• Preproduction Base, Design
• Interim:
  –   Panzer Elite Spec. Edition
  –   Community Management: Jowood forums
  –   Managing Online Games (WW3, Cultures2 etc.)
  –   Being Development Director of Jowood
History: 2002
• Jowood Crises:
   – Cash flow limits to Salaries only, if at all
   – 25% of employees have to go, incl. Co-founder
• Team Change:
   – Programmers: -3
   – Admin: -2
• Living with the crisis: month to month worries
   – Too many tasks on some resources
   – Planning failsafe: looking for alternatives
   – Effectivity drops by 30%
History: 2003
•   JoWooD: surviving! Replanning, Cash still tight
•   Second growth: ordering process, planned 12/2003 release
•   Oops #1: QA? -> Build your own
•   Oops #2: no Marketing (Spellforce took it all)
    – We had to do Online Marketing
    – We had to do Offline Marketing
    – Cover and Packaging
• Oops #3: We had to do Worldwide Serverdeals
    – Not task of Partners (BigBen or Encore)
• Oops #4: Release of public beta bombed
• 16. December 2004: delivery of beta candidate
History: 2004
• QA starts slow
• Soon being discovered: we need more time
   – > rescheduled March 2004
• Delivering Master Candidate each month
• BigBen insists on 2 more polishing months
   – > Rescheduled May 2004
• Release end of May/04 (EU)
• Release end of June/04 (US)
• Since then monthly updates as planned
Launch
•   We were caught unaware how big the hype was
•   Bad reviews
•   Bugs
•   Problems we never seen before in testing
•   Angry customers
•   Sales ok in GAS, bad in EU/US

• But WHY?
Examining Design

Gamestar quote: It takes forever until you have
enough money to afford a vehicle for the huge
                     maps

Wrong: with $3500 startmoney you can always
    afford a Lada ($800) and equipment
Design
• Dropping weapons: it’s so much easier not to
• ADS, changing environments
   – Lightning the scene is hard, no precalculated lightmaps
   – No static mesh in 3d cards (speed issues)
• One world design
   – no level load/reset means errors add up on time (and
     servers run days/weeks)
   – all real time, no pause (even in menu’s)
   – Remember: this tech was designed to be MMO
• No experience in Avatar games means cyclic
  approach to design
Design: good
• Overall design worked: people believed in
  the product, expectations high
• Generating content: settlement, landscape
  and foilage generation works
• Combined Arms works
• ADS, UCS, unique player names, no
  preselected classes: work together well
Design: Bad
• Designer overloaded since 2002: Producer,
  Designer, Managing Director, Marketing
  Support, Mastering
• Design not complete at stages due to this
• Cyclic design and re-evaluation took too
  long: too many cycles
• This led to unnecessary code rewrites
• Wanting too much: (see next slide)
Single Player             Components
 Multiplayer             Soldiers       UCS         Portal

Client:                 Wheeled       SWS: 70    Unique Name
3d Renderer
Physics
AI                        Tanks       AFV: 60     Serverlists
Netcode
Particle/FX
Scriptsystem           Helicopters     TCM           Chat
Overlay/GUI
Terrain
Landscape Generators
Scenatio Generator
                           Jets       AGS: 200     Patcher
Debug tools, etc.

                          Server        CCP       CD Keys

                       Linux Server    CMM        Messenger
Experience & Complexity

   GameStar: “... In game xyz you had intuitive
   controls like “wasd” etc. but in Söldner ...”

Comment: Söldner has the identical interface, it has
        been adapted to “the” standard
Experience: Avatars
    • Animation & Avatars: we rewrote soldier code 5 times. Still
      not right.


