4
LottoLogic IQ — The Institutional Lottery Terminal
SYSONLINE
DRAW INDEX75,408▲
GAMES98
RETAILERS13,432
PWRBALL$325M+12.0%
MEGA$190M+5.5%
FL LOTTO$2.5M
F5 MID$200K
F5 EVE$220K
JTP$650K
C4L$1K/WK
PB HURST0.49
PB χ²1068.87df=68
P3 ADF-22.42
P4 ADF-20.75
F5 ADF-15.40
FL LOTTO HHI0.020
F5 SKEW+0.12
PB Z-MAX+2.14σ #32
ρ PB↔MM0.62
ρ P3↔P40.71
SHARPE TOP0.18 #1627
KELLY f*0.034
EDGE SCORE40.0
#1627EDGE α+1.187
#1614FADE α-7.472
RETAILERFT LAUDERDALE 33301
SYNC06:00 ET
SYSONLINE
DRAW INDEX75,408▲
GAMES98
RETAILERS13,432
PWRBALL$325M+12.0%
MEGA$190M+5.5%
FL LOTTO$2.5M
F5 MID$200K
F5 EVE$220K
JTP$650K
C4L$1K/WK
PB HURST0.49
PB χ²1068.87df=68
P3 ADF-22.42
P4 ADF-20.75
F5 ADF-15.40
FL LOTTO HHI0.020
F5 SKEW+0.12
PB Z-MAX+2.14σ #32
ρ PB↔MM0.62
ρ P3↔P40.71
SHARPE TOP0.18 #1627
KELLY f*0.034
EDGE SCORE40.0
#1627EDGE α+1.187
#1614FADE α-7.472
RETAILERFT LAUDERDALE 33301
SYNC06:00 ET
Live · Florida Terminal/Quantitative Lottery Analytics/v2.1 · 12 Stat Models
Stop buying the popular ticket. Buy the smart one.
We rank every active Florida scratch-off and draw game by which ones still have top prizes left, which ones are fresh out of the box, and which ones are statistically due. One click, one ranked list, zero spreadsheets. Free to try, no card.
Treat the lottery like luck. We model it.
LottoLogic IQ is the first institutional-grade analytics terminal for the Florida Lottery.
We score every ticket on EV Yield, ROI, Sharpe, CAPM α, and Edge Score — and run a 12-metric
statistical lab (Hurst, χ², ADF, HHI, Skew, Kurtosis, Z-Score) over 75,408+
indexed draws to surface mean-reverting anomalies.
EXAMPLE SNAPSHOTReal-time index · growing every draw
Every draw, every game, every prize — indexed and growing daily.
75,442▲
Draws Indexed
+17 today · since 05/07/26
98
Games Monitored
88 scratch · 10 draw
13,432
Retailer Records
Updated 06:00 ET
12
Stat Models
α · f* · H · χ² · Z · ρ
01 The Terminal · Cinematic Walkthrough
A 90-second feature reel. Press play.
Fourteen scenes shot inside the LottoLogic IQ terminal. Cold open, data ingestion,
market overview, statistical formula ignition, deep analytics table, Hurst mean-reversion,
cointegration matrix in 3D, hot/cold/due engine, scratch-off ranking, AI backtester,
Kelly optimizer, strategy center, retailer geofence, and final title card. Rendered live
in your browser — every frame is the working engine.
7
23
42
11
36
3
19
52
8
63
14
29
5
47
21
38
17
55
9
68
LottoLogic IQ
FLORIDA TERMINAL · v2.1 · 12 STAT MODELS
ACT ISCENE 01 · COLD OPEN — Twenty balls cascade. One ignites the terminal. Welcome to the quant lab.
DATA INGESTIONFlorida Lottery PipelineLIVE
> connectionfl_lottery_api
> powerball2,035 draws ingested
> mega_millions2,487 draws
> fl_lotto3,355 draws
> fantasy_510,942 draws
> pick_3 / pick_419,838 / 19,180
> cash_pop_5x7,695 draws
> scratch_catalog88 / 88 games
> total█ TERMINAL ONLINE
Indexed Corpus75,408
ACT ISCENE 02 · INGEST — Edge worker pulls every active game on a 06:00 ET cron. 75,408 draws live in cloud database.
ACT ISCENE 03 · DASHBOARD — Six headline metrics. #1627 leads on EV ($35.77 / $50) with all top prizes intact.
DEEP ANALYTICS · STATISTICAL LABTwelve formulas. One operating picture.
α
CAPM Alpha
+1.187
f*
Kelly Fraction
0.087
SR
Sharpe Ratio
0.18
H
Hurst Exponent
0.487
χ2
Chi-Squared
p < 0.05
ADF
Aug. Dickey-Fuller
−3.42
HHI
Herfindahl Index
0.168
γ1
Skewness
+0.34
γ2
Kurtosis
2.91
Z
Z-Score
+1.96
RSI
Relative Strength
62.4
ρ
Cointegration
0.62
ACT IISCENE 04 · FORMULAS — Twelve academic statistical models. The same battery used in quant trading research. Each formula computed for every game daily.
