Every rule we teach is backed by a published study.

Adding more events to an accumulator tends to increase the house edge versus a single-match bet.

In "request-a-bet" soccer parlays, the bookmaker profit margin was 43.7% overall and 74.6% on the longest-odds bets — roughly eight times the margin on conventional single bets.

US regulatory data show parlay hold percentages at least double the overall sportsbook hold — one state recorded an 18.75% parlay win rate versus 6-9% overall.

Bookmakers can increase profitability by offering more multiples and accumulators — confirmed in theoretical models of expected bookmaker profit.

Long-odds, multi-event bets are precisely where bookmakers embed the highest house edge versus standard single-game markets.

A 16-game parlay built from singles with a 10% house edge yields an 81.5% effective house edge — expected loss per dollar is vastly worse than singles.

Only 9.7% of multi-leg request-a-bet parlays won; bookmaker profit margin reached 74.6% on the longest-odds combinations.

In practice, multiplicative payouts mean sportsbooks take a higher percentage on parlays than on the underlying singles.

Single-bet markets in major sports can price close to fair; complex and long-shot parlays carry large negative expected returns.

From 2003-2017, NFL bettors were more likely to bet on teams that recently won; bookmakers earned "over the odds" specifically because of this quasi-rational recency bias.

NFL bettors are erroneously influenced by recent team performance, creating profitable opportunities for more patient bettors who resist recency bias.

In NFL markets using actual bet-percentage data, teams on against-the-spread winning streaks attract more bets and cold teams attract fewer — direct evidence of perceived hot-hand bias.

DFS participants pick "hot" players despite no true performance hot-hand effect existing — exactly the psychology that loads up same-game parlays with recent winners.

In an elite NFL gambling contest, average entrants performed no better than chance and showed clear recency bias; the skilled bettors who outperformed showed less recency bias. Skill = resisting the bias.

Risking $1.10 to win $1.00 makes 52.38% the mathematical breakeven win rate. Below that, the vig wins.

With more conservative vig assumptions ($1.20 loss to $1.00 win), required win rate climbs to 54.5%.

The -110 commission is the built-in house edge that converts posted odds into implied win probabilities.

Standard fixed-odds sports products embed a margin that requires a sustained edge above 52.4% to overcome.

Across 40,499 online sports bettors, the median fixed-odds bettor lost 29% of total amount wagered; live-action bettors lost 18%.

A model-based over/under strategy generated about 0.8% average profit per bet over 12 years — but profitability slowed in recent seasons as markets tightened.

Six previously published "profitable" NFL betting rules failed to remain profitable when retested on different time periods — most published edges do not survive out-of-sample.

MLB early-season underdog strategy generated persistent profits — until it was published. After publication, the edge dissipated.

Simulations combining advanced prediction models and betting strategies produced short-term gains, but consistent long-term profit was "elusive" under most conditions.

In a sample of 40,499 online sports bettors, the median fixed-odds bettor lost 29% of total wagers over about four months.

In a contemporary European cohort of 32,262 bettors, the median net loss was 25 euros over 8 months. A small "disproportionately high net winnings" group exists but is rare.

Operator data from an Australian online sportsbook showed 77.6% of bets were losses; odds are structured so customers are "unlikely to make a profit" over time.

Conventional soccer bets carry 5-7% bookmaker margins; custom long-odds request-a-bet products carry 44-75% margins.

Across multiple sports, no odds-based strategy produced statistically significant long-term profits in a large multi-sport dataset.

Adding an explicit drawdown constraint to Kelly produces stakes that outperform simple fractional Kelly at the same drawdown risk.

A weaker prediction model with a strong bankroll strategy can beat a stronger model with poor staking — staking discipline matters as much as picks.

"Unit-impact" staking mimics Kelly under uncertainty and beats flat-staking when the bettor has a real edge.

Setting deposit or loss limits is associated with lower daily wager, lower net loss, and lower betting intensity in online sports/race wagering.

