New here? Going Yard ranks today's hitters by their recent power form against the pitcher they face. Tap any row for the full breakdown.
A daily board that ranks hitters on how likely they are to go yard, using free Statcast + MLB data. Everything updates itself through the day — no input needed. It's a research tool, not betting advice.
That last point is the whole thing. A great hitter at a bad price loses money. A mediocre hitter at a great price makes money. The app exists to tell you which is which — the Heat number finds the hitters, the price decides the bet.
Most days, the honest answer is "no bet." Books price home runs pretty well. Passing is a decision, not a failure.
Five tabs across the bottom: Board (every hitter ranked by Heat), Scan (the slate by game — tap a game for its arms, hitters, weather, team stack and a +EV cross-check against the model), Edges (the experimental signals), Props (hits, HRR, pitcher Ks, +EV, prop parlays, game lines) and Bets (the Edge Finder, your logged bets, and the jackpot cards).
Tap More for the full navigation drawer, which lists every view grouped by what you're trying to do:
DAILY — Board, Scan, Props, Bets. TARGETS — Vulnerable Arms (HR Vuln scores + zone maps), Quick Target (dangerous lineup spots per arm), Bullpens, and Parlays & Jackpots. CONDITIONS — Weather & Parks (wind arrows over each park plus tonight's best/worst HR parks), Microclimate (temperature-sensitive hitters tonight) and Late-HR Spots (short starter meets gassed pen). RESULTS — Tracker and this help page. If you can't find a feature, it's in the drawer.
Heat (0–100) ranks each hitter on six power signals over the last 2 weeks — Pull-air%, Barrel%, ISO, Exit Velo, Ideal Attack Angle%, SLG — then nudges for the opposing pitcher's vulnerability. Higher = better HR spot. The chip runs blue (cold) → orange (elite).
The four cells (Pull / EV / Brl / IAA) show those signals at a glance — orange means it cleared the "good" mark, blue means below. Tap any column header (HEAT, PULL, EV, BRL, IAA) to re-sort the whole slate by that stat — that's how you find a low-Heat guy who's elite in one area.
Tap any row to open full detail: recent/season/career for every signal, the trend windows, the opposing arm, bullpen, pitch matchup, and luck.
The pill next to a hitter's name isn't about sample size — it counts how many of the six signals cleared their "good" anchor. LOADED = 5–6 cleared, STRONG = 4, SOLID = 3, LEAN = 2, THIN = 0–1.
Heat tells you how hot; the tier tells you how many independent things agree. A 68 Heat that's LOADED is a hitter where nearly everything is firing. A 68 Heat that's LEAN is one or two signals carrying the whole score — same number, more fragile. Use it as a breadth check before you commit.
Thin-data warning. A signal can be missing rather than just low — a just-called-up or returning-from-injury hitter may not have data for every metric yet. When fewer than all six have real numbers, the score fills the blanks with a cautious "poor" baseline, and the card shows a Note: thin-data warning. Treat that Heat as softer than it looks — it's resting partly on absence of evidence. This is why two hitters at the same Heat can be very different bets.
When a hitter clears a high bar on multiple fronts at once — strong heat, several confirming reasons, a genuinely vulnerable spot — he gets flagged SMASH and appears in the morning briefing. It's deliberately rare (top few per slate). Treat it as "the model is unusually confident," not a guarantee. Still accumulating sample; check the Tracker before sizing anything off it.
Robbed flags hitters who hit a ball in the last 3 days that would have left most parks — a 400+ ft drive that died at the wall or got caught. It's a "the process was right, the result wasn't" signal. The ROBBED ×N badge shows how many times. Tap the card for the spray chart.
Milestone Watch surfaces hitters sitting near a round-number career HR mark (100, 150, 200…) or within a few HRs of passing a franchise leader. Not a model signal — a context signal. Some bettors like the narrative spot; the model doesn't care either way.
Star any hitter to add him to My List, then tap the ★ My List chip to filter the board to only those hitters. The list auto-clears when the slate date changes — yesterday's picks never bleed into today. A Clear button appears next to the chip whenever the list has anything in it, if you want to wipe it early.
