Richard Feynman gave science its sharpest self-warning. The first principle, he said, is that you must not fool yourself, and you are the easiest person to fool. Apply to fantasy football drafting

Don’t fool yourself in the draft room—you’re the easiest mark. 

Feynman’s rule is the single best mental model for fantasy football drafting. 

The game is mostly noise, incomplete information, and motivated reasoning dressed up as analysis. 

Your brain will invent stories, protect ego, and chase feelings of certainty. The first principle is to catch yourself doing it.

Common self-deceptions and how to interrupt them

• “I know this guy is going to break out.”
  You almost never do. Breakouts look obvious only in hindsight. Demand evidence stronger than a single strong training camp report, a new offensive coordinator, or “he finally looks healthy.” Ask: What would have to be true for this projection to fail, and how likely is that failure mode? If you can’t answer clearly, you’re storytelling.

• Recency and highlight bias.
  Last year’s fantasy points, one spectacular playoff game, or a viral training-camp clip feel more real than three years of data or age curves. Force yourself to look at multi-year trends, situational splits, and projected opportunity (targets, carries, red-zone work) rather than the narrative you just watched.

• Favorite-team and name-value inflation.
  Drafting the star from the team you watch every Sunday feels right. It is usually wrong. Separate fandom from expected points. If the player’s Average Draft Position (ADP) or projected points don’t support the pick, you are paying a tax for emotion.

• Overconfidence in your process.
  You have a “system,” a spreadsheet, or a favorite projection model. So does everyone else. The consensus ADP is not perfect, but it is the market’s current best guess incorporating thousands of drafts. Deviating sharply requires a specific, falsifiable reason—not “my model likes him more.” Track your big deviations and their outcomes over seasons; most people discover they are worse than the market.

• Sunk-cost and roster-attachment bias (in-season extension of the draft).
  You reached for a player in the third round, so you keep starting him through mediocrity. The draft is over. Every week is a new decision based on current expected value. Pretend you just acquired the roster and rank the players cold.

• Ignoring base rates and variance.
  Most second-year wide receivers do not become WR1s. Most aging running backs decline. Most tight ends outside the top tier are interchangeable. Extreme outcomes happen, but planning your entire draft around them is how you fool yourself into reaching. Build a portfolio that survives ordinary outcomes.

Practical anti-self-fooling habits

1. Write down your concrete reasons for every significant reach or fade before the pick. Revisit them after the season.

2. Rank players by projected points (or VORP) first, then layer in risk/upside. Do not start with “guys I like.”

3. Actively seek the strongest counter-argument to your favorite sleepers. If you cannot steel-man the case against them, you are not ready.

4. Treat ADP as a prior. Update only with information the market is likely underweighting (role clarity, injury recovery specifics, scheme fit that is not already priced in).

5. Accept that a large fraction of your “brilliant” picks will look average or worse, and a large fraction of “safe” consensus picks will bust. The goal is process, not the illusion of control.

Fantasy football rewards preparation and discipline more than genius.

The sharpest edge available to most managers is simply refusing the stories their own brain wants to tell. You will still be wrong often—that is the nature of the domain. The difference is whether you were honestly wrong or whether you fooled yourself first.

The sharpest edge available to most managers is simply refusing the stories their own brain wants to tell. How?

Refusing the stories is a deliberate practice, not a personality trait.

Here is how most managers can actually do it.

1. Force the story into writing before it becomes a pick

Before every significant reach or fade, open a note and answer three questions in complete sentences:

• What exact evidence makes this player better (or worse) than his current ADP?

• What is the strongest counter-argument, and why am I dismissing it?

• If this pick fails, what will the most likely reason be?

If you cannot finish the note cleanly, you are still inside the story. Do not make the pick until the note is done. The act of writing slows the emotional leap and exposes thin reasoning.

2. Separate “I like him” from “I project him”

Create two ranked lists:

• One pure ranking by projected points or value over replacement (your model, consensus projections, or a simple average of several sources).

• A second list of players you personally enjoy watching or root for.

Draft almost exclusively from the first list. The second list is entertainment, not strategy. Crossing from one to the other requires an explicit, written justification that survives the three questions above.

3. Install a pre-draft “bias checklist”

Keep a short list visible during the draft:

• Am I overweighting last season or one recent game?

• Am I drafting a player from my favorite team or a player whose highlights I just watched?

• Am I ignoring age, injury history, or opportunity trends because the narrative feels good?

• Is my deviation from ADP larger than the uncertainty in the data justifies?

Run the checklist on every non-consensus pick. It takes ten seconds and catches most self-deception in real time.