Full Kelly maximizes long-term log growth of wealth but produces aggressive volatility. Fractional Kelly reduces risk and drawdowns at the cost of slower growth.

When win probability is uncertain (always, in practice), shrinking Kelly bets improves out-of-sample performance on real tennis betting data.

Modified Kelly approaches typically recommend smaller bet fractions than full Kelly when win probability estimates are noisy.

Adaptive fractional Kelly that responds to performance and risk performs well across sports and betting settings.

Sportsbook spreads and totals explain 79-86% of variability in median outcomes; only small biases (about one point) may exist, and detecting them requires substantial sample sizes.

NHL and MLB exhibit substantially higher randomness in outcomes than NBA — more games are required in "noisy" sports to be confident records reflect skill rather than luck.

Small samples push bettors toward gambler's fallacy reasoning; larger observed samples reduce the bias for moderate-probability outcomes.

A strategy using sportsbooks’ own closing odds to flag mispriced events produced sustained profits in 10-year simulations and real-money tests — until accounts hit limits.

Across six European football books, arbitrage opportunities are rare but highly profitable when found (12-200% per trade).

Inter-market arbitrage between bookmakers and exchanges appeared in roughly 19% of matches studied.

Price misalignment between books is frequent, but stake limits and account restrictions constrain how much a sharp bettor can extract.

Blindly betting NFL underdogs on the moneyline approached break-even overall and returned 2-7% in certain spread ranges — direct evidence of exploitable mispricing.

In NBA spread markets, line changes from open to close improve forecast accuracy — informed traders push prices toward truer probabilities.

MLB moneyline forecasts are mostly reliable but not fully efficient — simple strategies exploiting overreactions in line movements can be profitable.

After house edge, a bettor may need >10,000 bets and >9% ROI to statistically prove skill — anything less is indistinguishable from luck.

NFL sportsbook spreads capture 86% of variability in the true median margin of victory — biases that enable positive expected profit are rare and small.

Self-identified "expert" hockey bettors’ perceived skills were largely cognitive distortions, not genuine edge over randomness.

Only 4-7% of online sports bettors recalled their 30-day net result and bet count within 10% of actual account data — most underestimated both losses and bet count.

Account-level tracking allows direct measurement of chasing losses, session length, and deposit patterns — markers self-report can’t capture.

Bet-by-bet tracking from major operators is the gold standard for measuring real betting intensity and outcomes (vs. self-report).

Government activity statements alone don’t fix the recall problem — active tracking and review (not passive summaries) is what changes behavior.

Analysis of 565,915 online sports bets: after wins, bettors chose safer odds and won more; after losses, riskier odds and lost more — gambler’s fallacy creates real hot/cold streaks.

Bettors expect reversals after streaks (gambler’s fallacy), but the behavior they adopt creates real winning and losing streaks instead.

Gambler’s fallacy and hot-hand belief coexist in the same gamblers, rooted in the "law of small numbers" heuristic.

DFS players chase "hot" players even though no real hot-hand edge exists; smart-money winning lineups actively fade the bias.

NFL bettors are biased by recent team performance (recency bias), creating exploitable mispricing for more patient bettors.

Loss chasing involves compromised executive control, urgency, and compulsivity — it’s a measurable brain pattern, not a character flaw.

Chasing shows up as bigger stakes, longer sessions, faster play, and returning sooner after losses — especially in people with gambling problems.

In real online eCasino data, losses were more likely than wins to trigger chasing — bigger bets and longer sessions consistent with the "break-even" motive.

Brain imaging shows chasing versus quitting losses activates opposing reward/motivation and conflict/anxiety systems.

Craving, short time-horizons, and distorted beliefs ("I can control outcomes") predict chasing regardless of baseline gambling severity.

Sports-betting tilt is a real, measurable state: bettors cluster into conscious, unconscious, and non-tilters, and conscious tilters have the highest problem gambling severity.