   Zur Anzeige wird der QuickTime™             Zur Anzeige wird der QuickTime™             Zur Anzeige wird der QuickTime™             Zur Anzeige wird der QuickTime™             Zur Anzeige wird der QuickTime™
Dekompressor „TIFF (Unkomprimiert)“         Dekompressor „TIFF (Unkomprimiert)“         Dekompressor „TIFF (Unkomprimiert)“         Dekompressor „TIFF (Unkomprimiert)“         Dekompressor „TIFF (Unkomprimiert)“
                benötigt.                                   benötigt.                                   benötigt.                                   benötigt.                                   benötigt.




                          Zur Anzeige wird der QuickTime™             Zur Anzeige wird der QuickTime™             Zur Anzeige wird der QuickTime™               Zur Anzeige wird der QuickTime™
                       Dekompressor „TIFF (Unkomprimiert)“         Dekompressor „TIFF (Unkomprimiert)“         Dekompressor „TIFF (Unkomprimiert)“           Dekompressor „TIFF (Unkomprimiert)“
                                       benötigt.                                   benötigt.                                   benötigt.                                     benötigt.
Experience: Multiplayer
• Multiplayer: no experience meant suboptimal netcode and
  side effects
   – Designing async game flows is a pain
   – Synced databases are the future, but why did we do our own?
• MP and changing environments: some things no longer
  work like “Bomb-Run”
• Server as master meant sacrificing speed and efficiency:
  server runs 32 clients (theoretically)
• Always developing 2 games: SP and MP
   – Coders always need 2 comps (server - client)
   – Many things need to be load tested (16 players min)
Experience: Physics
• Physics: meant adjusting whole project to it
• Taiming physic engines is an artform by itself
• Collision code was finally working when we
  enslaved it by the physic system
• We used ODE (open source physics engine) which
  helped a lot
Experience: Portal
• Portal: no GameSpy, meant an additional whole project:
   –   Login/Logout
   –   Unique Name handling
   –   Server Lists
   –   Load Balancing
   –   Messenger
   –   Auto Patch System
   –   Cd Key Handling
• You ask why? GameSpy makes business with YOUR
  game and YOUR customer ... And charges YOU for it!
   – No secure cd keys
   – Endusers don’t like it
   – Many features we needed not available in 2001/2002
Code

GameStar: “The game has serious problems with
 water: any vehicle touching water explodes”

Comment: Your beloved Bf1942 does the same
   and you weren’t bitching at all about it
Code: Good
• Open Source Physics were painless more or less
• Landscape Renderer: streets, dynamic texture
  generation, 25km View distance
• Auto Generation of content: Terrain, Roads,
  Foilage, Settlements saved thousands of level
  editing manhours
• ADS: simply worked
• Data driven design: Integration without coders
• Re-iterating 3d Tech (although FarCry hurt)
• Late Optimization: 12fps to 50+ fps in 3 month
• Localisation Technology: mastering with no code
Code: Bad
• Missing code reviews in 2001/2002 nearly killed the
  project (-> lead was fired)
• AI: late design, monster of state machine
• Tools: not enough (balancing 130 game units)
• Lack of performance sensivity/control
• Lack of consistent preplanning and structuring
• Applied standards weren’t strict enough: bad resources
  spoil code quality and stability
• Submit discipline -> stable builds (much better in 2003)
QA
Gamestar: “.. If they now get their
bugfixes done and balance a little
             then ...”
      Petra: “IF! If ever ...”

Statement: Professional Journalism?
QA: Good
• Internal QA increased productivity, short
  feedback loop, build discipline
• Automating builds: from 42 down to 12
  minutes
• Closed beta: feedback, organisation
QA: Bad
• External QA not up to the task:
   – Load tests: 16 testers minimum needed
   – Testplans: too many things to test
   – They were functionality control, not quality control
• Beta
   –   Hosting at foreign providers: not good
   –   No dedicated beta manager: not good
   –   Automation feedback of logs was missing
   –   Only 10% give proper feedback, only 1% are perfect
Marketing
   Florian Stangl (ex-Gamestar):
     “The vehicles don’t work”

