Editing Topic

{{Quality|Unrated}}

{{av}}

There are around 300 innovations that can be discovered. Many of them have requirements, and many do not, so it's kind of a tech forest/grassland rather than a tree. Notably missing are innovations related to practical labors in the game (which will be added later in development). Only the discovery of [[Book#Literary_Forms|literary form]] topics has any effect on gameplay, as it allows scholars to write new types of books.

For unskilled roles, some initial back-of-the-envelope calculations and some rough estimates observed through dfHack gives a basic idea of just how long research can take.  Note that player testing had found that the skill roll contributes a lot to the breakthrough credit.  Even dabbling dwarves can gain over 2000 breakthrough credit at once and may average around 1000 breakthrough credit.  Additionally the formula for the skill roll seems to have a very large spread, players have reported as low as 3000 breakthrough credit and as high as 25000 breakthrough credit awarded to a dwarf with legendary skill.  Analytically calculating from the algorithm Toady gave, it takes an average of 58.54 cycles to get a breakthrough credit. Player testing has confirmed breakthrough credits typically take 55-65 cycles, so this calculation seems plausible.  Likewise, player testing has not shown any correlation between skill level and number of cycles to get breakthrough credit.  An unskilled dwarf might average around 120+ cycles to get their first breakthrough die for a level 1 topic. With only a single research die,  it would take an average of 2528 cycles to get a breakthrough. Pondering takes slightly over 2 days per cycle, so this would be around 14 years.  In a real fortress, the research dice and skill level will be increasing during the process.  However, in a real fortress, the scholars will also need to take breaks to eat, drink, and meet other needs. A more reasonable estimate is around 5-10 years for a breakthrough starting with no skills.  In short, research takes an unreasonably long time with an unskilled dwarf.

To optimize for research, there are several ideas worth considering.  Discussion is faster, taking only a single day, but requires at least 2 dwarves.  [http://www.bay12forums.com/smf/index.php?topic=174850.msg8061426#msg8061426 Player testing] has found discussion also awards more XP, (0-5 for pondering vs. 15-30 for discussion).  So to optimize research, dwarves should be separated by topic/skill, with unskilled dwarves paired with skilled dwarves.  This way, discussions will be equally divided between increasing the skilled dwarf's cycle count and the unskilled dwarf's cycle count. Only half the cycles will go to the skilled dwarf's cycle count, but discussion only takes a single day, so this is at net the same speed for the skilled dwarf with the additional bonus of the unskilled dwarf's cycle count also being increased.  Both dwarves benefit from the increased XP gain.

Note that several topics benefit from skills that can be trained through faster means.  Diagnostician and Wound Dresser can be trained through intentionally causing injuries or simply engaging in dangerous activities that require regular dwarf healthcare.  Organizer can be trained through assigning make work to the manager, or Mechanic can be trained by spamming the creation of mechanisms.  Skill level will increase much faster through these activities than through pondering or discussion.  Thus, to optimize for producing breakthroughs in engineering topics, mechanics could first be quickly trained to a legendary within a year or two at only the cost of the stones for the mechanisms, and then assigned as a scholar.

[http://www.bay12forums.com/smf/index.php?topic=174850.msg8061426#msg8061426 Player testing] has also found that [[visitors]] can come with research credits already done, thus, for forts aiming to maximize research, it makes sense to prioritize attracting wandering scholars.