Positive moods and induced emotion both reliably push bettors toward more risk-seeking, strategically weaker bets.

Even positive emotions like hope increase gambling via "feeling-based" processing, despite no real change in win probability beliefs.

Hazardous gamblers show amplified emotional swings — stronger highs after wins, stronger lows after losses and near-wins — even when their bet choices look normal.

Smartphone betting integrates sports betting into daily routines (in bed, commuting, at work), increasing bet frequency and making bets more spontaneous and less researched.

Sessions emphasizing mobility, privacy, and promotions led to more impulsive betting, more inducement use, and betting with multiple operators in the same window.

Nearly half of past-year sports bets in one sample were impulsive, and over three-quarters of bettors made at least one impulse bet.

App-based gambling is specifically associated with higher odds of problem gambling symptoms in young adults — above and beyond other modalities.

Using wagering inducements (bonuses, refunds) is followed in the same week by higher wager amounts, more betting days, and more chasing.

Frequent use of live/in-play betting, instant cash-out, and in-app live streaming is more common among bettors at higher risk of harm.

In a 2022 US survey of sports bettors: 47% no-risk, 20.6% low-risk, 12.3% moderate-risk, 20.1% high-risk on PGSI. About 1 in 3 sports bettors shows moderate or high problem-gambling risk.

After 2018 legalization, US sports wagers grew from $4.9B (2017) to $121.1B (2023); 94% of 2023 bets were online. Internet searches for "gambling addiction" rose 23% nationally after legalization.

International meta-analysis: about 1.3% of adults meet criteria for problem/pathological gambling and another 2.4% are at moderate/at-risk levels.

Sports betting co-occurs with other addictive behaviors over time, including problem drinking — supporting screening of sports bettors as a priority group.

Greater breadth and frequency of betting, money-motivated wagering, gambling urges, distorted beliefs about control, and lower self-control are the strongest predictors of sports-betting problems.

Account-tracking data confirms chasing as a measurable behavioral pattern: frequent within-session depositing after losses, repeated returning to "get even," and co-occurrence with high net losses.

Chasing is a formal Gambling Disorder criterion and may be the most distinct behavioral marker of the condition.

Within-session chasing shows up as increased bet size, longer sessions, and faster speed of play after losses — measurable in real-time platform data.

Multidimensional loss-chasing metrics predict gambling harm above clinical screening tools — early warning is technically measurable from account data.

Chasing is one of the most stable and prognostic gambling-disorder symptoms over time. People who chase are more likely to progress to severe problems.

Umbrella review: self-exclusion generally reduces gambling and gambling-related harm during the exclusion period, with downstream benefits to mental health and finances.

Self-exclusion use is low in the general population (~0.26%) but higher among problem gamblers (~15%). Programs are under-utilized due to complex enrollment, incomplete coverage, and low awareness.

A 7-day temporary online self-exclusion had no short-term behavioral effect; at 2 months it reduced gambling cognitions and craving but behavioral change was modest. Short time-outs are too short to matter.

Even with Sweden's nationwide self-exclusion system, 68% of self-excluders continued to gamble via unlicensed websites. Enforcement is the bottleneck.

Self-exclusion is a useful individual-level harm reduction tool but a weak standalone public health measure without higher uptake and stronger regulation/enforcement.

After US sports betting legalization, states with legal (especially online/mobile) betting saw small credit score declines and sizable increases in bankruptcies and debts sent to collections — concentrated among already financially vulnerable consumers.

Online casino plus live sports betting is strongly linked with over-indebtedness and expecting future over-indebtedness in general-population data.

College students at a US public university report sports-betting behaviors directly tied to financial strain, lost study time, and academic impact.

Among Rhode Island young adults, 11.5% had any gambling problems; sports betting participation roughly doubled the odds of having gambling problems.

Problem gambling in emerging adults is associated with consumer debt, defaults, and debt-enforcement actions; debt mediates the link to anxiety and depression.