How are NBA Draft Lottery odds calculated?

The NBA assigns each of the 14 lottery teams a fixed share of 1,000 numbered combinations, weighted by reverse regular-season record. The worst-record team receives 140 combinations — a 14.0% share — while the best-record lottery team receives just 5 combinations. Four lottery picks are drawn in sequence from those 1,000 combinations; the four winning teams take the first four draft positions. The remaining ten lottery teams fill picks 5 through 14 in reverse order of regular-season record without another drawing. Per the NBA's published lottery probability table, the Top 4 probability for the worst team is 52.1%, while the first overall pick probability is 14.0%.

TankOdds uses these same combination weights as the foundation for every simulation. The odds shown in the standings table come directly from the NBA's official lottery probability table, not from the simulator — they are deterministic inputs, not simulation outputs.

How does TankOdds simulate NBA Draft Lottery outcomes using Monte Carlo?

Monte Carlo simulation estimates NBA Draft Lottery odds by running a random process repeatedly and measuring the frequency of outcomes rather than deriving the answer analytically. Each TankOdds simulation runs one complete lottery drawing — using the official NBA combination weights (140 combinations each for seeds 1 through 3, decreasing to 5 combinations for seed 14, out of 1,000 total, per NBA official lottery rules) — and records where the donor team's final position landed. As simulations accumulate, the resulting estimates grow more stable with each additional trial. TankOdds can run up to 1,000 trials per session to estimate draft probabilities.

What does the Draft Pick Distribution show?

The histogram shows the distribution of lottery outcomes across all completed simulations. Each of the 14 bars corresponds to one possible draft position. The decimal below each bar is the proportion of simulations that produced that position as the lottery result for the seed being analyzed. Bars are color coded by team: positions within the protected range where the pick is retained use the donor team’s color, while positions outside the protected range where the pick conveys use the recipient team’s color. When two teams share a similar color, the simulator automatically substitutes gray for the donor and green for the recipient so the two groups remain visually distinct. As simulations accumulate, the bars form a distribution shape that reflects how outcomes are concentrated. A tight cluster around a few positions signals low variance in the lottery result for that seed. A spread across many positions signals high variance.

What does the Simulation Outcome Strip show?

Each simulation result is plotted as a vertical rectangle colored by outcome: the recipient team’s color when the pick conveys, the donor team’s color when it is retained. As results stack up, the strip becomes a color band. The density of recipient-colored versus donor-colored rectangles across that band gives a direct visual read on overall conveyance probability. A band dominated by the recipient’s color signals high probability of conveyance; a fragmented or mixed band shows how closely the two outcomes compete. When the two teams share a similar color, the simulator substitutes gray for the donor and green for the recipient automatically.

How does streaming conveyance probability work?

The streaming probability chart plots a smoothed version of the running conveyance estimate as simulations accumulate. The smoothing averages each new result against approximately the most recent 13 simulations, with recent trials weighted more heavily than older ones. This filters trial-to-trial noise while remaining responsive to genuine directional shifts in the estimate.

What is Lottery Seed Movement and why does it matter?

Seed Movement adds a random adjustment to the donor team's lottery position before each simulation draws. Shift magnitude is sampled from an exponentially decaying distribution, making moves of 0 to 1 seeds far more probable than jumps of 3 to 4. Direction is randomized each trial. This models remaining-season uncertainty in the donor's final standing rather than treating the current position as fixed.

Simulation Controls

Run starts the simulation. Stop interrupts it at any point. Lottery Seed sets the donor team's projected lottery position, 1 through 14. Protected picks are entered via the Protected Picks field using interval notation such as 1-4 or 1-4,10-14. Both fields are pre-filled automatically when clicking a standings row but can be manually adjusted to explore custom scenarios.

Simulation Totals and Conveyance Probability

The Simulations and Prob to Convey counters reflect the accumulated totals across all completed trials. The probability shown is the running conveyance estimate. With sufficient simulations, it approximates the true long-run probability of the pick conveying under the current seed and protection settings.