4. Treat ADP as the default and demand a specific reason to leave it

The market price (ADP) already averages thousands of other managers’ information. Your edge is not “being smarter than the room.” 

Your edge is noticing when the room is systematically missing something measurable (clear role change, scheme fit not yet priced in, recovery data, etc.).

If you cannot name the missing information in one sentence, stay near ADP.

5. Keep a decision journal and review it ruthlessly

After the draft, and again at season’s end, go back to every note you wrote. Score each decision:

• Process good / outcome good

• Process good / outcome bad

• Process bad / outcome good (this is the dangerous one—luck that reinforces the story)

• Process bad / outcome bad

Most managers only remember the wins. The journal forces you to see how often the story felt true and was still wrong. Over two or three seasons the pattern becomes impossible to ignore.

6. Practice the uncomfortable pause

When you feel the surge of certainty—“this is the guy”—stop for 30–60 seconds. Look at the next three players on your board. Ask whether the surge is coming from data or from the narrative your brain just constructed. Certainty is usually the warning sign, not the green light.

These habits do not eliminate bias. They make the bias visible and expensive to follow. That visibility is the entire edge.

Your edge is noticing when the room is systematically missing something measurable (clear role change, scheme fit not yet priced in, recovery data, etc.). 

Discuss deep

Your only durable edge is spotting where the consensus systematically underweights measurable information that has already occurred or is highly likely. Everything else—talent evaluation, “eye test,” gut feel for breakouts, or believing you see traits the market ignores—is mostly noise dressed as insight. 

The ADP (and the live draft room that converges on it) is an efficient-enough aggregator of public knowledge, historical production, and collective projection models. Beating it consistently requires identifying discrete, observable gaps in that aggregation process, not superior forecasting of future randomness.

Why the "Draft Room" systematically misses things

The consensus is slow or incomplete in three recurring ways:

1. Discrete role changes lag behind narrative inertia
   A free-agent departure, a coaching staff’s stated commitment, or a depth-chart clarification creates a sudden jump in expected opportunity (snaps, targets, carries, red-zone share). Opportunity is the dominant driver of fantasy points; talent is secondary once a player is on the field.
   The room often prices the player rather than the role. It remembers last year’s production or the veteran’s name value longer than the new depth chart justifies. The measurable signal is the projected change in volume, which can be estimated from prior-year usage of the vacated role, coaching history, and early practice reports. If ADP has not yet moved proportionally to that volume increase, the gap is real.

2. Scheme and situational fit are under-modeled until proven
   Certain offensive systems historically inflate specific positions (e.g., heavy play-action boosting YAC receivers, two-tight-end sets, or zone-heavy runs creating backfield opportunities). When a player moves into such a system, or a coordinator with a clear historical profile arrives, the expected efficiency or volume shift is measurable from past data.
   Consensus projections often average the player’s prior efficiency with a generic team projection rather than fully applying the new scheme’s base rates. The lag exists because most public models and casual drafters update slowly; they wait for preseason games or early-season results. The edge is applying the historical scheme multipliers before the results confirm them—provided the fit is concrete (personnel, philosophy, not vibes).

3. Recovery and availability data are noisy and emotionally processed
   Injury recovery is not binary. Practice participation trends, specific medical timelines, surgically repaired structures with known return-to-play distributions, and workload ramps produce probabilistic availability and effectiveness curves.

The room tends to overreact to the injury label itself (“he’s injury-prone”) or underreact to granular positive signals (full participation in OTAs/minicamp, cleared without limitations, historical recovery norms for that procedure). 

Because availability multiplies every other projection, small improvements in expected games played or early-season workload can create large expected-point differences that ADP has not fully absorbed if the narrative remains “risky.”

These are not the only categories, but they share the same structure: a relatively hard, observable input (role vacancy, scheme history, recovery metrics) whose fantasy impact is large and whose incorporation into the market price is incomplete or delayed.

How to distinguish a true gap from self-deception

This is where Feynman’s rule becomes operational.

• The information must be public or nearly public. If you are the only one who “knows,” you are probably wrong, or the knowledge is not actionable. True edges come from better weighting or faster integration of shared facts.

• It must be quantifiable in expected points or value over replacement. “He’ll be more involved” is a story. “The departed player handled 18% of targets and 25% of red-zone looks; the new player is the clear primary replacement and ADP still prices him as a WR4” is a calculation. Run the numbers. If the implied point delta does not exceed the risk and the cost of the reach, there is no edge.

• The market must demonstrably not have adjusted yet. Check current ADP movement against the date of the information. If the news is two weeks old and ADP has already shifted 15–20 spots, the gap is closed. If it has barely moved, the inertia is still present.

