# PrizePicks Demons and Goblins Explained — How to Use Them in 2026

> What PrizePicks demons and goblins are, how the multipliers work, when each is +EV, and how to find demons that hit and goblins that fail.

**Date:** 2026-04-16  
**Author:** Jason Bowman  
**Tags:** PrizePicks, DFS, Strategy, Guide  
**Full article:** https://heatcheckhq.io/blog/prizepicks-demons-goblins-explained  
**Live picks & dashboards:** https://heatcheckhq.io

---


If you have spent any time in the PrizePicks app, you have seen the little red and green icons next to some projections. The red demon and the green goblin are not cosmetic. They are PrizePicks' way of letting you trade probability for payout in either direction — and most players trade in the wrong direction.

This guide walks through exactly what demons and goblins are, how the multipliers change your expected value, and when each one is worth taking.

## What Are Demons and Goblins on PrizePicks?

PrizePicks lets you pick whether a player will go OVER or UNDER a projected stat line. Standard lines are designed to sit near a 50/50 coin flip. Demons and goblins are modified versions of those same lines.

- **Demons** are **harder** picks. PrizePicks shifts the line against you — a scorer's 22.5 points line becomes a 24.5 demon, for example. You need a better outcome than the standard line would require. In exchange, PrizePicks pays you a **higher** multiplier if your full entry hits.
- **Goblins** are **easier** picks. The line shifts in your favor — that 22.5 points line becomes 20.5. You need a worse outcome to cash. In exchange, PrizePicks pays you a **lower** multiplier.

Put simply: demons are high-risk, high-reward. Goblins are low-risk, low-reward. Standard picks sit in the middle.

The trap is that both feel intuitive. Goblins look safe, demons look juicy. But the multipliers are the whole game, and they are not priced symmetrically.

## The Multipliers Explained

PrizePicks uses a flex-pay and power-pay structure. The payouts depend on how many legs you play and how many demons or goblins you mix in. The numbers below reflect PrizePicks' 2026 power-play (all-or-nothing) structure for standard entries.

| Entry Size | All Goblins | All Standard | 1 Demon Mixed | All Demons |
|------------|-------------|--------------|----------------|------------|
| 2-pick     | 2x          | 3x           | ~3.5x          | ~4x        |
| 3-pick     | 2.25x       | 5x           | ~6x            | ~8x        |
| 4-pick     | 5x          | 10x          | ~12x           | ~15x       |
| 5-pick     | 10x         | 20x          | ~22x           | ~25x       |
| 6-pick     | 15x         | 25x          | ~30x           | ~37.5x     |

Numbers drift slightly by sport and season, but the ratios hold. A 6-pick all-demon entry pays 1.5x what a standard 6-pick pays; a 6-pick all-goblin entry pays only 0.6x, even though each leg is easier. That asymmetry is the entire story.

## The Math: When Demons Are +EV

A standard PrizePicks projection targets roughly a 50% hit rate per leg. To break even on a 2-pick at 3x, each leg must hit ~57.7%. On a 6-pick at 25x, each leg must hit ~59% (25^(1/6) ≈ 1.698, so required probability is 1/1.698 ≈ 58.9%). Every pick needs a real edge on top of the coin flip.

A demon shifts the line by 0.5 to 1.5 stat units depending on the sport. A 0.5-unit shift typically drops the true hit probability from ~50% to somewhere around 38-44%. Here is the key: the demon multiplier goes up by **more** than the probability goes down — if the underlying line was fair and you have a real model edge.

### Worked example — a single demon leg

Say you are looking at a player's strikeout OVER. Standard line is 6.5 K's, pricing in roughly 50% both ways. The demon version is OVER 7.5 K's.

- Standard leg true probability: 50% (fair line)
- Demon leg true probability: ~42% (harder threshold)
- Standard leg multiplier contribution: roughly 1x (neutral)
- Demon leg multiplier contribution: adds ~15-20% to the final payout on a 4-6 pick entry

Now add a real edge. Your model says the pitcher faces a heavy-strikeout lineup and has a 32% whiff rate over his last five starts, translating to a 56% true probability of OVER 7.5 K's — a +14% edge on the demon.

Plug that into a 5-pick entry where four legs are standard (assume each is 55% true) and one is the demon (56% true). All five legs independent:

- Combined probability: 0.55^4 × 0.56 = 0.0915 × 0.56 ≈ 5.13%
- Multiplier with one demon mixed in: roughly 22x
- Expected return per $1: 0.0513 × 22 = **$1.13**, a ~13% edge

Swap that demon for a standard fifth leg at 55% true:

- Combined probability: 0.55^5 ≈ 4.98%
- Multiplier: 20x
- Expected return per $1: 0.0498 × 20 = **$0.996**, a ~0.4% loss

The demon flipped a break-even entry into a +EV one — but only because the model had a real read. Take the same setup without an edge (demon at its true ~42%), and you go from $1.00 to $0.92. Bleed.

The takeaway: **a demon is +EV when your true probability beats the demon's implied probability by roughly 4 percentage points or more.** Below that, the extra multiplier does not cover the added variance and reduced hit rate.

## The Math: When Goblins Are -EV

Goblins are the inverse trap. They look safe because the line is easier, but the multiplier drop is punishing.