DEEP ANALYTICS · ALL DRAW GAMESThe full coverage tableEXAMPLE SNAPSHOT
GAME
DRAWS
HURST
χ²
ADF
HHI
SKEW
RSI
SIGNAL
POWERBALL
2,035
0.49
1068.87
-3.46
0.018
+0.04
52.3
TOP TIER
MEGA MILLIONS
2,487
0.50
159.51
-2.84
0.017
-0.02
49.8
MIXED
FL LOTTO
3,355
0.49
51.41
-7.92
0.020
+0.07
54.1
TOP TIER
JACKPOT TRIPLE
745
0.51
46.27
-3.96
0.022
+0.09
56.0
TOP TIER
CASH4LIFE
2,674
0.49
46.89
-6.79
0.017
+0.01
50.5
MIXED
FANTASY 5
10,942
0.48
28.01
-15.40
0.029
+0.12
53.7
TOP TIER
PICK 4
19,180
0.48
7.40
-20.75
0.103
+0.02
51.2
TOP TIER
PICK 3
19,838
0.48
19.86
-22.42
0.103
+0.04
52.8
TOP TIER
ACT IISCENE 05 · COVERAGE — Every draw game scored on 12 metrics. Three EDGE flags align: Pick 3, Pick 4, Fantasy 5.
HURST EXPONENT · RESCALED-RANGE ANALYSISPowerball — 1,803 draws
⚠ OBSERVATIONAL DATA · NOT A PREDICTION TOOL · Lottery draws are statutorily i.i.d. random
H = log(R/S) / log(N)
Hurst 1951 · Mandelbrot & Wallis 1969
0.49
R/S statistic · frequency-space
STATISTICAL OBSERVATION
Single-window R/S over ball-frequency dispersion. Not a time-series Hurst signal. Lottery is statutorily i.i.d. — observed dispersion ≈ 0.5 ± sampling noise.
Rescaled-range walk · normalized R/S
ACT IISCENE 06 · HURST — Live H value updates from the data. Statistical observation, not a tradable signal.
High-ρ Pick-game cells (🔒) require Pro+Elite. Visible cells show low-signal benchmarks. PB ↔ MM ρ = 0.62 live · refreshes daily.
ACT IISCENE 07 · COINTEGRATION — 10×10 matrix · top-signal cells locked. Cross-game research surfaces; subscribe to act on them.
FREQUENCY ANALYSIS LABORATORY · 14,601 DRAWSHot / Cold / Due — Pick 5EXAMPLE SNAPSHOT
+EV GAMES TODAY
3
games crossed positive expected value · roll triggers · jackpot growth
LIVE · 14,601 analyzed
PBMMFL LOTTOJTPC4LF5P4P3P2PICK 5
046.9%
148.2%
248.5%
349.1%
447.8%
546.9%
647.9%
749.7%
848.2%
948.5%
🔥 HOT · above avg
7
Ball #7 · 7,254 hits
49.68% · +3% avg
❄ COLD · below avg
0
Ball #0 · 6,852 hits
46.93% · -3% avg
⚡ DUE · longest gap
8
Ball #8 · gap 39 draws
48.19% · 0% avg
ACT IISCENE 08 · HOT/COLD/DUE — Three lenses on every game. Ride the trend, fade the crowd, or chase the gap.
SCRATCH-OFF MARKET · 88 ACTIVE GAMES · EV THE NORTH STARTop plays · sorted by Edge ScoreEV = $ EXPECTED · YIELD = EV ÷ TICKET
01
💎
#1627500X THE CASH$50 · 4.05 odds · launched Feb 2026 · 2 of 2 jackpots intact
$35.77
EV
71.5%
YIELD
+1.187
α
40.0
SCORE
TOP TIER
02
⚡
#1684$2M FORTUNE$10 · launched Mar 2026 · top prizes intact
$6.90
EV
69.0%
YIELD
+0.572
α
31.2
SCORE
TOP TIER
03
🏆
#1623200X THE CASH$20 · 3.11 odds · launched Jan 2026 · 3 of 4 jackpots intact
$10.31
EV
51.6%
YIELD
+0.084
α
22.8
SCORE
MIXED
04
💵
#1622100X THE CASH$10 · 3.35 odds · launched Jan 2026 · 3 of 4 intact
$4.87
EV
48.7%
YIELD
-0.213
α
14.5
SCORE
SKIP
05
🎰
#1614WIN IT ALL$2 · 5.20 odds · launched Dec 2025 · most top prizes claimed
$0.30
EV
15.0%
YIELD
-7.472
α
3.1
SCORE
SKIP
ACT IIISCENE 09 · SCRATCH RANKING — All 88 games scored on published return rates. TOP TIER = best published RTP, SKIP = worst. Published return rates vary between scratch games.
Hot Bias · Conservative · PB→H = 0.4873 of 5 balls show +2σχ² rejects uniform · p < 0.05
ACT IIISCENE 11 · FORWARD LEDGER — Every pick logged 06:00 ET, graded after the draw. Wins are rare. Losses are real. The receipts don't lie.