How the NCAA Champion Probability Simulator Works

The NCAA Champion Probability Simulator runs Monte Carlo simulation from the 2026 Final Four through the national title game, producing a probability distribution over the four remaining teams rather than a single simulated outcome.

Championship Odds as a Probability Foundation

Each team enters the NCAA Champion Simulator with a base probability derived from its opening Final Four championship odds from March 29, 2026. American odds are converted to implied probability using the standard formula: probability equals 100 divided by the sum of the odds and 100. These figures reflect the market's assessment of each team's realistic path to the title once the Final Four field was set.

Those championship odds are still a title market, not a single-game market, so the simulator uses them as a team-strength signal rather than treating them as literal semifinal win probabilities. Any matchup is rated by comparing the two teams' implied title probabilities directly, which keeps the model anchored to the same Final Four market without the extra square-root compression layer that used to inflate underdogs.

Win Probability and the Upset Factor

When two teams meet in a simulated matchup, the simulator compares their implied title probabilities directly. If Team A has implied championship probability pA and Team B has implied championship probability pB, Team A's base chance of winning the game is pA / (pA + pB). That keeps the favorite's edge tied closely to the opening Final Four title market while still translating the title board into a game-level comparison.

The Upset Factor slider blends the computed probability toward a pure 50/50 coin flip. At zero percent, the model follows the title-market baseline directly. At one hundred percent, every game becomes a coin flip regardless of odds. The default setting is now 10 percent, which keeps the model market-led while still allowing a little extra variance.

Final Four Simulation

Each simulation now walks only the final three games left in the tournament: the two national semifinals and the championship game. Every matchup is independently simulated using the per-game win probability for the two teams who reached that stage in that particular trial. Because one semifinal result changes the championship matchup, the field of likely finalists still shifts from trial to trial even with only four teams left.

The simulator runs 1,000 trials per session. Each trial produces exactly one champion. The resulting distribution shows how often each team won across all trials, which is the simulator's estimate of each team's probability of winning the 2026 NCAA Tournament under the opening Final Four championship-odds baseline and the current Upset Factor setting.

Champion Probability Bar Chart

The horizontal bar chart shows the simulated championship probability for the four remaining teams, sorted from least likely at the bottom to most likely at the top. Each bar is colored with the team's primary color. The percentage at the end of each bar is the fraction of all completed simulations in which that team won the championship. As simulations accumulate, the bars stabilize. Early in a run, results are noisy because each individual trial carries more weight in the average. By 1,000 trials, the estimates are stable enough for a clean Final Four probability read while keeping the app more responsive.

Simulation Outcome Strip

The outcome strip below the bar chart shows every completed simulation as a thin vertical column, colored by the team that won that trial. Reading left to right, the strip is the sequential history of all simulation outcomes. Runs where a single team dominates appear as dense bands of one color. More evenly distributed results produce a mosaic of color bands across the strip. The color key below the strip identifies which color corresponds to which team. Teams that never won any simulation during the current run do not appear in the key.

How the NBA Playoff Bracket Simulator Works

The NBA Playoff Bracket Simulator pulls live conference standings from the ESPN API and uses DraftKings Sportsbook championship odds to simulate the full 2026 NBA Playoffs, including the play-in tournament. Those standings determine the seeds, while the title market provides the strength signal for each best-of-7 series.

Live Standings and Seeding

Conference standings are fetched from the ESPN API each time the page loads, so seeds always reflect the current regular season. The simulator populates all 20 bracket slots — ten per conference including the play-in field — directly from the live standings. No manual seed entry is required. As the regular season progresses and teams move in the standings, reloading the page updates the bracket automatically.

Championship Odds and Win Probability

Each team enters the simulation with a base probability derived from its current DraftKings championship odds. American odds are converted to implied probability using the standard formula, then any head-to-head series is rated by comparing the two teams' implied championship probabilities directly. In other words, if Team A has implied title probability pA and Team B has implied title probability pB, the base chance of Team A winning the series is pA / (pA + pB).