 Comment: So how else could you
complain about the Vehicle Physics?
Marketing: Good
•   Our own Online Hype exceeded expectations
•   Marketing Online Games Online
•   Community Management before Release
•   Getting Game and Fan-Sites into Beta
•   Cover
•   Packaging
•   PR (germany only)
Marketing: Bad
• Up to 12/2003 we were doing it ourselves
   – So Marketing started late: requests during crunch added load
     on team
• Germany: perfect PR
   – France: horrible
   – UK: arrogant, incompetent
   – US: lets not speak about it
• 1st MP Game of Jowood: lack of customer knowledge
   – No real online campaign despite Wings concepts & plans
   – No understanding of internet marketing as a tool
What we learned
  Gamestar, Commenting on bugs they
copy/paste into their review from our tech
             support forum:
“I don’t want to repeat what’s being said
 in the forums, we are building our own
           opinion after all ...”

           Comment: none ...
What we learned
• Code pre planning for all modules
• Code reviews for all modules
• Optimize early: implement automatic performance
  tracking
• Without Perforce we wouldn’t be here
• Automate build process early: saves hours of time
• Incorporate debug output early and automate
  feedback: makes reproducing much easier
• Automating localization works: export, installer,
  done
What we learned
• Generating content works ... To a limit
• Be sensible to resources leaving
  – Restructure
  – Transfer orderly
• Be honest: no experience in one field: hire
  contractor for counsel
• US Server providers suck
What we learned
• Multiplayer Games have their own rules
  – Linux Server is a must
  – Server must not use DirectX
     • And many other specs we won’t reveal here
  – Clan Support is #1 priority
     • When Clans don’t play your game noone will
  – Server Provider Support is a necessity
     • Their demands differ from your view
     • Understand their business and you win
Top Reasons Söldner is like it is
• Code planning/structuring: should have
  acted sooner
• JoWooD Crisis and it’s consequences
• Lack of proper QA for MP games
• Lack of resources in critical time
• Wanting too much
Q&A
Gamestar: “money has nothing to do with
              reviews”

             Comment: LOL!
   (GameStar canceled a Söldner preview after
      they learned from PC Games deal)
10
                                            15
                                                                  20
                                                                                         25




                    0
                        5
            may
             jun
              jul
            aug
            sep
             oct
            nov




Year 2000
            dec
             jan
             feb
            mar
            apr
            may
             jun
              jul
            aug
            sep




Year 2001
             oct
            nov
            dec
             jan
             feb
            mar
            apr
            may
             jun
              jul
            aug
            sep




Year 2002
             oct
            nov
            dec
             jan
             feb
            mar
            apr
                                                                       Wings Resources




            may
             jun
              jul
            aug
            sep
Year 2003
             oct
            nov
            dec
             jan
             feb
            mar
            apr
            may
             jun
              jul
Year 2004




            aug
                                   Misc
                                                           Prog




                            Sum
                                                  Artist
                                          Admin
Söldner Sales: Deutschland
14000


12000           11974


10000

                        8372
 8000
                                                                                Sales
 6000
                               5453

 4000                                 3782

                                             2596 2404
 2000
                                                         1069 880
                                                                    731   727
   0       6
        kw21 kw22 kw23 kw24 kw25 kw26 kw27 kw28 kw29 kw30 kw31
Söldner Facts
• Measured in week 22-31, Coverage 82%:
  44.844 sold units in Germany
• For weeks 22-31 Söldner is second best
  selling title in Germany (#1: Panzers)
• Söldner spend 2 months in the top 10
• Info: Joint Operations sold 13227 (D)
• Info: Joint Operations sold 27800 (US)
Stückzal Vollpreis
                38
                     20




                                            10.000
                                                           15.000
                                                                    20.000
                                                                             25.000




