Onboarding wizard: simplify the reference-database step to three options (none / free-managed / bring-your-own)

[jozef2svrcek]

Onboarding wizard: collapse the reference-database step to three options

Improvement of the existing onboarding wizard (#98, #124). The deep-history source analysis (#142) and cut-off analysis (#148) established sensible optimal defaults — so the wizard should encode those defaults rather than ask the user to reason about sources, windows, and FIDE coverage. A new user doesn't know TWIC from Lichess from Ajedrez; the app has now done that analysis, so the choice should be reduced to intent, not mechanics.

Proposed: exactly three options

1. No reference database — just my own games. No deep-history base, no reference auto-updates.
2. Free reference database, kept up to date (recommended default) — ships the public-domain deep-history base (Ajedrez) + live updates (TWIC, and Lichess Broadcasts if it earns a default-on slot — pending the overlap analysis in Determine the optimal TWIC cut-off date once the deep-history source is chosen #148), with the cut-off date set optimally under the hood (Ajedrez ≤ 2012, TWIC ≥ 2013 per Determine the optimal TWIC cut-off date once the deep-history source is chosen #148). Zero source/window decisions surfaced to the user.
3. Bring your own (e.g. Megabase) + updates — import a commercial database the user owns as the deep-history base (converted locally), plus live updates. Cut-off date TBD — depends on the imported DB's coverage/quality; determined at import (its own follow-up).

Why

• The prior wizard exposed source/window/commercial choices (Multi-source Phase C2: onboarding wizard — populate-vs-empty + deep-history choice #98/Onboarding UX: "I own a commercial database" path has wrong defaults (My games / Bundesliga) — offer Megabase/Bigbase instead #124) that require exactly the analysis we just did to answer. Users can't make that call; defaults should.
• Three intent-level options (nothing / free-managed / bring-your-own) map cleanly to real user situations and avoid decision paralysis.
• The "optimal cut-off" and "which updates default-on" are baked into option 2, not asked.

Depends on / relates to

• Choose the deep-history reference source(s) for LPDO (FIDE coverage, licenses) #142 — deep-history source selection (Ajedrez default + BYO-Megabase).
• Determine the optimal TWIC cut-off date once the deep-history source is chosen #148 — cut-off date (Ajedrez↔TWIC handoff ~2013) + whether Lichess Broadcasts is default-on (pending full TWIC+Lichess overlap data).
• Multi-source Phase C2: onboarding wizard — populate-vs-empty + deep-history choice #98 — wizard populate-vs-empty / deep-history choice (the step being simplified).
• Onboarding UX: "I own a commercial database" path has wrong defaults (My games / Bundesliga) — offer Megabase/Bigbase instead #124 — "I own a commercial database" path → becomes option 3.

Notes

• Option 2 must apply the complementary windows automatically (Ajedrez --to 2012-12-31, TWIC --from 2013-01-01).
• Option 3's cut-off is a separate determination per imported DB; can start with "no window" and refine.
• Keep an "advanced / Maintenance → Sources" escape hatch for power users to override, so simplification doesn't remove capability.

[jozef2svrcek]

Surface the expected update volume in the wizard

For the "free, kept up to date" option, show the user concretely what "kept up to date" means — the expected new games per month (measured, #148):

• TWIC only: ~28,500 / month
• Lichess Broadcasts only: ~29,400 / month
• TWIC + Lichess (default): ~49,400 / month (Lichess adds ~+20,900 unique on top of TWIC)

Displaying this in the option's description helps a new user understand the trade-off (and why both feeds are on by default) without needing to know what TWIC or Lichess are. Same idea applies to option 3 (BYO) once its update set is chosen. Numbers should ideally be computed/refreshed rather than hard-coded, since they drift with the tournament calendar.

[jozef2svrcek]

Closing — not needed. The wizard's History step already presents the three intended options (Free historical base / I own a commercial database / None), so the reference-database step is already simple enough in practice. The real remaining value in this area is the TWIC cut-off detection (#148), not further restructuring of this step. Reopen if the step proves confusing in use.