R / L / S (orange/blue/purple) = a hitter bats right, left, or switch; pitchers show RHP / LHP in the same colors. The trait badges flag why a hitter is interesting — matchup first: WEAK ARM (facing a shellable starter), PLATOON (strong HR history vs this arm's hand), PITCH EDGE (barrels the pitch this arm throws most), POWER (top-end exit velo, 112+ mph).
Then one form read per hitter — a simple spectrum so you never see a contradiction: DUE (hitting it hard but unlucky — results lag his contact), WARMING (contact quality trending up), HOT (elite contact and results match — genuinely hot, earned), MAY COOL (results have run well above his contact and may regress). ▲ / ▼ = power-form trend with % change. proj = lineup not posted, using last game's order. #1–#9 = batting-order spot.
Heat / Trend / BP / Bomb / Park / Name / Time sort the list — BP ranks by how vulnerable the opposing bullpen is to each hitter's hand, Bomb ranks by raw long-ball ceiling (season max exit velo nudged by recent form; the "NNN EV" chip shows the ceiling), and Park ranks by today's park & weather HR boost (the chip shows the % swing plus wind).
Edge filter chips ( Elite · Power+ · ◆ Zone edge · Climate+ · ◎ Mix+ · Late-HR) narrow the board to hitters carrying specific edges, and stack with AND logic — tap several to drill to exactly the spot you want. They're dashed until active, solid orange when on, and clear with "Clear all."
A separate, physics-based read — its own lens that never touches the Heat score. It flies each hitter's actual batted balls (exit velocity, launch angle, spray direction) through the real wall distances and heights of today's park, under today's temperature and wind, and counts how many clear the fence versus a neutral park. That gives a per-hitter boost, not a flat park factor: a dead-pull hitter gets more from a short porch on his side, and wind/heat scale with how often he already lives near the wall. Tap any row for the full read — expected HR per 100 batted balls here vs neutral, plus the conditions used. Two honest limits, shared by every public version of this (Ballpark Pal included): batted-ball backspin is estimated from launch conditions, not measured, and closed retractable roofs can't be detected, so treat it as a strong secondary signal — and for the final per-hitter precision, Ballpark Pal is still the authority.
The 30-park read. On the same expanded card, a 30-PARK block takes the hitter's recent deep contact and flies it through all 30 MLB parks' walls under neutral air — pure geometry, no weather. It tells you how homer-friendly his contact is park to park: a guy whose clearing balls average 28 of 30 parks is "out almost anywhere," while one averaging 8-12 "needs a friendly park." It also shows how many of his deep balls clear tonight's park specifically, and how many are park-dependent (out in some parks but not others — the ones where the venue decides his night). Same principle as the boost above, but framed as fit rather than a percentage — useful for spotting a hitter whose power plays up or down badly in tonight's building.
In Filters: Games and Teams narrow to specific matchups or clubs, Heat floor (40/55/70+) cuts the long tail, Opposing starter drops strong SP so you only see hitters facing a vulnerable one, Bats shows one handedness, and Traits lets you tap any badge to show only hitters carrying it (tap several to stack them). Every filter combines, so you can stack (say) a vulnerable starter + weak bullpen sort + a trait + a team at once. Hide small samples drops hitters under ~25 PA in the window or whose score leans on missing signal data (fewer than 4 of the 6 signals have numbers).
Top Plays (panel up top on the Board) is your fast scan, now grouped by angle so you can see why each hitter is here. Categories: SMASH (everything converges), ◈ ZONE (HRs in this arm’s meatball zones — the #1 signal), Graded (4+ matchup factors aligned), Bomb Score (best composite matchups), plus Top Heat, Mix Edges, Owns the Arm, Risers, Complete and Due. Every row carries its signal chips — matchup grade, ZONE count, plate-discipline icons and Bomb Score — so you can judge a play without opening the card.
