GhatGPT prompts for sports betting

Should have clarified this.

I have instructed it to NOT use betting lines found online or social media.

Strictly data driven from sites like Baseball Savant and Baseball Reference

eta: is it still pulling from random social media accounts? Who knows, but I find it spits out a lot of data.

I refer to them as Clankas

quote:

---

Core Moneyball Prompts

“Moneyball me a parlay” ? Build a single-game parlay using pitcher vs. batter matchups, splits, and trends.

“BUILD AWAY MONEYBALL” ? Same as above but triggered when you give me a specific matchup.

“I want a 3-pick parlay around even odds” ? Focus on runs, team totals, strikeouts; avoid HR props unless specified.

“I want a 4-pick HRR prop” ? Hits + Runs + RBIs prop parlay, usually with favored or hot teams/players.

“Pull starters and lineups” ? Confirm listed pitchers and probable batting orders.

“Look at batter heat maps AND pitcher splits” ? Advanced matchup data, left/right splits, wOBA, OPS, K%, etc.

“How does [team] match up against [pitcher/team]?” ? Full breakdown of lineup vs. opposing starter.

“Team run totals / alternate runs” ? Evaluate realistic scoring totals based on averages, pitching, and ballpark.

---

Nerds have ruined baseball already, now they are trying to ruin betting.

The best sports bets are made on vibes.

quote:

---

Anyone else tinker with this?

---

I have extensively, for both MLB and WNBA.

I use a custom GPT for sports betting that is supposed to help narrow the focus to "legitimate" sources.

The first few times I used it in-depth, it was 67% on picking one player from each game to get an RBI for example, and was pretty good on prop bets in general.

Lately, like you mention, it will get the starting pitcher wrong (repeatedly), or list players off different rosters, even after I correct it multiple times.

The last time I tried it with WNBA, it was recommending ATS picks based on the anticipated performance of players who aren't even in the league.

I'd almost swear that AI is getting stupider.

quote:

---

parlays

---

Here is what it spit out for the Cubs @ Giants game this afternoon

quote:

---

Got it — Cubs (Shota Imanaga) @ Giants (Logan Webb). Let’s Moneyball this matchup.

Here’s the quick framework we usually roll with:

Pitcher splits (how each starter has fared lately + vs opponent type)

Batter matchups (who looks good against that type of pitcher)

Team scoring trends (runs per game, recent momentum, ballpark factor)

Possible parlay legs (hits/runs/RBIs, K props, team totals, alt totals)

Pitchers

Imanaga (CHC)

Rookie sensation, still carrying a sub-3.00 ERA late in the year.

Strikeouts: tends to sit in the 6–8 K range, but the Giants are middle-of-the-pack in K%.

Weakness: occasional long ball, especially vs right-handed bats.

Webb (SF)

Ground-ball machine, tough to homer off.

Lower K totals (usually 4–6 per outing), Cubs’ lineup doesn’t whiff a ton outside of a couple guys.

Can get nickeled-and-dimed by singles and doubles if BABIP runs against him.

Batter Trends

Cubs:

Nico Hoerner: makes a ton of contact, sneaky hit/RBI play.

Seiya Suzuki: hot bat lately, matches well vs sinker/GB guys like Webb.

Cody Bellinger (if in lineup): good against RHP, though Webb suppresses HRs.

Giants:

Wilmer Flores: one of their better righty bats against lefties.

Jung Hoo Lee (if healthy/playing): contact-first, can slap a hit.

Matt Chapman: power/swing-and-miss, but could connect off Imanaga.

Trends & Ballpark

Oracle Park = pitcher-friendly, keeps HRs down.

Both teams skew to unders in SF because of the park.

Cubs bullpen sometimes shaky, Giants bullpen up-and-down.

Parlay Angles

Imanaga 6+ Ks (Giants whiff enough)

Webb under 5.5 Ks (Cubs make contact)

Hoerner 1+ hit (contact king)

Flores or Lee 1+ hit for Giants side

That would give you a mix of pitcher props + batter contact props, in line with how we usually build.

---

quote:

---

Seiya Suzuki: hot bat lately

---

He is 3/16 over the last 5 with 4 Ks 1 RBI and 5 BB

quote:

---

I find it fun

---

Me too, and I do enjoy some of the breakdowns (when you ask), because it can provide some surprising insight about matchups, tendencies, etc.

I just don't find it as consistent anymore, and when I have to spend so much time verifying and correcting, it feels pretty inefficient.

Sometimes I'll provide a URL or screenshot with something like, "use this to confirm the MLB games officially scheduled for today, along with each team's starting pitcher," blah blah, and it will still get some of the games wrong or two or three starting pitchers that have never been with those teams.

Weird anomalies, but great when it actually works.