STRATEGY CENTERFour lenses. One terminal.
🔥
HOT BIAS
Ride the momentum.
Weights numbers above their long-run mean. Best in trending games (Hurst > 0.50) and rising χ².
Hurst H0.512χ² p-value< 0.05
❄
CONTRARIAN
Fade the crowd.
Weights numbers below mean. Best in mean-reverting games (Hurst < 0.49) with strongly negative ADF.
Hurst H0.487ADF stat−3.42
⚡
DUE NUMBERS
Chase the gap.
Ranks balls by gap-since-last-hit ÷ expected gap. Z > 2 = statistically overdue.
Top gap39 drawsZ-score+2.41σ
⚖
AI BALANCED
Blend all eight signals.
Equal-weighted: full freq · S20/S50/S200 · gap · χ² · pair co-occurrence · cross-game ρ.
Cross-game ρ0.62Skew γ₁+0.34
ACT IIISCENE 12 · STRATEGY — Each lens defined by its statistical regime. Pick the one that fits the game's signature.
SCRATCH-OFF GEOFENCE · INTACT TOP PRIZES NEAR YOUFlorida — scratch retailers · last 30 days13,432 RETAILERS
2 of 2 intact · #1627
3 of 4 intact · #1623
top prizes intact · #1684
all claimed · #1614
3 of 4 intact · #1622
2 of 3 intact · #1601
88Active scratch games
$0Edge on draw games (uniform odds)
13,432Retailer records tracked
Scratch onlyRetailer-level edge applies
ACT IVSCENE 13 · GEOFENCE — Scratch-off only. Buy where intact top prizes remain. Draw games have uniform odds at every retailer.
Your Shopping List
Wednesday, May 6, 2026 · Diversified AI Portfolio · Target Area: 33131
LottoLogic IQ
v2.1 · FL TERMINAL
SCRATCH-OFFS
2×#1579 JACKPOT JUBILEE$10
2×#1297 $50K FLAMINGO MULT$4
1×#1630 $2,000,000 FORTUNE$10
1×#1619 10X THE CASH$1
DRAW GAMES · AI Balanced · Balanced Tier
1×POWERBALL [QP] 5 36 42 52 63 · PB 6$2
1×MEGA MILLIONS [QP] 4 47 53 55 68 · MB 18$5
1×FL LOTTO [QP] 1 10 17 21 46 52$2
1×FANTASY 5 EVE [QP] 9 12 13 25 36$1
TOTAL SPEND$35
Budget: $50 · Used: $35 · Surplus: $15
🧠 AI Analyst Notes:
Draw picks sourced from today's Forward Ledger — the same picks tracked server-side and graded against official FL Lottery results. Discipline check · i.i.d. random · past performance ≠ future returns.
🔒 PRINT 1-PAGE SHOPPING LIST · ELITE
The lottery is luck. We bring the math.
Launch Terminal · Free →
No credit card · Guest mode in 1 click · 3-day Pro & Elite trial
ACT IVSCENE 14 · LAUNCH — One printable list every morning. See the numbers. Decide for yourself.
0:00/ 1:30
LIVE
01Cold Open
02Ingest
03Dashboard
04Formulas
05Coverage
06Hurst
07Cointegration
08Hot/Cold/Due
09Scratch Rank
10Backtester
11Kelly
12Strategy
13Geofence
14Launch
02 Statistical Lab
Twelve metrics. One operating picture.
Every draw game runs through a 12-metric statistical battery — the same family of tests used
to validate quant trading models. Hurst quantifies mean-reversion. χ² flags non-uniform
distributions. ADF rejects unit-root. HHI measures concentration. Skew and kurtosis check shape.
Z-scores surface anomalies. Every test, every game, every day.
In Plain English
Most "lucky number" advice is vibes. This is a battery of academic statistical tests applied to over
150,000 indexed lottery draws. We're not predicting winners — we're showing you which numbers have
drifted from where they should be, and which games have the least-bad odds. Use it like a
research dashboard, not a crystal ball.
α
CAPM Alpha
f*
Kelly Fraction
SR
Sharpe Ratio
H
Hurst Exponent
χ²
Chi-Squared
ADF
Aug. Dickey-Fuller
HHI
Herfindahl Index
γ₁
Skewness
γ₂
Kurtosis
Z
Z-Score
μ
Mean Deviation
ρ
Cointegration
α
CAPM ALPHA
Excess EV over state-wide market.
Adapted from the Capital Asset Pricing Model. Each game's α measures excess EV vs. the FL Lottery aggregate yield. Positive α = mispriced opportunity.
Plain English: α tells you which games are the "least bad bet." Higher = the state lost more edge on that ticket than usual.
Best game#1627 +1.187
f*
KELLY CRITERION
Optimal stake fractions.
Computes the bet size that maximizes log-utility growth: f* = (bp − q) / b. Fractional Kelly with budget & risk constraints, never overbets a negative-EV asset.