The whole slate grouped by game, each scored 0–100 for how much edge/action it holds. Chips flag what's inside at a glance: hot hitters, +EV plays, a vulnerable arm, a gassed pen, climate boosts, late-HR spots, a high total, warm weather, wind blowing out. Tap a game and you get everything for that matchup in one screen: the arms ranked by vulnerability, the top hitters with their edge tags, a full weather & park block (temp, wind interpreted as helps/suppresses HR, precip, humidity, roof), the team stack signal (which side has the denser cluster of hot bats), and the best +EV plays. This is where the old Weather and Stacks views now live.
The experimental signals — spatial and situational matchups the market doesn't price directly. Four sub-views:
Arms — today's starters ranked by zone-matchup exploitability: the average xwOBA the top opposing hitters post in exactly the cells this arm throws to most (only hitters with real sample count). This is a different lens than "Arms Today" on the board, which ranks by the season HR-vulnerability model — they measure different things and won't match. Tap an arm for its interactive zone map: cells 1–9 colored by where he lives. Tap a hitter below and the map repaints to show that hitter's xwOBA in each zone vs this arm, with barrel% and distance, and the pitcher's usage % in the corner of each box. ◆ = he crushes it and the pitcher throws there (the danger overlap); ● = he crushes it elsewhere. A number badge by each hitter's name = how many zones he crushes.
Pens — every bullpen on the slate ranked most-exploitable to toughest, built on pen ERA, season HRs allowed (HR/9), how worn down the pen is (recent pitches thrown + arms unavailable) and its platoon split (which hand it gets crushed by), plus a supporting statcast contact-quality read. Each row shows those numbers and a bar of the five components with the biggest weakness highlighted. Micro — hitters flagged by tonight's temperature vs their personal temperature sensitivity, ordered by actual impact tonight (sensitivity × how far tonight's temp is from their median). Late — games where a short starter meets a gassed pen, raising late-inning HR expectancy.
Every hitter's card now carries two HR-specific lenses, both separate from Heat. Zone profile — a mini 9-cell grid lighting up the zones he crushes (xwOBA ≥ .400) vs today's arm, so you see his damage zones without opening Edges. HR power profile — his raw over-the-fence contact: barrel%, HR-swing% (how often he takes a home-run-shaped swing), average fly-ball distance, and max distance. Flagged PLUS when the power is strong. Barrel rate and carry are the leading HR indicators, shown plainly.
An ELITE / ELITE+ chip appears by a hitter's name app-wide when he clears the four headline thresholds (pull-air ≥33%, EV ≥90, barrel ≥10%, plus ideal attack angle for the +). On the Board, the edge filter chips ( Elite · Power+ · ◆ Zone edge · Climate+ · ◎ Mix+ · Late-HR) let you narrow the slate to hitters carrying specific edges. They stack with AND logic — tap several to drill to your exact spot (e.g. Elite + Power+ + Zone edge = only the hitters clearing the gate who also have plus power and crush a zone their pitcher lives in). None of these touch Heat.
Arms Today ranks starters by how shellable they are. Form labels: SHELLABLE (bad & getting worse — prime), STEADY-BAD (consistently hittable — still a target), SLIPPING (was fine, now cracking), BAD-IMPROVING / BOUNCING-BACK (downgrade — numbers improving), DEALING (avoid). Arm badges spell out the specific vulnerability behind the label: HR PRONE (gives up bombs), BARRELED (barreled ≥10%), LOUD (hard contact/high EV allowed), PULL-AIR (pulled fly balls allowed), WEAK vs R/L (which hand's taken him deep over 2yr — target those hitters), VELO ↓ (velo trending down), SLIPPING (recent worse than season), TOUGH (low vulnerability — avoid). Platoon shows whether the arm gets hit harder by RHB or LHB, and the score vs your hitter's hand. Opposing bullpen shows the pen's HR-vulnerability and its split vs your hitter's hand.
Tap any arm to expand it — you'll see every batter facing him tonight, ranked by Heat, plus a link to filter the board to just those bats. The Sort toggle at the top of the panel switches between "By vulnerability" and "Most hittable." The expanded view also shows a HRs allowed by hand table with two windows: YTD (this season only) and 2szn (this season + last, regular season only). Reference tools like PropFinder default to season-only, so if a number ever looks off, check which window you're reading — they're both there on purpose.
vs Pitch Mix toggle (top controls): re-ranks the board by each hitter's EV/launch-angle/whiff weighted to the opposing starter's pitch mix over the last 2 weeks — a guy who crushes fastballs rises against a fastball-heavy arm. Toggle off to return to standard Heat instantly. Barrel% stays overall. Each card shows the mix-adjusted EV and how Heat shifts.
