…
)”. So we need to include those WP block comments. Probably we need to wrap each paragraph in those comments. Also headings: maybeetc with comments. We must not include placeholders. Must be concise, each sentence adds value. End with promotion paragraph with link. We need to count words between 450-500 inclusive. Let’s aim for ~470 words. We need to include title line: “Title: …” then newline then HTML content. Title line not counted? Probably part of article but we need to count words of the article only? Safer to count everything after Title line? The instruction: “Start with the title as a plain heading (e.g., “Title: …\n\n”), then HTML content.” So we include Title: line then blank line then HTML. We’ll count words in the whole output? Probably they’ll count the article content (including title?). Safer to keep total words ~470 including title line. We need to include the e-book promotion paragraph at end. Let’s draft content. We need to talk about Mining for Gold: Identifying Feature Requests and Balance Issues, using AI automation. We need to use facts from e-book: core signals, examples, key phrases, scaling, separating novelty, surfacing silent majorities, define categories, examples quotes. We need to include Prompt Pattern for Balance Issue Detection and Prompt Pattern for Feature Request Mining (maybe just mention them). Not required to write the patterns explicitly? The facts list includes those headings. We can incorporate. Let’s craft. We’ll need to count words. Let’s write then count. Draft: Then blank line. Then HTML content. We’ll use WP block comments. Let’s write paragraphs. We’ll need to be concise. I’ll write and then count. Paragraph 1: Introduction. Paragraph 2: Core signals. Paragraph 3: Key phrases. Paragraph 4: Scaling perception. Paragraph 5: Separating novelty from need. Paragraph 6: Surfacing silent majorities. Paragraph 7: Defining categories. Paragraph 8: Examples of feature requests. Paragraph 9: Examples of balance issues. Paragraph 10: Prompt patterns. Paragraph 11: Workflow tip. Paragraph 12: Conclusion. Paragraph 13: Promotion. Now count words. Let’s write actual text. I’ll write without counting first, then count. Title line: “Title: Mining for Gold: Using AI to Spot Feature Requests and Balance Issues in Indie Games” Now blank line. Now HTML:
Indie developers drown in playtest feedback, yet hidden gems of insight sit untapped.
The first step is to recognize the two core signals that separate useful noise from actionable data.
One signal addresses the perceived fairness, effectiveness, or “feel” of an existing element – a balance issue.
The other signal expands the game’s systems, scope, or narrative – a feature request.
Key phrases such as “I wish…”, “It would be cool if…”, “You should add…”, “Can we have…?”, and “The game needs…” reliably flag these signals.
Spotting them manually works for a handful of comments, but an AI can read thousands in minutes, applying the same criteria every time.
This scaling lets you separate novelty from need, distinguishing a fun “wouldn’t it be neat” idea from a widely‑requested solution to a real friction point.
AI also surfaces silent majorities by finding patterns across Discord, forums, and surveys that no single human could correlate.
Before running any analysis, write your own game‑specific definitions for “Feature Request” and “Balance Issue” so the model knows exactly what to look for.
For example, a feature request appears in statements like “A map for the forest dungeon would be so helpful.” or “I wish I could re‑spec my skill points after level 10.”
Balance issues surface in comments such as “Grinding for leather takes too long; the drop rate feels bad.” or “The Frost Staff is useless compared to the Fireball.”
Other typical balance cues include “The final boss’s second phase is impossible without the rare potion.” and “Grinding for leather takes too long; the drop rate feels bad.” (already used) but we can add “The final boss’s second phase is impossible without the rare potion.”
We need to avoid duplication. Let’s adjust. We’ll produce distinct examples. Let’s rewrite. We need to ensure each sentence adds value. Now prompt patterns. We’ll include a short description. Paragraph for Prompt Pattern for Balance Issue Detection: maybe give a template. Paragraph for Prompt Pattern for Feature Request Mining. Then workflow tip. Then conclusion. Let’s write. We’ll need to count words now. Let’s write full content then count. I’ll write in a text editor mentally. Title line: Title: Mining for Gold: Using AI to Spot Feature Requests and Balance Issues in Indie Games Blank line. Now paragraphs. I’ll write each with the WP comment. Let’s go. — start counting manually. I’ll write sentences and count words. I’ll keep a running total. Initialize total words. We’ll count words in the title line as well? Probably yes. Let’s include. Title line: “Title: Mining for Gold: Using AI to Spot Feature Requests and Balance Issues in Indie Games” Words: Title:(1) Mining(2) for(3) Gold:(4) Using(5) AI(6) to(7) Spot(8) Feature(9) Requests(10) and(11) Balance(12) Issues(13) in(14) Indie(15) Games(16) So 16 words. Now blank line not counted. Now paragraph 1:Indie developers drown in playtest feedback, yet hidden gems of insight sit untapped.