Plain English: Kelly answers "how much should I spend?" If a game is bad, Kelly says zero. If decent, it says a small fraction of your budget.
Top f* today0.034
SR
SHARPE RATIO
Risk-adjusted return.
Sharpe = (EV − Rf) ÷ σ. Higher = better risk-adjusted yield. Note: lottery σ is huge from rare big prizes — realistic Sharpes are tiny (0.02 – 0.20).
Plain English: Two games with the same EV aren't equal — one might have wilder swings. Sharpe rewards steady EV over spiky EV.
Top Sharpe0.18 #1627
H
HURST EXPONENT
Mean-reverting vs. trending.
Rescaled-range analysis: H ≈ 0.5 = random walk, H < 0.5 = mean-reverting (anti-persistent), H > 0.5 = trending. Lottery should be ~0.5 — deviations are signals.
Plain English: When a number's been "hot" for a while, does it stay hot? Hurst answers that. Below 0.5 means hot streaks tend to cool.
Powerball H0.49
χ2
CHI-SQUARED
Uniformity test.
Tests whether numbers appear at expected rates (no bias). Large χ² with deep history may indicate distributional quirks worth tracking — even fair draws have local clusters.
Plain English: If the lottery were perfectly fair, every ball appears equally often. χ² measures how far reality is from that ideal.
Pwrball χ²1068.87 · df=68
ADF
AUG. DICKEY-FULLER
Stationarity check.
H₀: series has unit root (non-stationary). Reject H₀ → frequency deviations mean-revert. Pick 3 ADF -22.42 (highly stationary) means deviations don't compound — they correct.
Plain English: ADF asks: "do streaks build up forever, or do they fix themselves?" A strongly negative ADF means they fix themselves.
P3 ADF-22.42 · p≈0.005
HHI
HERFINDAHL
Concentration index.
Borrowed from antitrust economics. Sum of squared frequencies. High HHI = a few numbers dominate. Low HHI = spread out. Confirms whether bias claims have any teeth.
Plain English: Are wins concentrated in a few balls, or spread evenly? Low HHI = spread out (which is what fair lotteries look like).
FL Lotto HHI0.020
Z
Z-SCORE
Standard deviations from mean.
Per-number Z-score reveals outliers. A ball trending +2σ above its expected frequency is statistically anomalous — flagged in Hot/Cold Lab and weighted in the AI engine.
Plain English: Z tells you "is this number unusually frequent?" A Z above +2 means that's about a 1-in-40 chance by pure luck.
Top Z (PB)#32 +2.14σ
Live Statistical Feed · All Draw Games
Updated 06:00 ET · synced from cloud database
GAME
DRAWS
POOL
HURST H
χ²
ADF
HHI
SKEW γ₁
KURT γ₂
VOLUME
SIGNAL
POWERBALL
1,803
1–69
0.49
1068.87
-3.46
0.018
+0.04
-0.18
2×/wk
TOP TIER
MEGA MILLIONS
1,563
1–70
0.50
159.51
-2.84
0.017
-0.02
-0.05
2×/wk
MIXED
FL LOTTO
2,130
1–53
0.49
51.41
-7.92
0.020
+0.07
-0.12
2×/wk
TOP TIER
JACKPOT TRIPLE PLAY
744
1–46
0.51
46.27
-3.96
0.022
+0.09
-0.21
2×/wk
TOP TIER
CASH4LIFE
2,674
1–60
0.49
46.89
-6.79
0.017
+0.01
-0.08
daily
MIXED
FANTASY 5
12,027
1–36
0.48
28.01
-15.40
0.029
+0.12
-0.34
2×/day
TOP TIER
PICK 4
21,868
0–9
0.48
7.40
-20.75
0.103
+0.02
-0.07
2×/day
TOP TIER
PICK 3
21,868
0–9
0.48
19.86
-22.42
0.103
+0.04
-0.06
2×/day
TOP TIER
PICK 2
14,046
0–9
0.49
21.06
-18.43
0.105
+0.06
-0.11
2×/day
MIXED
PICK 5
14,602
0–9
0.48
93.65
-7.94
0.108
+0.08
-0.19
2×/day
TOP TIER
CASH POP
5,847
1–15
0.50
14.22
-9.18
0.072
+0.03
-0.14
5×/day
MIXED
Reading the Live Feed
Each row is a draw game. Hurst < 0.5 means deviations correct themselves (good for "due number" plays).
Strongly negative ADF confirms the same thing — frequency drift mean-reverts. Low HHI means winnings are
spread across many balls, not a few favorites. The TOP TIER badge appears when at least 3 of those signals align.
Fantasy 5 with 12,027 draws and ADF -15.40 is a textbook EDGE setup.
03 Cross-Game Research Matrix
When games move together.
Amateur tools stop at correlation (Pearson's ρ): two games happen to co-move in the historical record. We add a second test — Engle-Granger 2-step — that asks whether the residual spread is stationary, the methodological prerequisite for any cross-game research signal. When a high-correlation, stationary-spread pair shows a wide divergence, it's a research-grade context flag for cross-game strategies in the AI Backtester. It is not a guarantee of future co-movement; lottery games are statutorily games of chance and every individual ticket remains negative-EV.