Pitch matchup = the hitter's barrel% by pitch family (Fastball/Breaking/Offspeed) crossed with the arm's usage. Contact luck = expected (xwOBAcon) vs actual; a positive gap means he's hitting it well but unlucky (DUE). Overperforming only flags MAY COOL when results run far above his contact; if the underlying contact is itself elite he reads HOT — earned, not lucky. Starter HRs by hand = actual HRs the starter allowed to RHB vs LHB, over a 2-year sample and this year. None of these affect Heat unless the mix toggle is on.
A panel on the board that polls MLB's live feed every 90 seconds while games are on. Every home run that lands shows up with the hitter's full board identity from this morning — his Heat chip, confirmation tier, the exact badges he carried, his board rank, all three prop scores, plus distance and RBI count. A ★ means he was on your list. Hitters who homer but weren't on the board show as "off board."
The point isn't the scores — it's the cross-reference. Over a night you can see whether the model's picks are the ones cashing, in real time, rather than waiting for the grader next morning. The dot pulses red while games are live; polling pauses when you close the tab and stops entirely once every game is Final.
Four prop models that run completely separate from the HR heat model — different signals, different scores, their own tracker and backtest. HR heat is never touched by any of them.
Hits — ranks hitters for 1+ / 2+ hits using xBA, hard-hit%, line-drive%, plate discipline (BB% minus K%), contact%, and the opposing arm's hittability. HRR — hits + runs + RBIs, which is hit skill plus a lineup-spot multiplier (top of the order gets more PAs and more RBI chances) plus an HR-upside boost, since one swing can produce all three. Pitcher Ks — ranks starters for their K-total prop using their own K% and SwStr% blended over 2 weeks, weighted against the opposing lineup's K vulnerability. Each card shows an estimated K total and target line ("est. 7.8 Ks · line target O7.5") so you can compare directly against the book before opening the odds screen. Openers get an automatic downgrade — they can't throw enough innings to clear a normal line.
K Under — sorted low to high on purpose. The lowest K-prone scores are the best "UNDER 0.5 strikeouts" bets: contact hitters facing low-whiff arms. Tags run STRONG UNDER (under 30) → LEAN UNDER → NEUTRAL → FADE (60+, the K-prone guys you'd avoid).
All four use the same 2-week trailing window as the HR model, with the same thin-sample discount on both the batter and pitcher side.
The most useful thing the backtest found is that signals compound. Hitters with no convergence families homer at 8.2%; with two families 14.5%; with three 17.0%; with four 21.2% — nearly double the 10.9% base rate, measured across 30,000 player-games. That ladder is what makes leg selection worth doing.
Two routes. Same game (SGP) — three hitters from one lineup; the legs move together, so a blow-up rescues all three at once. This is the only version where correlation helps, and correlation is the only thing that narrows the gap to a book price. One per game — three hitters from three matchups; each leg is individually stronger because you take the best spot on the slate rather than the third-best in one lineup, but nothing saves the ticket if a game goes quiet. Route A is the better bet in theory; Route B is easier to fill honestly when only one or two hitters qualify per game.
Board buttons, in order. Tap ▣ 250+ BBE first — the same convergence signal reads 16.1% on 250+ batted balls and 9.1% under 120, so this chip alone separates a real read from a coin flip. Then ◎ 4+ families, then ◇ 2+ near miss. For a same-game ticket, add the game chip for your matchup. If fewer than three hitters survive, drop to ◎ 3+ families; if still short, change games rather than filling the ticket with a name that did not qualify.
