How lottery draws are conducted

North American major lotteries use one of two kinds of mechanical ball machines: air-mix machines (most common, used by Powerball and Mega Millions) and gravity-mix machines (used by Lotto Max and some smaller state lotteries).

An air-mix machine is a clear cylindrical chamber with numbered ping-pong-sized balls inside. A high-velocity fan at the base blows the balls into chaotic motion. A small opening at the top releases one ball at a time into a transparent collection tube, drawn by a separate smaller air flow.

Powerball uses two air-mix machines per draw — one for the 5 white balls (1-69) and a separate smaller machine for the single red Powerball (1-26). Mega Millions similarly uses two machines: one for the 5 white balls (1-70) and one for the Mega Ball (1-25). The Megaplier / Power Play number is drawn from a third smaller air-mix machine.

A gravity-mix machine uses an enclosed drum that rotates on a horizontal axis, with the balls tumbling under gravity. After ~30 seconds of tumbling, balls drop one at a time into a collection tray through a release mechanism.

Gravity-mix machines are typically used for smaller-pool lotteries (e.g., 6/49 style games) where the tumbling chamber doesn't need to handle very large ball populations. Lotto Max uses gravity-mix for its main 7-from-50 draw and Maxmillions selections.

Both styles have been studied extensively by independent statisticians who test the actual draw distributions against expected uniform distributions. The published analyses consistently find no statistically significant deviation from random for any current production-grade machine.

The numbered balls aren't ping-pong balls — they're purpose-built solid balls (typically rubber-core with a painted exterior) machined to extremely tight tolerances. The relevant tolerances:

A single set of balls is used per machine, with multiple identical sets kept in secure storage. The lottery rotates sets between draws to prevent any single set from developing wear patterns that could bias outcomes. Used sets are retired and tested at intervals to confirm they still meet tolerance specs.

Before each draw, an independent auditor inspects the ball machines and the ball sets. The auditor verifies:

A trial draw is run with the balls (different from the live broadcast draw) and discarded. This is a calibration check — the lottery's data team logs the trial outcomes and compares them statistically over time against expected distributions. Anomalies in trial draws trigger investigation before any live broadcast.

The live draw happens with multiple witnesses present:

The drawing host operates the machines following a published protocol. For Powerball: white balls drawn first, one at a time, with each ball lifted clear and confirmed on camera before the next ball is released. The 5 white balls take roughly 60 seconds total. The Powerball red is drawn from the second machine afterward. Power Play number is drawn last from the third machine.

After the draw:

The draw is video-recorded from multiple angles and archived. Any disputes after the fact can be resolved by reviewing the recording.

The auditor is the key trust anchor in the system. They are an external CPA firm (not a lottery employee, not paid per-prize-distributed) hired on a fixed-fee contract to attend and certify draws. Their job is to confirm the mechanical procedure was followed — not to verify the underlying randomness, which is the lottery's responsibility.

Independent auditors typically rotate between firms every 5-10 years to prevent complacency or conflict-of-interest buildup. Drummond & Goldberg has audited Powerball and Mega Millions for much of the past decade; Smith Bucklin has audited various state lotteries. The Lotto Max draws are audited by KPMG.

Beyond per-draw verification, lotteries continuously monitor the long-run distribution of each number. Over thousands of draws, each number should come up roughly its expected frequency (e.g., for Powerball white balls, each number 1-69 should appear in about 5/69 ≈ 7.25% of drawings). Significant deviations trigger machine inspection.

Modern lottery data teams use chi-square tests, Kolmogorov- Smirnov tests, and other distribution-fit measures against the published draw histories. Public researchers do the same — there's a small academic literature on lottery fairness analysis, and to date no published study has found statistically significant bias in any major U.S. or Canadian lottery's draw distribution.

The flip side: this is also why frequency analysis isn't forecast-grade. If the draws were biased, frequency analysis could detect and exploit it. Since they aren't, frequency differences between numbers are random variation, not signal.

The most famous lottery integrity failure in U.S. history involved Eddie Tipton, the security director of the Multi-State Lottery Association from 2005-2017. Tipton installed code in the random-number generators used by smaller state lottery games — Hot Lotto, Wisconsin Megabucks — that produced rigged outcomes on specific dates. He won (or arranged for associates to win) several jackpots before being caught when an unclaimed Hot Lotto jackpot was traced back to him.

The Tipton case illustrated the risk of any digital random number source touched by an insider. After 2017, U.S. multi-state lotteries (Powerball, Mega Millions, etc.) reaffirmed their use of only physical ball machines for the headline games — no RNGs in the loop. Some smaller state lottery games still use RNGs but with stricter oversight: dual independent codebases, code review by external security firms, continuous statistical monitoring, and air-gapped deployment.

The headline takeaway from the Tipton case: physical mechanical draws under multi-party observation are still considered the most-trustable randomness source for very large prize games. RNGs are cheaper and more flexible but introduce a single point of failure (the code) that's hard to audit even with multiple reviewers.

For a player asking "are the draws fair," the answer is functionally yes. The physical machines, the audited procedure, the continuous distribution monitoring, and the multi-witness ceremony together make insider manipulation extraordinarily difficult. The Tipton case is the exception that proves the rule — and after Tipton, the protocols tightened further.

Specifically you cannot:

For lottery players, the only mathematically rational optimizations relate to shared-jackpot risk and secondary tier value:

State-run lotteries in the U.S. began with New Hampshire in 1964. Before that, lotteries in America were primarily private operations — and many were rigged. The 1969 federal interstate lottery ban was partially in response to mob-affiliated private lotteries that systematically defrauded customers.

The modern state-run lottery model emerged specifically to provide a trustworthy, audited alternative. The Multi- State Lottery Association (MUSL) was formed in 1987 to consolidate operations across states, share machine infrastructure, and standardize audit procedures. Powerball launched as a MUSL product in 1992. Mega Millions launched as a competing consortium in 1996.

The integrity protocols described in this article are the products of 30+ years of iteration. Each notable failure (Tipton 2017, several smaller state game irregularities in the 1990s, occasional Canadian retailer fraud cases involving ticket-buying employees) has tightened the protocols. The system as it exists today is conservative, redundant, and surprisingly robust given the stakes.

We build JackpotX on the assumption that the underlying draws are fair. That assumption holds — every major U.S. and Canadian lottery draw is audited, witnessed, and statistically monitored. The randomness is real.

What that means for our analytics: frequency analysis is descriptive — it describes past draws, not future ones. Hot/cold patterns are real features of the data — random sampling produces variation — but they don't tell you anything about future draws, because the past does not constrain the future. See /about-frequency-analysis for the deeper treatment.

If you ever see a lottery "system" that promises to forecast the next draw, or a "guaranteed winners" service, the seller is exploiting the fact that the draws are fair to sell you a product that does not work. JackpotX deliberately avoids the marketing language that scammers use, precisely because of how misleading it is in the lottery space.