Pearson ρ on frequency-drift series · Engle-Granger 2-step (MacKinnon 1994 critical values) · 1,803–21,868 draws each
Values shown are Pearson ρ between game-pair frequency-drift series. Cyan ≥ 0.50 = significant. Green ≥ 0.65 = highly significant. The Engle-Granger 2-step is run separately and verifies that the residual spread is stationary — a sanity check that rules out spurious correlation from drifting series.
The Quant Advantage: Correlation vs. Cointegration
Most lottery software stops at simple correlation (Pearson's ρ). If Pick 3 frequencies and Pick 4 frequencies move together, they are correlated — but correlation can show up by chance and disappear the moment you act on it. LottoLogic IQ adds the Engle-Granger 2-step on top: an OLS regression of one game's frequency-drift series on the other, then an ADF test on the residuals to verify the spread is stationary. What that gets you is a defensible filter: pairs where co-movement in the historical record was not just drifting noise. How we use it is as one of the eight signals fed into the multi-game AI Backtester — never as a standalone "edge." The lottery remains a game of chance and every ticket is negative-EV; the matrix is a research lens, not an arbitrage. If a tool ever tells you a lottery strategy "must" mean-revert, walk away — that is the gambler's fallacy in finance vocabulary, and we don't trade on it.
04 The AI Engine
Three engines. One terminal.
Behind the dashboard sits a continuously running compute layer. Three engines feed into every recommendation: a Hot/Cold/Due frequency analyzer that classifies every digit and ball position by its statistical regime; an AI Combo Optimizer that brute-forces millions of combinations against historical draw distributions; and a Mean-Reversion module that estimates Hurst H and half-life on every drift series. All three update on the 06:00 ET cron and stream to your terminal in milliseconds.
01LIVE
Hot / Cold / Due
Z-score classification of every digit · weighted by recent vs. long-run frequency.
7HOT
3HOT
5
0COLD
8DUE
1
2COLD
9HOT
4
6DUE
PICK 3 · LAST 50Z = 1.84
Hot · z > +1Cold · z < −1Due · gap ≥ 2σ
02ELITE
AI Combo Optimizer
Brute-forces millions of combinations against the 8-signal scoring stack. Ranks by composite Edge Score.
Estimates Hurst exponent (H) and half-life on every digit's frequency-drift series. H < 0.5 = anti-persistent.
H0.487
half-life14 draws
verdictanti-persistent
R/S analysis · Augmented Dickey-Fuller verified
How to read this: The three engines surface statistical patterns in the historical record. Patterns are not predictions. Lottery games are statutorily games of chance and every individual ticket remains negative-EV. We use these tools to find the least bad plays — never to claim a guaranteed win.
04 Strategy Center
Six lenses. Pick your edge.
Same data, six ways to read it. The Strategy Center compiles AI-ranked combinations under each
philosophical lens — Hot Bias rides momentum, Contrarian fades it, Due Numbers chases gaps,
Statistical Edge exploits deviation, Gemini Quant targets sleeping momentum, and AI Balanced blends all signals.
🔥 HOT BIAS
Ride the momentum.
Weights numbers above their long-run frequency mean. Best when Hurst > 0.50 (slight trending) and recent χ² is rising.
When to use: Short-cycle games (Pick 3/4) where you've seen a number hit 3+ times in 14 draws.
729
Hurst H0.512
χ² p-value< 0.05
❄ CONTRARIAN
Fade the crowd.
Weights numbers below their frequency mean. Best when Hurst < 0.49 (mean-reverting) and ADF strongly negative.
When to use: Most picks games. Crowd plays hot numbers, so cold numbers split prizes less when they hit.
058
Hurst H0.487
ADF stat−3.42
⚡ DUE NUMBERS
Chase the gap.
Ranks balls by draws-since-last-hit divided by expected gap. Z-score > 2 = statistically overdue.
When to use: Pick 5 / Fantasy 5 with confirmed mean-reverting ADF. The longer the gap, the stronger the signal.
When to use: Default for most users. Stable across game types. Best long-run performance in our 12-month backtest.
379
Cross-game ρ0.62
Skew γ₁+0.34
📊 STATISTICAL EDGE
Exploit deviation.
Weights numbers by χ² deviation from expected uniformity. Best when χ² rejects randomness (p < 0.05).
When to use: Games showing severe non-uniformity across large sample sizes.
417
χ² stat42.7
p-value< 0.05
🌙 GEMINI QUANT
Sleeping momentum.
Targets the 'sweet spot' gap window (3-15 draws) combined with cross-game cointegration state boosts.
When to use: Multi-state draw games (Powerball/Mega) where cross-correlation provides hidden edge.