The method. 1. Pick the game first — a vulnerable starter in Edges → Arms, a worn pen in Bullpens, a park that carries in Weather & Parks. 2. Filter the board to that game and tap ◎ 4+ families; drop to 3+ if nobody clears it, and change games if nobody clears 3. 3. Rank what's left by barrel% 12+ (19.5%, the strongest single metric), heat 70+ (16.4%), 2+ near misses (12.2% vs 9.3% for none), then batting spot 1–4. 4. Check sample size — a convergence read on under 120 batted balls hits 6.5%, worse than random, while the same signal on 250+ hits 16.1%. 5. Pair hitters batting close together (2-3-4 or 3-4-5) so they face the same pitcher in the same inning. 6. Confirm the four families are genuinely different kinds of evidence, not one signal counted four times.
What it's worth. Three legs at 21.2% is 0.95% if the legs are independent — fair odds around +10,400. Books pay near +2,500, so the math still says no. What the process buys you is roughly 7× better odds than picking three hitters at random, and steps 1 and 5 exist to create the same-game correlation that is the only real argument for the ticket. Bet it because you want the shot, not because it's +EV.
In the Props tab → Parlays sub-tab. Builds multi-leg parlays from tonight's priced +EV prop legs (hits, HRR, pitcher Ks, HR, game lines). Three cards: Safest 2-leg (the two likeliest props to both hit), Best value 3-leg (highest-EV legs that are each still genuinely likely), and Longshot 3-leg (bigger payout, only genuine edges). Each shows the combined odds and the honest combined probability so you see the real risk, not just a flashy payout. Legs are spread one-per-game to avoid correlation. Parlays are structurally −EV for most bettors — these only surface when the individual legs each carry an edge, and even then the combined chance is small.
In the Parlays view (More → Parlays). The 3 hitters most likely to hit a grand slam tonight, ranked by traffic × punish: traffic = how likely the bases are loaded when he bats (the on-base bats ahead of him + the opposing starter's wildness), punish = can he golf a grooved in-zone fastball (barrel/EV/pull + the elite gate). A slam needs both — a masher batting leadoff scores near zero. This is the highest-variance thing in the app: it ranks relative likelihood, not a real probability. A lottery ticket, sized accordingly.
In Bets. The slate ranked for the single longest home run — but not on raw distance alone. Each hitter gets two scores that are blended: distance (how far if he connects — career-max ceiling over the last two seasons, plus fly-ball carry, barrel and exit velo) and convergence (how many proven-lift signals agree tonight). Convergence rewards signals that actually moved HR outcomes in the backtest and your tracker — power badges, zone-fit edges, heat band, barrel band, matchup grade and park factor — each weighted by its measured lift. So a 470ft ceiling with a POW badge (real +76% lift), a zone edge and a hot park outranks a 470ft ceiling with nothing corroborating it. Each row shows the final score, the distance/convergence split, and the converging signals with their measured lift %. A row tagged “ceiling only — no signals converge” has the raw pop but nothing backing it up tonight.
The per-hitter grade for today's specific matchup, by factor convergence rather than a weighted average. Five factors either align or they don't: Power (ISO — HIGH at .200, ELITE at .250), Zone (3+ overlap dots viable, 5+ premium), Arm (the opposing pitcher's HR Vulnerability tier), Park (1.05+ boost, 1.10+ premium), and Form (bat trending up). 4+ aligned = STRONG · all five with premium levels = ELITE. The card shows a five-cell grid of exactly which factors hit. Plate discipline nudges the grade one step (elite eye up, chaser down) but never inflates a spot with nothing else going for it. Sort the board by Grade, or use the Graded filter for ELITE/STRONG only.
The composite batter-vs-pitcher matchup score, 0–100 — this replaced the old raw max-exit-velo "Bomb" sort. Combines ISO, SLG, zone overlap (weighted heaviest), park factor, platoon edge, the opposing starter's ERA, a hot-bat bonus, and the pitcher's times-through-order fade. ≥55 high-priority, ≥65 elite. Every component's 0–100 subscore is shown on the card so you can see what's driving the number.