                                   5.000




                               0
                          02
                40
                     20
                          02
                42
                     20
                          02
                44
                     20
                          02
                46
                     20
                          02
                48
                     20
                          02
                50
                     20
                          02
                52
                     20
                      02
                 2
                   20
                      03
                 4
                   20
                      03
                 6
                   20
                      03
                 8
                   20
                      0
                10 3
                   20
                      0
                12 3
                   20
                      0
                14 3
                   20
                      0
                16 3
                   20
                      0
                18 3
                   20
                      0
                20 3




Kalenderwoche
                   20
                      0
                22 3
                   20
                      0
                24 3
                   20
                      0
                26 3
                   20
                      0
                                                                                      Battlefield 1942 - Verkaufsentwicklung seit Release




                28 3
                   20
                      0
                30 3
                   20
                      0
                32 3
                   20
                      0
                34 3
                   20
                      0
                36 3
                   20
                      0
                38 3
                   20
                      0
                40 3
                   20
                      0
                42 3
                   20
                      0
                44 3
                   20
                      03
Misc Facts
• 3200 Bugzilla entries     • 1544 xml files
                            • 717 textures
   (08/2002-05/2004)
                            • 651 Animation files
• 18054 Perforce submits    • 722 x-files (Objects)
• 656 test builds           • 1991 wav files
• Full master build: 20     • 18 mio square km terrain
  minutes                     generated
                            • 225 villages generated
• Master from submit to     • 775.000 lines of code
  CD: 1 hour                • Languages: 10,
• Patch from submit to        De,Fr,Sp,It,Uk,Us,Jp,Cz,Hu,Ru
  exe: 30 minutes
100
                                 200
                                                   300
                                                         400
                                                               500




-100
               0
       Ja
          n
       Fe
          b
       M
         ar
       Ap
          r
       M
        ay
       Ju
          n
        Ju
           l
       Au
          g
       Se
          p
       O
         ct
       N
        ov
       D
         ec
       Ja
          n
       Fe
          b
       M
         ar
       Ap
          r
       M
        ay
       Ju
          n
        Ju
           l
       Au
          g
                                                                     Söldner Task/Bug Hunt*




                         Delta
                                 Fixed
                                         Created
Tools used
•   Visual .Net          •   3D Studio Max 4/5
•   Perforce
                         •   Character Studio
•   Incredibuild
•   Visual Assist        •   Photoshop
•   Granny               •   Kaldera
•   ODE
•   Miles
•   zLib
•   Python
•   DoxyGen
•   BugZilla
Thanx very much

