The Same Skill Behind Consistent Parenting and Sharp Predictions

What each one is actually for

Triple P

Triple P breaks "be a good parent" down into specific, learnable pieces: descriptive praise, telling a child exactly what they did well — "I like how you put your shoes away without being asked" — rather than a generic "good job" that doesn't tell them which action to repeat; planned ignoring, deliberately withholding attention from a behavior that's mainly happening to get a reaction, like a tantrum with an audience, while staying nearby and still responding warmly to everything else the child does; and predictable routines, doing the same sequence of events — dinner, bath, story, lights out — in the same order, so a child isn't spending energy testing what happens next.

Superforecasting

Calibration training takes "I'm pretty sure" and forces it into a number: not "probably," but 70%, and not for a single big life question but across many smaller, checkable ones — will this project ship by Friday, will it rain tomorrow, will a specific event happen this quarter. The number gets written down before the answer is known. Once the answer is known, it gets scored: did the thing you called 70% likely actually happen close to seven times out of ten across many similar predictions, or does your "70%" really behave more like a 40% or a 90% once it's checked against enough outcomes.

What you get out of pairing them

What Triple P's consistency buys a child is predictability in the one place it matters most: a behavior reliably produces the same response, which is precisely what teaches a child what to expect and what to do, rather than what to test. An inconsistently applied response — ignored on a calm day, reacted to on a hard one, for the exact same behavior — doesn't just fail to work; researchers studying reinforcement schedules have found that occasional, unpredictable attention to a behavior can strengthen it more than either consistent attention or consistent ignoring would, the same mechanism that makes a slot machine hard to put down. What calibration training buys a forecaster is something almost nobody gets by default: actual knowledge of whether their gut sense of confidence is accurate. Most people go their whole lives never finding out whether their "pretty sure" means they're right sixty percent of the time or ninety percent of the time, because they never wrote the number down and checked it. Calibration training turns a vague, untestable feeling into a specific, improvable skill.

What Triple P rests on

• Consistency strengthens behavior change clinical practice
  Applying a behavior-change strategy consistently, without exceptions, determines whether it produces lasting change.
• Extinction reduces attention seeking behavior randomised trials
  Withdrawing attention from a behavior that is maintained by attention reduces its frequency over time.
• Positive reinforcement strengthens behavior randomised trials
  Reinforcing a desired behavior with specific descriptive praise increases how often that behavior occurs.
• Predictable routines reduce conflict clinical practice
  Predictable structured routines reduce child behavioral problems by lowering uncertainty and testing.

What Superforecasting rests on

• Decomposition improves estimate accuracy observational studies
  Breaking a complex uncertain question into smaller more tractable sub-questions produces more accurate estimates than judging the whole question at once.
• Explicit probability estimates improve calibration randomised trials
  Forcing a numeric probability rather than a vague verbal hedge makes a prediction falsifiable and trainable.
• Outcome feedback corrects judgment bias randomised trials
  Regularly scoring predictions against actual outcomes and reviewing the misses corrects systematic biases in judgment over time.

Where they actually meet

Set them side by side and the same move shows up twice: both replace a single global judgment — "I'll react however feels right to this behavior," "I feel pretty confident about this" — with a specific, small, repeatable unit of action, done the same way every time, and then checked against what actually happens. Triple P checks a parent's consistency against whether the target behavior changes; calibration training checks a forecaster's stated number against the real outcome. Neither method works from a single instance — one well-handled tantrum, one correct guess — because a single instance can't be checked against a pattern. Both need repetition to generate anything worth learning from.

The principle underneath both

A specific behavior only reliably changes, and a judgment only reliably sharpens, when the exact same action gets repeated the same way across many instances and then compared against what actually resulted — a single good-faith attempt, applied differently each time depending on mood or convenience, doesn't accumulate into a pattern either the child or the forecaster's own brain can learn from. The repetition and the checking aren't supporting details around some other technique; in both cases they're doing the actual work, which is why a good parenting strategy applied inconsistently, or a smart guess never checked against an outcome, produces far less than the same action done the same way and tracked.

What that makes the real goal

If repeated, checked, specific action is what makes the difference, then the target both are aimed at is getting a specific outcome — a child's behavior, a forecaster's stated confidence — to reliably track something real: the actual consequence a parent wants a behavior to be met with, or the actual base rate an event occurs at. That target holds regardless of the domain in play; the method for reaching it doesn't change whether the thing being trained is a five-year-old's bedtime routine or a professional analyst's confidence in a geopolitical forecast — repeat the specific action, check it against the real result, adjust.