The big badge on each arm in Edges — a 0–100 read on how susceptible that pitcher is to giving up a home run today. Point budget: ERA 30 · park 20 · HR-splits by batter hand 15 · WHIP 15 · zone damage 12 · dangerous bats faced 8. ≥70 = ELITE TARGET, ≥50 = STRONG. Tap an arm for the full breakdown — a bar per component with its earned points and a plain-English note. HR splits use the pitcher's worse handedness side, since that's the exploitable one. If ERA/WHIP aren't available for a starter, the score rescales over what could be measured and tells you what's missing.
Tap any arm in Edges for its full card: the VULN badge and tier up top, then ERA / WHIP / K-9 / BB-9, season HR allowed, the GB% / FB% split (a high fly-ball rate is a HR red flag), and the contact quality he allows (AVG EV / BRL% / HH%). Below that: HR allowed by batter hand over two seasons — target the side he bleeds — and HR HOT SPOTS, the specific hitters who have taken him deep. The pitch mix cards show each family’s velocity and usage, with the fastball flagged VULN when he’s fly-ball prone. Then the interactive zone map and the batters who punish him, ranked by the ZONE signal.
A 0–100 quality-of-contact score: how consistently a hitter squares the ball up. Built from Sweet Spot%, Barrel%, Hard Hit%, and average exit velo. ≥75 Elite · ≥60 Strong · ≥45 Average. Separate from both Heat and HR power — it measures contact quality, not home-run shape specifically.
Three icons next to a hitter's name, from chase rate and in-zone contact percentiles: Eye (chase rate in the top 25% — rarely chases, forces pitchers into the zone), ⊕ Crosshair (in-zone contact in the top 25% — squares up strikes), Note: Warning (bottom 25% chase rate — swings at everything). Eye and Crosshair together is the top-tier confirmation. Percentiles are true MLB percentiles computed over every qualified hitter in the season pull. These now adjust HR probability in the parlay math — Eye ×1.06, Crosshair ×1.05, Warning ×0.94.
A Savant-style bar chart in the expanded card ranking the hitter against every qualified MLB hitter — not just today's slate — across nine categories: Exit Velocity, Barrel%, Hard Hit%, xwOBA, Walk Rate, Chase Rate, Whiff%, Zone Contact% and K%. Computed from the full season pull, so these are real league percentiles. Chase, Whiff and K% are inverted so higher is always better. Red/orange across Exit Velo, Barrel% and Zone Contact% together = an elite power-contact hitter.
In Edges — the batting-order spots that have done the most damage against each arm (HR + ISO + SLG at that slot), cross-referenced with who's actually hitting there today. Every pitcher bleeds more to certain spots in the order; this surfaces which, so you can check the lineup and target the hitter standing in that spot. Tap a name to open their card.
ZONE 5 means this batter has hit 5 home runs in the exact zones this pitcher is most hittable in. Not contact quality — actual home runs, landing in the pitcher's meatball cells. Meatball zones are where a pitcher gives up the most damage (highest xwOBA allowed — his worst few cells). 3 is the minimum threshold · 5+ is premium · no badge means no matchup. On the card's grid, amber cells mark the overlap with the HR count on them; on the big Edges heatmap, ◆ marks a HR hit in a meatball zone and ● a HR elsewhere, each labeled with how many. HRs over the heart of the plate carry more confidence than corner overlaps.
Today's game condition (from MLB's official day/night flag) checked against the hitter's own historical day-vs-night splits. The card shows both splits and whether tonight's condition favors him, with the OPS gap. A meaningful gap in the favorable direction is a confidence add; splits only appear when there's enough sample on both sides.
Grades every model against reality each morning. Model vs baselines is the one that matters: does Heat's top 10 homer more than just sorting by ISO or barrel%? If yes, the model is genuinely finding HR hitters better than the simple stats. The rest shows HR rate by Heat tier, by arm form, top-N success, which signals correlate to HRs, and a daily log.
Parlay performance shows how the auto-generated parlays actually converted — the leg hit rate is the honest number; if it isn't beating the base rate, that strategy isn't earning. Props performance does the same for hits, 2+ hits, HRR, pitcher Ks, and K Under, each with its own base rate derived from your own data (never a hard-coded league average).