This keeps the model tied to the live title market instead of falling back to seeding alone. If a team is missing from the championship-odds table, the simulator uses seed-based fallback logic only for that matchup. The Upset Factor then blends the base series probability toward 50/50, and the slider now starts at 0 percent so the market baseline is preserved unless the user chooses to add extra chaos.

Play-In Tournament

Seeds 7 and 8 in each conference are not guaranteed — they are earned through the play-in tournament, which runs in April before the main bracket opens. The simulator includes the full play-in field for both conferences below the main bracket. Game A (#7 vs #8): the winner locks in the #7 seed immediately. The loser is not eliminated — they fall to Game C. Game B (#9 vs #10): the winner advances to Game C. The loser is eliminated. Game C (loser of Game A vs winner of Game B): the winner earns the #8 seed and enters the first round against the #1 seed. The play-in must be simulated before the first round unlocks, because the #7 and #8 slots in the main bracket depend on those results.

Round Structure

After the play-in, 16 teams advance — 8 per conference, seeded 1 through 8. The first round pairs 1v8, 2v7, 3v6, and 4v5 in each conference across 8 best-of-7 series. Conference semifinals reduce each conference to two teams across 4 series. Conference finals produce one champion per conference. The NBA Finals is a best-of-7 between the West and East champions. Home-court advantage belongs to the higher seed in every round. The simulator resolves each round in sequence, advancing winners and locking in results before moving to the next stage.

Simulation Controls

Any individual matchup can be simulated by clicking it to open the series popup and selecting Simulate This Series. The slot machine animation resolves the series before the result locks in and dependent matchups unlock. Simulate All runs every available game from the earliest unlocked round forward, completing the entire bracket automatically. Reset Sim clears all simulated results while keeping live standings and seeds in place, allowing a fresh run with the same or adjusted Upset Factor. The Upset Factor slider (0 to 100 percent, default 10 percent) controls how much the underlying title-market edge is blended back toward random chance. On desktop, scroll to zoom the bracket and drag to pan; on mobile, pinch to zoom and drag to pan. Reset View returns the bracket to its default full view.

Export Options

The Export dropdown provides four options: Print opens a clean print layout with no controls or dark background. Save HD PNG downloads a high-resolution image of the full bracket suitable for sharing. Copy Image to Clipboard puts the bracket on the system clipboard for pasting into messages or social posts. Copy Image and Post to X copies the image and opens X with a pre-filled link. The bracket can be exported at any stage of simulation — partially or fully completed.

How the Lottery Forecast Works

The Lottery Forecast runs a full-lottery simulation that models all 14 NBA Draft Lottery picks simultaneously. Unlike the Team Pick Sim, which focuses on one team's seed and one pick, the Lottery Forecast uses live standings to build the full lottery field, applies current protection rules, and produces a projected order for every lottery team in a single pass.

Full-Lottery Simulation

Each simulation draws all four lottery picks in sequence using the official NBA combination weights — 140 combinations each for seeds 1 through 3, decreasing to 5 combinations for seed 14, out of 1,000 total, per the NBA's published lottery rules. The four lottery winners are drawn without replacement, so winning the first pick removes a team from consideration for picks 2 through 4. The remaining ten teams fill picks 5 through 14 in seed order. The result of each simulation is a complete 14-pick order assigning one team to every draft slot.

The app pulls the lottery field from current ESPN standings and maps known 2026 first-round conveyance rules onto those teams before the run begins. Users can choose how many runs to perform, from 1 to 100, with the current default set to 20.

Projected Pick Order Chart

The vertical bar chart shows one bar per pick slot (1 through 14). Each bar represents the projected team for that slot based on the observed full lottery orders in the current run. Bar height is the number of times that exact team-slot pairing occurred in the sample, not a site-wide fixed probability. Bars are colored with each team's primary color, with similar-hued teams shifted to distinct visual spectrums so all 14 bars remain distinguishable.