285
Gap window8 draws
Cross-ρ+0.62
AI Combo Optimizer · top 6 by signal score
PICK 3 · BALANCED · 14d window · last run 06:00 ET · ILLUSTRATIVE
#
COMBO
SIGNAL SCORE
SIGNALS ALIGNED
χ² DEVIATION
GAP DAYS
TIER
01
729
94.2
7 of 8
+2.41σ
14
TOP TIER
02
359
87.6
6 of 8
+1.92σ
11
TOP TIER
03
912
82.4
5 of 8
+1.74σ
9
TOP TIER
04
835
71.2
4 of 8
+1.21σ
6
MIXED
05
274
68.8
4 of 8
+1.08σ
5
MIXED
06
548
42.1
2 of 8
−0.42σ
2
SKIP
How to Read the Optimizer
Each combo is scored by all eight AI signals on a 0–100 scale, derived from statistical observations: χ² deviation from uniform expected frequency, draws-since-last-hit (gap days), and signal-alignment count. Higher score = stronger statistical observation, not a prediction. The Optimizer reports observation strength only — it does not claim that these combinations will hit. Lottery games are statutorily i.i.d. random; any individual ticket remains negative-EV regardless of signal score.
05 Backtester · Results Snippets
What the engine looks like in production.
The AI Backtester runs your chosen strategy against the last N draws and reports total spent, total won,
net P&L, dollar-positive play rate, and the variance distribution. Two snippet snapshots from the in-app terminal — Pick 3
and Powerball strategies. Both show the typical loss curve a real strategy produces: most plays lose money,
a small minority break even, the rare hit drives outliers.
⚠ ILLUSTRATIVE SNIPPET · Synthetic test dates · Not a verified prospective backtest run · Lottery is statutorily negative-EV
BT-RUN · 0427-001
PICK 3 · AGGRESSIVE · 18 plays · $1 stake/play
Total Spent
$18
18 × $1
Total Won
$80
single $80 box hit
Net P&L
+$62
single-hit driven
Dollar-positive plays
1 / 18
5.6%
Strip-the-hit P&L
−$17
17 losses, 0 wins
Best play
2 / 3 box
$80 partial
DRAW
PICK vs DRAW
RESULT
WON
NET
TST-12 EVE
729
0/3
$0
−$1
TST-11 MID
359
0/3
$0
−$1
TST-10 EVE
832
2/3 box
$80
+$79
TST-09 MID
146
0/3
$0
−$1
TST-08 EVE
912
0/3
$0
−$1
TST-07 MID
675
0/3
$0
−$1
Snippet · 6 of 18 plays shown17 losses · 1 partial · 0 straight
BT-RUN · 0427-002
POWERBALL · BALANCED · 50 plays · $2 stake/play
Total Spent
$100
50 × $2
Total Won
$23
small partial hits
Net P&L
−$77
−77% of stake
Dollar-positive plays
3 / 50
6.0%
vs random baseline
−77% vs −50%
underperformed
Best play
3 white
$7 (lost net $-$13)
DRAW
HITS
TIER
WON
NET
TST-23 SAT
3
Match 3
$7
−$13 (won < stake× count)
TST-21 WED
2 + PB
Match 2+1
$7
−$13
TST-19 SAT
1 + PB
Match 1+1
$4
−$16
TST-18 MON
0
Miss
$0
−$2
TST-16 SAT
0
Miss
$0
−$2
TST-14 WED
0
Miss
$0
−$2
Snippet · 6 of 50 plays shown47 losses · 3 small partial wins · 0 jackpot
What This Snippet Shows (and Why You Shouldn't Trust One Snippet)
The Pick 3 snippet shows the kind of variance that makes single-snippet backtests unreliable: a single $80 partial hit in 18 plays produced a net positive, but pull that one play out and the run is −$17 (a 94% loss rate). The Powerball snippet is more representative of expected behavior over a longer window: 47 of 50 plays missed, the 3 partial-match plays still cost more than they paid back, and the run finished at −77% — worse than the −50% house-edge baseline. The backtester's job is not to demonstrate winning strategies — it is to let you observe how a strategy behaves before you risk real money. Run a full backtest in the in-app terminal to see the loss curve over hundreds of plays, not six.
06 Methodology
Six formulas. Every signal we publish.
These are the foundational equations behind the Statistical Lab and AI Engine. Same primitives that
underpin modern portfolio theory and quant trading. Adapted to a discrete, finite, negative-EV asset class.
EQ. 01
αi = Ri − [Rf + βi(Rm − Rf)]
CAPM ALPHA · GAME-RELATIVE EXCESS YIELD
Each game's α measures EV beyond what the FL Lottery aggregate market would predict given the game's beta. Positive α isolates mispriced opportunities — games where the state's edge is structurally smaller than the headline odds suggest.
Maximizes long-run log-utility. b = net odds, p = win prob, q = 1−p. We apply fractional Kelly with budget caps and risk tiers (Conservative / Balanced / Aggressive). Returns 0 on negative-EV plays — never overbets.