Historical replay checks how model scores relate to recorded outcomes. It uses actual participants and starters, which were not always known before games. Older exports may also contain full-season inputs. Treat these results as exploratory; they do not establish a betting edge.
Checking the estimated chances. Historical outcome rates can inform a starting estimate. They do not validate the final estimate after matchup adjustments. The stronger check is predictions saved before games and graded afterward, across enough future games to see whether their stated chances hold up.
Rows marked thin and dimmed mean the sample is too small to trust yet (under 30 for hitters, 20 for pitchers). The number is real math, but it could easily be noise. Wait for it to fill in.
The HOW TO BET button in the header opens a guide that rewrites itself from your own tracker data. It sorts every signal into buckets — use (big enough sample, real lift), tiebreak (real sample, thin edge), watch (interesting but not enough data yet), fade (actually negative), ignore (too thin to say). As your sample grows the guide changes. It is not static advice; it's whatever your data currently supports.
The tab that turns the model into a decision. Every hitter shows the model's probability and the worst price you can take and still break even. Type in what your book is posting and it computes the expected value instantly, then sorts everything by it.
Why this exists: beating the base rate is not the same as making money. A hitter the model likes at 17% needs +488 or longer to break even. Books post home run props at +300 to +450. So a 17% hitter at +400 is a losing bet — you'd be paying 20 cents for something worth 17. Meanwhile a 12.5% hitter at +800 is profitable, because the book is paying you for 11% when the true odds are 12.5%.
Price research compares listed prices with model estimates. It shows every matching candidate, labels large disagreements for review and suggests no stake until the final probability model has been validated on predictions saved before games.
That inversion is the entire game, and it's why the hottest hitter on the board is frequently the worst bet on the board. Prices you enter clear automatically each new slate.
The caveat that matters: a positive EV number is only as trustworthy as the model's calibration. Until the Tracker shows that a stated 17% actually converts near 17%, these are candidates, not green lights.
Your private bet log — stored only on this device, nothing leaves your phone. Log each bet, pick the player(s) from today's board (their badges and Heat are captured automatically), then after games mark each as HR ✓ or no HR and set the result. Over time it shows your record, net, and ROI — plus the payoff feature: HR rate by badge, so you learn which badges actually convert for you (does HOT homer more than MAY COOL?) and which are noise.
Auto-grading. While games are live, the app watches the home run feed and fills in your legs for you — when a hitter you bet goes deep, his leg flips to HR ✓ with a small auto tag so you know it came from the feed, not your own tap. You can always override it. Players who get scratched from the lineup are flagged SCRATCHED — at most books an HR prop on a scratch is voided and refunded, so that leg probably doesn't count.
Closing Line Value (CLV) — the number that actually matters. After a line closes, type the closing price into any logged bet. CLV measures whether you got a better price than where the market ended up. If you took +500 and it closed +400, you beat the close — that's positive CLV, and it's the single most reliable sign your edge is real. Here's why it matters more than wins: you can lose a bet you were right to make, and win a bet you were wrong to make — over a small sample, win/loss is mostly luck. But consistently beating the closing line can't be luck. If your CLV is positive across a few dozen bets, the profits will follow even if your early record looks rough. If it's negative, the market is telling you your reads are off, no matter how many you happen to win. Watch this number more than your record.
The dot by the title shows build health: green = fresh, amber = a few hours old, red = likely stale or showing an older slate. The board builds at 6:30am, 10am, 12pm, 3pm, 5pm and 7pm ET; if one is missed or you're viewing yesterday's slate, a banner warns you before you bet.
The board rebuilds through the day (6:30am, 10am, 12pm, 3pm, 5pm, 7pm ET) as lineups post; the grader runs each morning after games are final. Expanded cards show a heat trajectory sparkline — the hitter's last 10 days of Heat, with dashed lines at the 55 and 70 tier marks — so you can see the shape of a trend, not just an up/down arrow. Park factor is shown for context but isn't baked into Heat — the data and matchup drive the ranking.
Every model here is honest about what it doesn't know. Thin samples get flagged, new models get labeled as unvalidated until the tracker and backtest say otherwise, and none of it is betting advice.