    cu online

2004: Söldner - a Post Mortem

  • 1.
    Söl dner -Secr et War s A Post Mortem Teut Weidemann Managing Director, Wings Direcor of Design, Jowood
  • 2.
    Overview • History • ExaminingIssues – Design – Experience – Code – QA – Marketing • What we learned • Q&A – Wings Resource count – Söldner Sales
  • 3.
    History: 2000-2004 Gamestar TestQuote: “When you die you loose all your hard earned money” Wrong: You always get a minimum of 3500 when you die, if you had more you even keep it.
  • 4.
    History: Year 2000 • Acquisition of Wings in Mai • Plans: Base Tech - Söldner - MMO • Pre Production, Planning, Staffing • Meanwhile: – Counseling: acquisition of Massive – Counseling: acquisition of Products (Spellforce) – Online Presence, Websites – Web/Ftp Server setup & management – Teamgrowth: +4 Artists (team 7 total)
  • 5.
    History: 2001 • TeamGrowth: – Programmers: +4 – Artists: +1 – Admin: +3, team total: 15 • Preproduction Base, Design • Interim: – Panzer Elite Spec. Edition – Community Management: Jowood forums – Managing Online Games (WW3, Cultures2 etc.) – Being Development Director of Jowood
  • 6.
    History: 2002 • JowoodCrises: – Cash flow limits to Salaries only, if at all – 25% of employees have to go, incl. Co-founder • Team Change: – Programmers: -3 – Admin: -2 • Living with the crisis: month to month worries – Too many tasks on some resources – Planning failsafe: looking for alternatives – Effectivity drops by 30%
  • 7.
    History: 2003 • JoWooD: surviving! Replanning, Cash still tight • Second growth: ordering process, planned 12/2003 release • Oops #1: QA? -> Build your own • Oops #2: no Marketing (Spellforce took it all) – We had to do Online Marketing – We had to do Offline Marketing – Cover and Packaging • Oops #3: We had to do Worldwide Serverdeals – Not task of Partners (BigBen or Encore) • Oops #4: Release of public beta bombed • 16. December 2004: delivery of beta candidate
  • 8.
    History: 2004 • QAstarts slow • Soon being discovered: we need more time – > rescheduled March 2004 • Delivering Master Candidate each month • BigBen insists on 2 more polishing months – > Rescheduled May 2004 • Release end of May/04 (EU) • Release end of June/04 (US) • Since then monthly updates as planned
  • 9.
    Launch • We were caught unaware how big the hype was • Bad reviews • Bugs • Problems we never seen before in testing • Angry customers • Sales ok in GAS, bad in EU/US • But WHY?
  • 10.
    Examining Design Gamestar quote:It takes forever until you have enough money to afford a vehicle for the huge maps Wrong: with $3500 startmoney you can always afford a Lada ($800) and equipment
  • 11.
    Design • Dropping weapons:it’s so much easier not to • ADS, changing environments – Lightning the scene is hard, no precalculated lightmaps – No static mesh in 3d cards (speed issues) • One world design – no level load/reset means errors add up on time (and servers run days/weeks) – all real time, no pause (even in menu’s) – Remember: this tech was designed to be MMO • No experience in Avatar games means cyclic approach to design
  • 12.
    Design: good • Overalldesign worked: people believed in the product, expectations high • Generating content: settlement, landscape and foilage generation works • Combined Arms works • ADS, UCS, unique player names, no preselected classes: work together well
  • 13.
    Design: Bad • Designeroverloaded since 2002: Producer, Designer, Managing Director, Marketing Support, Mastering • Design not complete at stages due to this • Cyclic design and re-evaluation took too long: too many cycles • This led to unnecessary code rewrites • Wanting too much: (see next slide)
  • 14.
    Single Player Components Multiplayer Soldiers UCS Portal Client: Wheeled SWS: 70 Unique Name 3d Renderer Physics AI Tanks AFV: 60 Serverlists Netcode Particle/FX Scriptsystem Helicopters TCM Chat Overlay/GUI Terrain Landscape Generators Scenatio Generator Jets AGS: 200 Patcher Debug tools, etc. Server CCP CD Keys Linux Server CMM Messenger
  • 15.
    Experience & Complexity GameStar: “... In game xyz you had intuitive controls like “wasd” etc. but in Söldner ...” Comment: Söldner has the identical interface, it has been adapted to “the” standard