Take the same 5-pick entry. Replace one standard leg with a goblin:

- Goblin true probability (easier line): ~60-62%
- Multiplier drop: from 20x standard to roughly 12x with one goblin (and 10x with two goblins, all the way down to 10x flat on all-goblin 6-pick entries)

Plug in numbers:

- Combined probability: 0.55^4 × 0.62 = 0.0915 × 0.62 ≈ 5.67%
- Multiplier: 12x
- Expected return per $1: 0.0567 × 12 = **$0.68**, a 32% loss

You added probability (5.67% vs 4.98%) but the multiplier collapsed from 20x to 12x. The net is catastrophic.

This is because PrizePicks effectively **doubles the vig on goblins.** The line gets easier by about 10-12 percentage points in true probability, but the payout drops by 40%+ on multi-leg entries. You need a truly massive edge — something like 70%+ true probability on the goblin leg — to break even versus just taking the standard line.

Outside of one narrow scenario (a goblin where your model says the line should be two full stat units easier), goblins are losing plays. The fact that they are marketed as the "safe" option is part of why PrizePicks promotes them.

## Identifying Demon Candidates with HCHQ Data

A demon only works if you have a genuine read that the standard line underpriced the OVER (or overpriced the UNDER, for demon UNDERs). HeatCheck HQ's dashboards are built around surfacing exactly those gaps.

**MLB hitter demons for hits, total bases, or HR props**

Look for batters whose recent underlying metrics outpace their surface stats. On our [hitting stats dashboard](/mlb/hitting-stats), sort by xBA minus actual BA — hitters with a 30+ point gap are running below their true skill level and are primed to regress up. Pair that with a barrel rate in the top 20% of qualified hitters and you have a demon candidate for total bases OVER. Use the situational splits (Day/Night/Home/Away OPS) to make sure the matchup context supports it.

**MLB pitcher demons for strikeouts**

On our [strikeout props page](/mlb/strikeout-props), we surface pitchers with rising whiff rates over their last five starts. A whiff rate climbing from 28% to 33% usually means the pitcher has found a new put-away pitch or is attacking a sequence that is working. That is a textbook demon candidate for K's OVER — the book's projection has not caught up yet.

**Cross-sport demon scanning**

The [Heat Score leaderboards](/mlb/leaderboards) surface the highest-conviction plays for today's slate. When a projection carries a Heat Score above 80 and the PrizePicks demon is in the same direction as our lean, our model has a stronger read than the market. Drop the exact prop into the [Prop Analyzer](/check) for a full breakdown — hit rates, recent trends, matchup factors, and our confidence score — before you submit.

## Common Mistakes

**Mixing too many demons in one entry.** Each demon adds variance. A 5-pick with four demons at 55% true each has a combined probability of about 9.2%, which is worse than it sounds because real-world outcomes are not perfectly independent — a rainout, an early pitcher exit, or a blowout can torch multiple legs at once. Keep demons concentrated where your edge is largest.

**Avoiding demons entirely.** The opposite mistake. Players who are scared off by the red icon leave +EV money on the table. If your model is right more than 54% of the time on a demon leg, you should be taking that pick every time — the multiplier makes up for the drop in hit rate.

**Not adjusting for stat volatility.** A pitcher's strikeout total is relatively stable (22-28 batters faced, whiff rate drives K's). A hitter's home run total is binary and noisy. Demons on stable stats are much safer than demons on volatile ones. Avoid HR demons unless the matchup is exceptional (power hitter, HR-prone pitcher, hitter-friendly park, favorable weather).

**Confusing "line moved" with "demon is +EV."** A shifted line does not automatically have a known true probability. Always run the demon's implied probability against your own model — if the gap is at least 4-5 points in your favor, take it. Otherwise, take the standard line.

**Stacking demons on correlated outcomes.** Two K demons from the same team's staff are not independent. Keep demon legs on different games or different stat types.

## Optimal Demon/Goblin Mix

Based on the math above, here is a practical framework for 2026:

- **Zero goblins ever, unless your model gives you a 15+ point edge on the goblin leg.** In practice, this almost never happens — if your edge is that big, you should be firing a standard or demon leg, not a goblin.
- **1-2 demons per 4-6 pick entry** is the sweet spot. Concentrated in your highest-confidence spots (biggest edge versus the demon's implied probability).
- **All-demon entries only on short slates** (2-3 pick), when each leg has a significant edge and the variance is manageable.
- **Pure standard entries are fine** for grinding. They do not produce the biggest scores, but they keep expected value positive when you have modest 3-5 point edges.

Think of it like poker. Standard picks are your solid value bets. Demons are your thin value raises — profitable only when you know you are ahead. Goblins are limping and then calling a raise with a weak hand. You do not limp.

## Bottom Line

Demons and goblins are not gimmicks. They are priced asymmetrically: demons can be +EV when you have a real edge, while goblins are almost always -EV because PrizePicks effectively doubles the vig on them. The red icon is not scary; the green icon is the one that quietly drains bankrolls.

If you take one thing from this guide: **never take a goblin without a huge model edge, and take demons aggressively when your data supports it.**

Ready to find demon candidates for tonight's slate? Hit our [Heat Score leaderboards](/mlb/leaderboards) for the highest-conviction plays across MLB, dig into [hitting stats](/mlb/hitting-stats) and [strikeout props](/mlb/strikeout-props) for the underlying metrics, and run your final entry through the [Prop Analyzer](/check) to confirm each leg's edge before you submit. Demons pay when you bring data. Bring data.


---

*Data powered by HeatCheck HQ — sports analytics platform. Free tools at https://heatcheckhq.io*