Kelly (1956) · Thorp implementations · capped-fractional variant
EQ. 03
H = log(R/S) / log(N)
HURST EXPONENT · MEAN-REVERSION TEST
Rescaled-range analysis. H ≈ 0.5 = random walk · H < 0.5 = mean-reverting · H > 0.5 = trending. Most lottery games sit at 0.48–0.51 (essentially random) — but slight anti-persistence (Pick 3/4/5 at 0.48) means frequency drifts do correct.
Hurst (1951) · Mandelbrot & Wallis (1969)
EQ. 04
SR = (E[R] − Rf) / σR
SHARPE RATIO · RISK-ADJUSTED YIELD
Expected excess return per unit of standard deviation. Lottery σ is dominated by rare large prizes, so realistic Sharpes are tiny (0.02 – 0.20). A higher SR for the same EV means steadier returns — useful for ranking mid-tier scratch-offs head-to-head.
Sharpe (1966) · high-σ asset reformulation
EQ. 05
χ² = Σ(O − E)² / E
ECONOMETRIC SUITE · χ² · ADF · HHI
Three orthogonal tests run together: χ² for uniformity, Augmented Dickey-Fuller for stationarity (rejects unit-root → mean-reverts), Herfindahl-Hirschman for concentration. EDGE signal fires when at least two agree.
Tests whether two games share a long-run statistical relationship despite each individually being non-stationary. P3↔P4 = 0.71 is the strongest pair — both 0–9 pools, twice-daily. The AI engine uses cross-game ρ to confirm signals from same-pool siblings.
Engle & Granger (1987) · Johansen (1991)
07 Dashboard Preview
Step inside the terminal.
A live snapshot of the analytics dashboard. Every panel below is rendered in real time inside the actual web app — no images.
DEEP ANALYTICSSCRATCH ENGINEDRAW LABAI BACKTESTERKELLY OPTIMIZERRETAILERS
LIVE
FLORIDA LOTTERY · DEEP ANALYTICS · STATISTICAL LAB
Powerball — 1,803 draws
Last sync · Apr 28, 2026 06:00 ET · WHITE-NOISE FOOTPRINT, mean-reverting bias
HURST H
0.49
mean-rev
χ²
1068.87
df=68
ADF
-3.46
stationary
HHI
0.018
low conc
SKEW γ₁
+0.04
symm
KURT γ₂
-0.18
platy
Frequency Distribution · all 69 main balls
Long-history mean reversion · normalized R/S
08 How it works
From raw draw to actionable signal.
Four steps. Every morning at 06:00 ET, on autopilot.
01
INGEST
Pull every draw + every retailer.
Our edge-worker pipeline ingests Florida Lottery data for the latest 14 days of draws across all 10 games, plus winner-claim PDFs for retailer geofencing. 75,408 rows live in our cloud database.
02
VALIDATE
Cross-check & deduplicate.
Each row passes a draw-date validator and ball-pool sanity check. Duplicates are merged. Failed rows go to a quarantine queue for next-day reprocessing.
03
COMPUTE
Run the 12-metric battery.
CAPM α, Kelly f*, Sharpe, Hurst, χ², ADF, HHI, Skew, Kurtosis, Z-Score, Mean Deviation, and pairwise Cointegration computed for every game. Persisted with timestamps.
04
DELIVER
Stream to your terminal.
Edge-cached responses hit your dashboard in <200ms. Top plays, EDGE/HOLD/FADE signals, AI combos, Kelly stakes — all updated and ready before you log in.
09 Modules
Eight modules. One terminal.
Every feature you'd build if you were running a quant desk on Florida Lottery data — already built.
01
Scratch-Off EV Engine
All 88 active scratch games scored on EV per dollar, EV Yield (positive %), ROI (true), and CAPM α. TOP TIER / MIXED / SKIP badges drive a sortable cover-page table you can act on in seconds.
88 gamesEV / Yield / ROI / αDaily refresh
02
Statistical Lab (Deep Analytics)
12-metric battery — Hurst, χ², ADF, HHI, Skew, Kurtosis, Z-Score, % from Mean, plus α / f* / Sharpe / ρ / E-G. Per-game cards, a live feed table, and the cross-game research matrix.
12 metrics10 draw gamesCross-Game Matrix
03
AI Backtester
8 orthogonal signals × 6 strategies × 3 risk profiles = 144 distinct configurations. Run any combo against the last N draws and see total spent, won, net, win rate, ROI, best hit.
8 signals6 strategies3 risk tiers
04
Daily Shopping List
One printable PDF every morning at 06:00 ET. Budget-aware allocation across scratch and draw games — set $25, $50, or $100 cap and the wizard splits it. AI Analyst Notes explain why each pick sits where it does, with discipline check + i.i.d. compliance line baked into every print.
Budget capScratch + Draw1-page print
05
Hot / Cold / Due Lab
Frequency, gap, and overdue analysis across all draw games. Three philosophical lenses: Hot Bias (ride momentum), Contrarian (fade), Due (chase gaps). Z-score and percentile context on every ball.
3 lensesZ-contextPer-game
06
Strategy Center
Pre-built combo lists per strategy (Hot Bias, Contrarian, Due, AI Balanced) with descriptive performance stats from the rolling backtest window. One-click apply to Kelly.