  • 16.
    Experience: Avatars • Animation & Avatars: we rewrote soldier code 5 times. Still not right. Zur Anzeige wird der QuickTime™ Zur Anzeige wird der QuickTime™ Zur Anzeige wird der QuickTime™ Zur Anzeige wird der QuickTime™ Zur Anzeige wird der QuickTime™ Dekompressor „TIFF (Unkomprimiert)“ Dekompressor „TIFF (Unkomprimiert)“ Dekompressor „TIFF (Unkomprimiert)“ Dekompressor „TIFF (Unkomprimiert)“ Dekompressor „TIFF (Unkomprimiert)“ benötigt. benötigt. benötigt. benötigt. benötigt. Zur Anzeige wird der QuickTime™ Zur Anzeige wird der QuickTime™ Zur Anzeige wird der QuickTime™ Zur Anzeige wird der QuickTime™ Dekompressor „TIFF (Unkomprimiert)“ Dekompressor „TIFF (Unkomprimiert)“ Dekompressor „TIFF (Unkomprimiert)“ Dekompressor „TIFF (Unkomprimiert)“ benötigt. benötigt. benötigt. benötigt.
  • 17.
    Experience: Multiplayer • Multiplayer:no experience meant suboptimal netcode and side effects – Designing async game flows is a pain – Synced databases are the future, but why did we do our own? • MP and changing environments: some things no longer work like “Bomb-Run” • Server as master meant sacrificing speed and efficiency: server runs 32 clients (theoretically) • Always developing 2 games: SP and MP – Coders always need 2 comps (server - client) – Many things need to be load tested (16 players min)
  • 18.
    Experience: Physics • Physics:meant adjusting whole project to it • Taiming physic engines is an artform by itself • Collision code was finally working when we enslaved it by the physic system • We used ODE (open source physics engine) which helped a lot
  • 19.
    Experience: Portal • Portal:no GameSpy, meant an additional whole project: – Login/Logout – Unique Name handling – Server Lists – Load Balancing – Messenger – Auto Patch System – Cd Key Handling • You ask why? GameSpy makes business with YOUR game and YOUR customer ... And charges YOU for it! – No secure cd keys – Endusers don’t like it – Many features we needed not available in 2001/2002
  • 20.
    Code GameStar: “The gamehas serious problems with water: any vehicle touching water explodes” Comment: Your beloved Bf1942 does the same and you weren’t bitching at all about it
  • 21.
    Code: Good • OpenSource Physics were painless more or less • Landscape Renderer: streets, dynamic texture generation, 25km View distance • Auto Generation of content: Terrain, Roads, Foilage, Settlements saved thousands of level editing manhours • ADS: simply worked • Data driven design: Integration without coders • Re-iterating 3d Tech (although FarCry hurt) • Late Optimization: 12fps to 50+ fps in 3 month • Localisation Technology: mastering with no code
  • 22.
    Code: Bad • Missingcode reviews in 2001/2002 nearly killed the project (-> lead was fired) • AI: late design, monster of state machine • Tools: not enough (balancing 130 game units) • Lack of performance sensivity/control • Lack of consistent preplanning and structuring • Applied standards weren’t strict enough: bad resources spoil code quality and stability • Submit discipline -> stable builds (much better in 2003)
  • 23.
    QA Gamestar: “.. Ifthey now get their bugfixes done and balance a little then ...” Petra: “IF! If ever ...” Statement: Professional Journalism?
  • 24.
    QA: Good • InternalQA increased productivity, short feedback loop, build discipline • Automating builds: from 42 down to 12 minutes • Closed beta: feedback, organisation
  • 25.
    QA: Bad • ExternalQA not up to the task: – Load tests: 16 testers minimum needed – Testplans: too many things to test – They were functionality control, not quality control • Beta – Hosting at foreign providers: not good – No dedicated beta manager: not good – Automation feedback of logs was missing – Only 10% give proper feedback, only 1% are perfect
  • 26.
    Marketing Florian Stangl (ex-Gamestar): “The vehicles don’t work” Comment: So how else could you complain about the Vehicle Physics?
  • 27.
    Marketing: Good • Our own Online Hype exceeded expectations • Marketing Online Games Online • Community Management before Release • Getting Game and Fan-Sites into Beta • Cover • Packaging • PR (germany only)