Sentence: Indie(1) developers2 drown3 in4 playtest5 feedback,6 yet7 hidden8 gems9 of10 insight11 sit12 untapped13. 13 words. Running total: 16+13=29. Paragraph 2:The first step is to recognize the two core signals that separate useful noise from actionable data.
Words: The1 first2 step3 is4 to5 recognize6 the7 two8 core9 signals10 that11 separate12 useful13 noise14 from15 actionable16 data17. 17 words. Total: 46. Paragraph 3:One signal addresses the perceived fairness, effectiveness, or “feel” of an existing element – a balance issue.
Words: One1 signal2 addresses3 the4 perceived5 fairness,6 effectiveness,7 or8 “feel”9 of10 an11 existing12 element13 –14 a15 balance16 issue17. 17 words. Total: 63. Paragraph 4:The other signal expands the game’s systems, scope, or narrative – a feature request.
Words: The1 other2 signal3 expands4 the5 game’s6 systems,7 scope,8 or9 narrative10 –11 a12 feature13 request14. 14 words. Total: 77. Paragraph 5:Key phrases such as “I wish…”, “It would be cool if…”, “You should add…”, “Can we have…?”, and “The game needs…” reliably flag these signals.
Words: Key1 phrases2 such3 as4 “I5 wish…”,6 “It7 would8 be9 cool10 if…”,11 “You12 should13 add…”,14 “Can15 we16 have…?”,17 and18 “The19 game20 needs…”21 reliably22 flag23 these24 signals25. 25 words. Total: 102. Paragraph 6:Spotting them manually works for a handful of comments, but an AI can read thousands in minutes, applying the same criteria every time.
Words: Spotting1 them2 manually3 works4 for5 a6 handful7 of8 comments,9 but10 an11 AI12 can13 read14 thousands15 in16 minutes,17 applying18 the19 same20 criteria21 every22 time23. 23 words. Total: 125. Paragraph 7:This scaling lets you separate novelty from need, distinguishing a fun “wouldn’t it be neat” idea from a widely‑requested solution to a real friction point.
Words: This1 scaling2 lets3 you4 separate5 novelty6 from7 need,8 distinguishing9 a10 fun11 “wouldn’t12 it13 be14 neat”15 idea16 from17 a18 widely‑requested19 solution20 to21 a22 real23 friction24 point25. 25 words. Total: 150. Paragraph 8:AI also surfaces silent majorities by finding patterns across Discord, forums, and surveys that no single human could correlate.
Words: AI1 also2 surfaces3 silent4 majorities5 by6 finding7 patterns8 across9 Discord,10 forums,11 and12 surveys13 that14 no15 single16 human17 could18 correlate19. 19 words. Total: 169. Paragraph 9:Before running any analysis, write your own game‑specific definitions for “Feature Request” and “Balance Issue” so the model knows exactly what to look for.
Words: Before1 running2 any3 analysis,4 write5 your6 own7 game‑specific8 definitions9 for10 “Feature11 Request”12 and13 “Balance14 Issue”15 so16 the17 model18 knows19 exactly20 what21 to22 look23 for24. 24 words. Total: 193. Paragraph 10:For example, a feature request appears in statements like “A map for the forest dungeon would be so helpful.” or “I wish I could re‑spec my skill points after level 10.”
Words: For1 example,2 a3 feature4 request5 appears6 in7 statements8 like9 “A10 map11 for12 the13 forest14 dungeon15 would16 be17 so18 helpful.”19 or20 “I21 wish22 I23 could24 re‑spec25 my26 skill27 points28 after29 level30 10.”31 31 words. Total: 224. Paragraph 11:Balance issues surface in comments such as “Grinding for leather takes too long; the drop rate feels bad.” or “The Frost Staff is useless compared to the Fireball.”
Words: Balance1 issues2 surface3 in4 comments5 such6 as7 “Grinding8 for9 leather10 takes11 too12 long;13 the14 drop15 rate16 feels17 bad.”18 orFor a comprehensive guide with detailed workflows, templates, and additional strategies, see my e-book: AI for Indie Game Developers: How to Automate Game Design Document Updates and Bug Report Triage from Playtest Feedback.