Training

Fading Treats the Right Way

"He only listens when I have food." That's not a stubborn dog — it's a lure that never got faded and a reward schedule that never changed. Here's the fix.

Article · Training Plan Companions series

black and brown long coated dog drinking beer — Photo: Nicole Calmet / Unsplash

The complaint is universal: 'She's great when I have treats, and deaf the second my hand's empty.' It feels like the dog gaming you, but it's actually a training mechanics problem with a clean fix — two of them, really, because the lure and the reward are different things, and both need to change on schedule. Fade the visible food from the cue, shift the payments from every-time to surprise-me, and you get a dog who works because the behavior pays — not because they can see the paycheck.

A dog who only works for visible food isn't disobedient. He's just never been shown the paycheck comes AFTER the work.

📋 Quick Read

Two different jobs: the LURE (the food that guides the behavior) fades early and fast; the REWARD (the pay for doing it) shifts from continuous to intermittent — later, and never to zero.
The lure becomes a hand signal: once the behavior's fluent, run the same motion with an empty hand and pay from the other hand or your pocket — the food leaves the cue but not the equation.
Intermittent reinforcement is the magic: unpredictable rewards (sometimes one treat, sometimes three, sometimes praise) produce STRONGER, more durable behavior than treating every single time — it's the slot-machine effect, working for you.

🧠 Why It Works'Only works for food' is two mechanics failures wearing one costume. First, the lure never faded: if food stays in the guiding hand, it becomes part of the cue — the dog learns 'sit' means 'sit when the hand has food,' so no food, no cue. The fix is fast and early: once the behavior happens reliably, the luring hand goes empty (becoming a hand signal) and the reward comes from elsewhere, breaking the food-in-hand dependency while keeping the food-as-consequence. Second, the schedule never thinned: continuous reinforcement (treat every time) is perfect for teaching but fragile for maintaining, because the dog notices instantly when it stops. Shifting to intermittent reinforcement — variable, unpredictable payoffs — exploits the best-documented finding in behavioral science: variable-ratio schedules produce the highest, most extinction-resistant response rates (the same mechanism that makes slot machines compelling). Done in order — fade the lure early, thin the schedule later, never to zero — you get a dog who performs on cue anywhere, paid on a schedule that keeps them guessing and trying.

The empty-hand deafness

Lure vs. reward

Fade the lure early

Thin the schedule late

Never zero

Week plan

In the app

Separate the lure from the reward (the whole insight)

The lure is the food you use to GET the behavior — the treat at the nose that guides the sit or the down. The reward is the food that PAYS for the behavior once it happens. Beginners collapse these into one — lure with a treat, hand over the same treat — and the dog reasonably concludes the visible food is the trigger. The unlock is to split them: use the lure only for the first several reps of a new behavior, then drop it, and let the reward come from a different source (other hand, pocket, treat pouch, the environment) and a different moment (after the behavior, sometimes after a delay). The food never leaves training; it just stops being the bribe the dog can see coming.

💡 Pak PrincipleLure to teach, reward to pay — and never let the dog see the pay before the work. Visible-food-first is the exact habit you're undoing.

📈 Track It In Pak SocialTrack the fade in Pak Social's Training Coach — the plans flag when a behavior is fluent enough to drop the lure and thin the schedule, so you fade on the dog's readiness instead of guessing.

a dog laying on the floor with a person holding a stick — Photo: Ayla Verschueren / Unsplash

Fade the lure early, then thin the schedule late

Order matters. Fade the lure fast — after just a handful of successful lured reps, run the identical hand motion with nothing in your fingers, mark the behavior, and pay from the other hand; the gesture is now a hand signal, and the food-in-hand dependency is dead before it forms. Keep rewarding every rep at this stage — you're still teaching. THEN, once the behavior is solid on cue in a context, start thinning the reward schedule: pay most reps, then most-but-not-all, then randomly (this one gets a jackpot, that one gets praise, the next gets nothing, the one after gets three). The unpredictability is the point — a dog who can't predict which rep pays keeps offering strong responses on all of them, which is exactly the durable, food-independent behavior you were after.

🐕 Read Your DogWatch the response quality as you thin: crisp, prompt, enthusiastic means your schedule's rich enough — keep thinning. Slowing down, hesitating, or checking your hands means you thinned too fast; bump the pay rate back up and thin more gradually. The dog will tell you the right pace if you read the responses.

⚠️ Don't Accidentally Teach ThisGoing cold turkey to zero food, 'because he should know it by now.' Cutting rewards entirely triggers extinction — the behavior weakens and dies, exactly the opposite of the goal. Fluent behaviors still get paid, just unpredictably and sometimes with life rewards instead of food. Maintenance reinforcement is forever; only its schedule changes.

Never zero — just smarter

The endgame isn't a food-free dog; it's a dog paid on a variable schedule with an expanded currency. Rewards graduate beyond treats into 'life rewards' — the door opening for a polite sit, the ball thrown for a solid drop, the freedom to go sniff after a check-in, the greeting granted for four-on-the-floor (the Premack principle: the behavior the dog wants pays for the behavior you want). Keep genuine food jackpots in rotation for hard asks and great responses, so the possibility of a big payoff keeps the whole system lively. The clicker or marker, incidentally, fades differently — it's a teaching tool you can simply stop using once the behavior's on cue, no schedule needed. Fade the lure, thin the treats, expand the currency, never hit zero: that's a dog who works anywhere, food or no food in sight.

🏆 Trainer's ChallengeThe Empty-Hand Week: pick one solid behavior and this week (a) run it purely on the empty-hand signal, and (b) thin its rewards to roughly every-other-rep, unpredictably, mixing in praise and one life reward. A crisp response with your hands empty is the whole thing working.

Illustrative

📈 The fade, in order

Lure leaves early; rewards thin late and never to zero. Most 'only-works-for-food' dogs had these two steps merged or skipped.

Lure the behavior

Empty-hand signal

Reward every rep

Reward most reps

Variable / random

+ life rewards

Step 1

Step 2

Step 3

Step 4

Step 5

Step 6

Illustrative sequence per AKC lure-reward and marker-training guidance; variable-ratio reinforcement is best-documented for durability.

🗓 This Week's Plan

Days 1–2: Pick one behavior; confirm the lure is already gone (empty-hand signal, pay from elsewhere) — fix it if not.
Days 3–5: Start thinning that behavior's schedule — most reps, then unpredictably; watch response quality.
Days 6–7: Run the Empty-Hand Week test and add one life reward (door, ball, or freedom) to the currency.

📈 Track It In Pak SocialPak Social's Training Coach flags when each behavior is ready to drop the lure and thin the schedule, and logs response quality as you go — so fading happens on the dog's readiness, and 'only works for food' never gets a chance to set in.

Start the fade in Pak Social

Empty your luring hand on one cue today, pay from your pocket, and start keeping the dog guessing. The paycheck-after-the-work dog was always trainable — this is the mechanics.

Get Pak Social — Free

➡️ Up nextMarker Words and Timing5 min read›

Sources: AKC — Tips for Fading the Lure in Lure-Reward Training · AKC — Clicker Training: Mark & Reward Dog Training · AKC — How to Teach Your Dog Emotional Self-Control