  • 28.
    Marketing: Bad • Upto 12/2003 we were doing it ourselves – So Marketing started late: requests during crunch added load on team • Germany: perfect PR – France: horrible – UK: arrogant, incompetent – US: lets not speak about it • 1st MP Game of Jowood: lack of customer knowledge – No real online campaign despite Wings concepts & plans – No understanding of internet marketing as a tool
  • 29.
    What we learned Gamestar, Commenting on bugs they copy/paste into their review from our tech support forum: “I don’t want to repeat what’s being said in the forums, we are building our own opinion after all ...” Comment: none ...
  • 30.
    What we learned •Code pre planning for all modules • Code reviews for all modules • Optimize early: implement automatic performance tracking • Without Perforce we wouldn’t be here • Automate build process early: saves hours of time • Incorporate debug output early and automate feedback: makes reproducing much easier • Automating localization works: export, installer, done
  • 31.
    What we learned •Generating content works ... To a limit • Be sensible to resources leaving – Restructure – Transfer orderly • Be honest: no experience in one field: hire contractor for counsel • US Server providers suck
  • 32.
    What we learned •Multiplayer Games have their own rules – Linux Server is a must – Server must not use DirectX • And many other specs we won’t reveal here – Clan Support is #1 priority • When Clans don’t play your game noone will – Server Provider Support is a necessity • Their demands differ from your view • Understand their business and you win
  • 33.
    Top Reasons Söldneris like it is • Code planning/structuring: should have acted sooner • JoWooD Crisis and it’s consequences • Lack of proper QA for MP games • Lack of resources in critical time • Wanting too much
  • 34.
    Q&A Gamestar: “money hasnothing to do with reviews” Comment: LOL! (GameStar canceled a Söldner preview after they learned from PC Games deal)
  • 35.
    10 15 20 25 0 5 may jun jul aug sep oct nov Year 2000 dec jan feb mar apr may jun jul aug sep Year 2001 oct nov dec jan feb mar apr may jun jul aug sep Year 2002 oct nov dec jan feb mar apr Wings Resources may jun jul aug sep Year 2003 oct nov dec jan feb mar apr may jun jul Year 2004 aug Misc Prog Sum Artist Admin
  • 36.
    Söldner Sales: Deutschland 14000 12000 11974 10000 8372 8000 Sales 6000 5453 4000 3782 2596 2404 2000 1069 880 731 727 0 6 kw21 kw22 kw23 kw24 kw25 kw26 kw27 kw28 kw29 kw30 kw31
  • 37.
    Söldner Facts • Measuredin week 22-31, Coverage 82%: 44.844 sold units in Germany • For weeks 22-31 Söldner is second best selling title in Germany (#1: Panzers) • Söldner spend 2 months in the top 10 • Info: Joint Operations sold 13227 (D) • Info: Joint Operations sold 27800 (US)
  • 38.
    Stückzal Vollpreis 38 20 10.000 15.000 20.000 25.000 5.000 0 02 40 20 02 42 20 02 44 20 02 46 20 02 48 20 02 50 20 02 52 20 02 2 20 03 4 20 03 6 20 03 8 20 0 10 3 20 0 12 3 20 0 14 3 20 0 16 3 20 0 18 3 20 0 20 3 Kalenderwoche 20 0 22 3 20 0 24 3 20 0 26 3 20 0 Battlefield 1942 - Verkaufsentwicklung seit Release 28 3 20 0 30 3 20 0 32 3 20 0 34 3 20 0 36 3 20 0 38 3 20 0 40 3 20 0 42 3 20 0 44 3 20 03
  • 39.
    Misc Facts • 3200Bugzilla entries • 1544 xml files • 717 textures (08/2002-05/2004) • 651 Animation files • 18054 Perforce submits • 722 x-files (Objects) • 656 test builds • 1991 wav files • Full master build: 20 • 18 mio square km terrain minutes generated • 225 villages generated • Master from submit to • 775.000 lines of code CD: 1 hour • Languages: 10, • Patch from submit to De,Fr,Sp,It,Uk,Us,Jp,Cz,Hu,Ru exe: 30 minutes
  • 40.
    100 200 300 400 500 -100 0 Ja n Fe b M ar Ap r M ay Ju n Ju l Au g Se p O ct N ov D ec Ja n Fe b M ar Ap r M ay Ju n Ju l Au g Söldner Task/Bug Hunt* Delta Fixed Created
  • 42.
    Tools used • Visual .Net • 3D Studio Max 4/5 • Perforce • Character Studio • Incredibuild • Visual Assist • Photoshop • Granny • Kaldera • ODE • Miles • zLib • Python • DoxyGen • BugZilla
  • 43.