4 strategiesTop-6 combosOne-click
07
Retailer Geofencing
FL Lottery winner-PDF parser indexes 13,432 retailer records by ZIP code. Heat-map of recent claims, top-prize hot zones, walk-up filter for your address.
13K+ recordsZIP geofenceHeat-map
08
Worker Health
Live ingest pipeline monitor — last scrape time, success rate, P99 latency, error log. Transparency on the data layer that powers every signal you act on.
Cron 6AM ETP99 <200msPublic log
10 Pipeline
Built on the open web.
Edge-deployed, always-on, no servers to babysit.
SOURCE
FL Lottery
flalottery.com APIs
3× DAILY
→scrape
COMPUTE
CF Worker
edge nodes · Workers KV
06:00 ET
→persist
STORE + SERVE
Cloud DB
Postgres · 150K rows
REST + RT
P99 < 200ms response
99.94% 30-day uptime
100% edge-cached reads
3× daily cron schedule
11 Execution
Built for serious capital deployment.
A research-grade tool for people who treat lottery as a statistical discipline, not a daydream.
"Finding asymmetric risk in scratch-offs."
01
The Value Hunter
Utilizing the Edge Score
Focuses entirely on the EV Yield column. Waits for a $50 ticket to reach 75%+ EV Yield due to top-prize stagnation, then deploys capital when the state's mathematical edge is at its absolute weakest.
"Automating portfolio stake sizing."
02
The Syndicate
Utilizing the Kelly Optimizer
Replaces emotional spending with the Kelly Criterion. Inputs a $500 session bankroll and lets the algorithm dynamically divide the cash across 4 different games based on their risk-adjusted Sharpe ratios.
"Trading cross-game frequency momentum."
03
The Quant
Utilizing Cointegration Matrices
Uses the Engle-Granger matrix to find highly correlated draw games (like Pick 3 and Pick 4). If ball #7 is burning hot in Pick 3, they use the statistical linkage to predict momentum spillover into Pick 4.
12 Pricing
Start free. Stay free if you want.
Try the full terminal in Guest mode in one click — no card, no email. Upgrade when you want the AI Backtester and the Kelly Optimizer. Subscription fees are for access to the analytics platform only and are not lottery purchases.
No, and any tool that claims otherwise is lying. The lottery is statutorily a game of chance — every ticket is negative-EV by design. What we do is surface the games where the published return-to-player rate is highest and where statistical observations diverge from uniform expectations. That's research and analysis, not a system that beats the lottery.
How do I actually buy these picks?
Every morning at 06:00 ET the platform generates a Daily Shopping List — a one-page printable PDF with your scratch and draw picks allocated to your budget. Set $25, $50, or $100 cap; the wizard splits between scratch and draw based on EV ranking. AI Analyst Notes explain each pick's rationale, with the i.i.d. compliance line baked into every print. Take the page to any FL Lottery retailer.
Why does Hurst matter for lottery?
Hurst measures whether a series mean-reverts (H<0.5) or trends (H>0.5). Most FL games sit at 0.48–0.51 — essentially random — but slight anti-persistence in Pick 3, 4, and 5 (H ≈ 0.48) means frequency drift corrects itself. That's the statistical justification for "due number" plays in those games.
What's the difference between EV Yield and ROI?
EV Yield = EV ÷ ticket cost (a positive %, e.g. $35.77/$50 = 71.5%). ROI = (EV − cost) ÷ cost (typically negative for lottery, e.g. ($35.77 − $50)/$50 = -28.5%). Both are correct — they answer different questions. Earlier versions of the dashboard sometimes labeled Yield as ROI; this is now fixed and the columns are separated.
Why are the Sharpe ratios so small?
Sharpe = (EV − Rf) / σ. Lottery σ is enormous because of the tail of rare large prizes — so realistic Sharpes are tiny (0.02 – 0.20). A "hedge-fund elite" Sharpe of 2+ is not achievable with negative-EV tickets and any earlier display showing values that high was a bug. The ranking is what matters: 0.18 is still the best in the cohort, even if the absolute number is low.
Where does the data come from?
Florida Lottery's public APIs and winner-claim PDFs. Our proprietary ingest pipeline scrapes them on a 06:00 ET cron, validates each row, and persists into a Postgres-backed cloud database. 75,408 draw rows indexed across 10 draw games, plus 88 active scratch games and 13,432 retailer records. All metrics are computed server-side — your terminal just reads the cached results.
Is this affiliated with the Florida Lottery?
No. LottoLogic IQ is an independent third-party analytics tool. We are not affiliated with, endorsed by, or sponsored by the Florida Lottery, the FL Department of Lottery, or any state agency. All data is sourced from publicly available APIs. Lottery games are gambling — please play responsibly. Florida Council on Compulsive Gambling 24/7 helpline: 1-888-ADMIT-IT (1-888-236-4848).
The lottery is luck. We bring the math.
Open the terminal in Guest mode in one click. See the numbers. Decide for yourself.