The projection is intentionally tied to complete observed lottery orders rather than stitching together 14 separate slot leaders. That prevents impossible combinations and keeps every displayed team-slot pairing grounded in a lottery order that actually occurred during the run. Early picks stay volatile because several teams have similar top-lottery odds; later picks become much more stable because seed order dominates once the four lottery winners are removed.

Mock Draft

Once the forecast completes, the projected team order populates a draft board with 14 empty pick slots. Clicking Generate Lottery Draft runs a mock draft that assigns one 2026 prospect to each pick using a weighted decision engine rather than a probability simulator.

The board blends three inputs stored in the current app logic: TankOdds prospect rankings, editorial team-need weights, and a lighter layer of recurring player-team fit boosts gathered from current expert mock drafts. Picks 1 through 5 are still driven mostly by top-of-board talent, but team need and fit signals have some influence there. From pick 6 onward, team need matters more and mock-fit signals get stronger, which makes the late lottery behave more like an actual front office board.

The result is intentionally deterministic for any given projected lottery order: if the same projected order appears again, the same weighted mock board will be produced. That keeps the tool consistent and makes it read more like an editorial best-fit draft board than a probabilistic pick forecast.

Players are revealed pick by pick with a short pause between each selection so the draft board fills in the way an actual draft room plays out. Teams marked with an asterisk (*) have picks that may convey if they fall outside protection, while double asterisks (**) indicate the current projected slot already sends the pick to the recipient shown.

NBA Standings

The NBA Standings page pulls live data from the ESPN API each time it loads and presents all 30 teams sorted best to worst by win percentage. Two views are available: Conference and Overall. Conference view stacks Western Conference standings and Eastern Conference standings in a single table, which makes it easy to scan playoff seeding, the play-in field, and the lottery line without switching pages. Overall view combines all 30 teams into one league-wide order.

The direct-playoff cutoff is highlighted in Conference view with a dotted separator line between seed 6 and seed 7. Seeds 1 through 6 advance directly to the playoffs, while seeds 7 through 10 enter the NBA play-in tournament. Teams 11 through 15 are outside the postseason picture and are primarily relevant for lottery positioning. That visual break is there to make the most important standings threshold easy to scan in a single glance.

The table columns extend well beyond a standard standings line. In addition to wins, losses, win percentage, games behind, home record, and road record, the page shows last ten games, current streak, points per game, opponent points per game, and point differential. Point differential is color-coded green for positive and red for negative values because it is one of the strongest regular-season indicators of underlying team quality. Home and road records matter because home-court advantage in the NBA Playoff Bracket flows directly from the regular-season standings shown here.

Games behind is handled differently in the two views. Conference view uses ESPN's conference-specific games-behind value. Overall view recalculates games behind relative to the team with the best record in the entire NBA, which creates a true league-wide measure instead of a conference-only one. Clinch markers from ESPN are normalized into a smaller set of badges: z for division clinch, y for conference clinch, x for playoff berth, p for play-in clinch, and e for eliminated. Eliminated rows are dimmed so the active race is easier to scan.

How is TankOdds different from other NBA draft lottery simulators?

Most NBA draft lottery tools run a single lottery spin and show one outcome. TankOdds runs up to 1,000 independent simulations per session and displays the full probability distribution — how often each pick position occurred across all trials. That distribution is more informative than any single spin because it shows the range of realistic outcomes and how tightly or loosely concentrated they are around a few positions.

TankOdds also resolves pick protection rules before each simulation draws, rather than applying them as a post-processing filter. That means every simulated outcome respects the current protection structure — no impossible lottery results appear in the distribution. The combination weighting system follows the NBA's official lottery rules, using the same published probability table (sourced from the NBA's official lottery documentation) that determines real draft positioning.