Basketball Ideas
Reading the almanac

Methodology & Sources

Start with the question each tool answers. Open a section for the rules, evidence, and limits behind its results. A historical comparison is a way to explore the record, not a prediction.

Trades

Trade Tracker follows an anchor transaction from one franchise’s perspective. The Archive finds transactions across the published record.

Time and lineage

The original cost stays fixed. Return follows verified descendants through the chosen checkpoint. Later Exits show where realized return moved on. A trade-day snapshot and a later snapshot answer different questions.

Changing downstream scope changes the displayed snapshot. It does not rewrite canonical history. Picks, protections, swaps, unresolved outcomes, and unavailable statistics keep their explicit status in the asset ledger.

Evidence and player context

Open an asset or transaction in Trade Tracker for its provenance and linked evidence. Those record-level references are the source for a particular claim; a general source list is not a substitute.

Player season statistics are full-season values, which may include games after a trade checkpoint. Season age is measured on February 1; arrival and departure ages use recorded movement dates. — means unavailable, not zero. NBA/BAA regular-season statistics run from 1946–47 through 2025–26.

Search the Archive → · Player statistics →

Trade Recall · Daily Game

Trade Recall uses real trades from the same published, verified transaction records as the Trade Archive. The launch collection contains 66 trades and 131 eligible receiving sides, spanning 1970–2023.

What is hidden

Exactly one player in the entire trade is hidden. Every other player, including those received by the same team, remains visible. One correct answer completes the trade and unlocks the next one. First-round picks, second-round picks, and swaps are shown as separate counts. Their years, protections, and eventual selections are not quiz answers. Cash is omitted. Deals involving draft-rights assets, unclear identities, unsupported pick rounds, or a side that becomes empty after removing cash are excluded from this launch collection.

A challenge for different kinds of fans

Difficulty belongs to the missing player, with the visible trade as context. Opening rounds hide recognizable headline players. Later rounds can hide a supporting player or test an older deal, with familiar questions mixed in. Targets are selected explicitly for each receiving side; a trade is not rated difficult merely because it once required several names.

A run starts with two warm-ups, then mixes familiar trades with deeper cuts and occasional easier rounds. The sequence rotates teams and eras when possible. The same transaction cannot appear twice in one run, including from the other side, and a required answer cannot have appeared in an earlier trade in that run.

The first two weeks are scheduled from the reviewed collection. Later editions use the same rules and collection. Familiar trades can recur across different days. Ratings are starting estimates, not measured solve percentages; aggregate results will guide future changes.

Runs and scores

Three mistakes end the entire run. Completing a trade does not restore mistakes. You also get three optional hints for the entire run: select a blank to see the player’s initials and recorded playing years. Each blank has one clue, and a hint neither fills it nor costs a strike. Hint use is saved across trades and refreshes and included in shared results. Playing years come from the player statistics catalog and describe the recorded career span, not the player’s tenure with either team. A run continues until the third strike or until no further eligible trade remains in that day’s sequence. The score is completed trades. Each correct answer counts as one completed trade.

Everyone receives the same edition, with a new one at midnight in America/New_York, including daylight saving time. Past-edition links preserve their challenge. Progress is stored for an anonymous browser session; clearing site data or using another browser creates a separate run. There is no public leaderboard.

After resolving a trade, open its Tracker or Archive link for the canonical record and evidence. Play Trade Recall →

Teams & Eras

Window Explorer compares one-to-ten-season franchise runs from 1950–51 through 2025–26. It compares enabled metrics with equal input weight, and weights seasons and year-to-year transitions equally.

Level and shape

Lₘ = 100 · (k/n) · clip(1 − |uₘ(Aₖ) − uₘ(Bₖ)| / Rₘ(n))
Tₘ = 100/(n−1) · Σ valid adjacent t clip(1 − |ΔAₘ,t − ΔBₘ,t| / Dₘ)
L = mean(Lₘ)     T = mean(Tₘ)
S = αL + (1−α)T
clip(z) = min(1, max(0, z))

u is the mean of aligned available season values, except series wins, which are summed. k counts aligned seasons with data on both sides. Missing transitions contribute nothing and retain their denominator share. No missing year is bridged.

Fixed level ranges use 1 for win fraction, the historical maximum group size minus 1 for conference-rank gap, 100 for offensive/defensive rating percentiles, and the highest observed series-win total for the selected run length. Filters never change those ranges.

Rating percentile = 100 × (N − canonical league rank) / (N − 1). Supplied tied ranks retain equal percentiles; rounded raw ratings are never reranked. Win percentage uses the arithmetic mean of exact season wins / games. Conference finish remains a raw rank.

Playoff shape compares changes in annual series wins; the chart displays cumulative series wins. A one-season run compares level only. A rounded 100% does not establish identical yearly values.

The tool’s How similarity works panel includes the active metrics and exact range values for the selected run. Open Window Explorer →

Historical artwork

Team identities follow the first season of a run; individual seasons use their own artwork. Credits: SportsLogos.Net, SportsLogoHistory.com, SportsTeamHistory.com, and the NBA.

Coaching

Each connection represents a documented working relationship. The forest is not a map of friendship, general influence, or today’s staff.

Standard Connection

A coach is connected to the first National Basketball Association head coach under whom Basketball Reference lists that coach as an assistant.

Starting a New Tree

A coach begins a new tree if Basketball Reference does not list that coach as a National Basketball Association assistant before their first head-coaching appointment, unless one of the following exceptions applies.

Exceptions

Mikan–Kundla Exception

A former player is connected to the coach they played for throughout their entire National Basketball Association career if they had no assistant-coaching experience before becoming a head coach.

Applies to: George Mikan, Jim Pollard, Bill Russell, and Tom Heinsohn.

American Basketball Association Exception

A coach is connected through their first professional assistant position if that position was in the American Basketball Association, the head coach they worked under also coached in the National Basketball Association, and the assistant later became a National Basketball Association head coach.

If the head coach never coached in the National Basketball Association, the assistant begins a new tree.

Applies to: Doug Moe and Del Harris.

Player-Coach Exception

For a player-coach, the connection is determined by the lead executive listed by Basketball Reference for the season in which the appointment was made.

If that executive had previously been a National Basketball Association head coach, the player-coach joins the executive’s coaching tree. Otherwise, the player-coach begins a new tree.

Applies to: Dave Cowens, Bill Russell, Carl Braun, and Walt Budko.

Records and imagery

Career totals and season records come from the Basketball Reference Coach Register and linked coach profiles. Portraits come from Basketball Reference coach profiles and Wikimedia Commons. Individual records preserve source links and available creator and license information.

Statistics recalculate head-coaching seasons inside your selected span. Tree averages use coaches with qualifying head-coaching results in that span. The Statistics Center includes a glossary and exportable tables.

Explore the Forest → · Statistics →

Help improve the record.

Send a correction or source with the relevant page URL. Patrick reviews submissions before historical data changes.