Other practices that serve the same goal

Applied behavior analysis, the clinical field that grew out of B.F. Skinner's research on reinforcement, has decades of work on exactly why inconsistent reinforcement is worse than either consistent reinforcement or consistent non-reinforcement, including research on variable-ratio schedules — the same unpredictable-reward pattern that makes gambling machines compelling — showing occasional, random reinforcement of a behavior can make it harder to extinguish than reliable reinforcement would. In judgment and decision research, psychologist Daniel Kahneman's later work on "noise," inconsistency in repeated professional judgments made by the same person or across different people facing the same case, found that inconsistency alone, separate from any bias, produces large and costly errors in fields from insurance underwriting to criminal sentencing, correctable mainly by replacing an individual's in-the-moment judgment with a consistent, structured decision rule. And in habit-formation research, work by psychologists Wendy Wood and David Neal on the role of stable context and cues in strengthening automatic behavior pursues a related target — consistency of the trigger, not just the response — through the lens of everyday habits rather than parenting or forecasting specifically.

A tradition that already does this

Gerald Patterson's research group at the Oregon Social Learning Center was applying Skinner's operant-conditioning research directly to parent training back in the 1960s and '70s, decades before Triple P was formally developed, and their core finding was the same one Triple P builds on: a parent trained to apply a specific, consistent response to a specific child behavior, and to track whether that behavior actually changed, produced far better results than general advice to "be more patient" or "set better boundaries." On the forecasting side, national weather services have scored meteorologists' probability-of-precipitation forecasts against actual rainfall using a formal accuracy measure, the Brier score, since the 1950s — long before Philip Tetlock's forecasting tournaments applied the same idea to individual judgment on world events. Calibration training, in a real sense, is the weather service's decades-old scoring method turned into a personal skill anyone can practice on any kind of prediction.

Turning a Parenting Instinct Into a Scored, Checkable Guess

Most parents already make informal predictions constantly — "this new bedtime routine will probably help within a week or two" — and never find out whether their instinct for what works is actually good, because the prediction was never written down or checked. Borrow calibration training's exact method here: before trying a new response to a specific behavior, write down a real percentage for how likely you think it is to produce a visible change within two weeks, then actually check back on that date and mark whether it happened. Do this across five or six different parenting decisions over a few months, and a pattern usually shows up — maybe your gut consistently overestimates how fast a new routine will take hold, in which case you now know to give any new approach more patience than your gut initially wants to.

Borrowing Triple P's Consistency Discipline for Your Own Predictions

Calibration training's number-writing habit is easy to do inconsistently — a person writes real numbers down for a week, gets busy, and drifts back to vague verbal hedges like "probably" for the next month. Triple P's core insight about consistency applies directly here: pick one specific, small class of prediction — say, whether a given work task will finish by its stated deadline — and commit to writing a real percentage for every single instance of that one kind of prediction, the same way a parent commits to the same planned response for the same misbehavior every time, rather than only when it's convenient. The skill improves from the repetition of the exact same tracked action, not from occasionally doing it well.

The Bedtime Routine That Only Worked on Good Nights

You've started a new bedtime routine with your daughter — bath, two books, lights out at the same time every night — and it's working great on the nights you have the energy to hold the line. On a hard night, tired, you let the routine slide: one more book becomes three, lights-out drifts forty minutes later, and you tell yourself it's just tonight. Within two weeks, bedtime has quietly gone right back to being a nightly negotiation, and you can't figure out why the routine "stopped working," when really it never got a consistent trial at all — it got applied on the easy nights and abandoned on the hard ones, which is close to the least consistent way to run it. You start again, but this time you write down, on the calendar, exactly what the routine is and commit to it on the hard nights specifically, since those are the nights the whole test actually depends on. Two weeks later, holding the line on the bad nights, not just the good ones, is what finally makes bedtime boring again in the way a working routine should be.

When this is worth reaching for

A strategy, a parenting approach, a professional judgment call, seems to work some days and fail on others, and the real suspicion is that the inconsistency itself is the problem, not the strategy. The same fix applies anywhere a rule of thumb gets applied differently depending on mood or convenience and has never actually been written down and checked against what happened.

Where this does not apply

If a child's difficult behavior has a cause outside a parent's control to consistently respond to — a medical issue, a developmental condition requiring a different kind of support entirely — applying planned ignoring or descriptive praise with more consistency won't reach the actual cause, and a professional evaluation matters more than tightening up the technique. And forcing a numeric percentage onto every small daily decision, rather than the kind of repeated, checkable prediction calibration training is built for, mostly just adds friction; the method only works on judgments you'll make many times and can actually check against a real outcome, not on one-off choices with no comparable instance to learn from.

What the evidence does and does not support

Triple P itself has been tested in numerous randomized trials across different countries and shows real, if moderate, effects on reducing problem behavior and improving parent-reported confidence, and the specific mechanisms underneath it, descriptive praise and planned ignoring, each have their own separate experimental support going back decades in behavioral psychology. Calibration training's evidence comes from a different source entirely: forecasting tournament research, most notably Philip Tetlock's Good Judgment Project, found that specific individuals could measurably improve their calibration with feedback and practice, tracked against real-world outcomes rather than a lab task. What's never been studied is whether deliberately applying a calibration-training mindset, writing down a number, checking it later, to parenting decisions specifically would improve a parent's ability to judge what's working; that's an inference from a shared mechanism, not a finding anyone in either research tradition has actually gone looking for.