• You must be able to state the failure mode cleanly. “This fails if the coaching staff splits the work more than historical norms or if the player’s efficiency collapses in the new scheme.” If you cannot define falsification, you are protecting a narrative.

• Base rates still apply. Most role changes under-deliver relative to the optimistic case. Most scheme fits produce regression toward the mean rather than pure multiplicative gains. Size the bet accordingly—usually a modest ADP deviation, not a massive reach.

The practical mechanics of hunting these edges

Treat the pre-draft period as an information-processing race against a slow-moving consensus. 

Map every relevant roster and coaching change against historical usage data. 

Maintain a simple expected-volume model: prior role shares × likelihood the new player captures them × team pass/run rate projections. 

Compare the resulting fantasy-point estimate to the points implied by current ADP. 

Large positive residuals that survive skepticism are candidates. 

For scheme, maintain a short library of coordinator tendencies (play-action rate, target distribution by alignment, TE usage, etc.). When a player or coach moves, apply the relevant historical deltas rather than waiting for new data.

For recovery, track participation logs and medical language with more granularity than headlines. Convert them into adjusted games-played probabilities using historical distributions for similar injuries. Then—critically—write the thesis exactly as described earlier: the measurable

Pre Mortems as a way?

https://www.scienceoffantasyfootball.com/2026-pre-mortem-tables-updated

Yes. Pre-mortems are one of the cleanest, highest-leverage ways to refuse the stories your brain wants to tell. A pre-mortem flips the usual post-mortem. Instead of waiting until the season ends and reconstructing why a pick failed, you assume right now that the decision has already failed, then force yourself to generate the most plausible reasons. 

This exploits a psychological asymmetry: people are better at explaining failure in hindsight than at anticipating it in advance. By artificially creating the hindsight, you surface risks, overlooked base rates, and self-serving narratives while you can still act on them.

How to run a pre-mortem on a fantasy football decision

1. State the decision in concrete terms

“I am drafting Player X at ADP +12 (or in the 4th round) because of [specific measurable edge: role change/scheme fit/recovery data].”

2. Assume the decision has failed

Fast-forward to the end of the season (or to Week 8 if you want an earlier checkpoint). The player finished as a clear bust relative to draft cost. Or your entire draft strategy underperformed. Do not soften it. Declare the failure as fact.

3. Generate the reasons why it failed

Force a list of plausible, specific causes. Aim for quantity and honesty:

• The role change never materialized; the committee approach persisted.

• Scheme fit was overstated; historical coordinator tendencies did not transfer.

• Recovery progressed slower than the optimistic medical timeline; early-season workload was limited.

• The measurable volume increase was real but efficiency collapsed (or the opposite).

• ADP had already partially priced the information and the remaining gap was smaller than I claimed.

• I overweight a single data point and ignored the broader base rate for similar players/situations.

• Opportunity was there, but the quarterback / offensive line / supporting cast underperformed in ways that were foreseeable.

• Simple variance/injury that had nothing to do with my thesis (important to include so you do not overfit).

4. Rank the reasons by likelihood and severity

Which failure modes are most probable? Which would hurt the most? This ranking becomes your risk register.

5. Decide whether the thesis still survives

After seeing the failure pathways laid out, ask:

• Does the expected-value gap still justify the cost of the reach?

• Are there cheap ways to mitigate the highest-probability risks (e.g., roster construction that does not depend on this player hitting, or a backup plan)?

• Am I still making the pick mainly because I like the story?

If the pre-mortem makes the pick look fragile, shrink the deviation from ADP or abandon it. That is the point.

Where pre-mortems add the most value in drafting

• Individual reach/fade theses: The exact place you are most likely to fool yourself.

• Overall draft strategy: “Assume this Zero-RB / Hero-RB / late-QB approach failed. Why?” Surfaces structural weaknesses before you lock them in.

• In-season extensions: When deciding whether to hold or trade a drafted player who is underperforming, run a mini pre-mortem on the original thesis.

Integration with the earlier process

The written note you make before a significant pick already contains a failure-mode section. The pre-mortem simply expands and stresses that section. It turns a polite “what could go wrong” into a more visceral “it has gone wrong—explain how.” The emotional distance of imagining failure already happened reduces the defensive urge to protect the original narrative. 

Used consistently, pre-mortems do not make you pessimistic; they make you calibrated. Most managers only update after the damage is done. You update while the decision is still reversible. 

That is a direct operationalization of Feynman’s rule: you deliberately make it harder to fool yourself by manufacturing the conditions under which self-deception is hardest to maintain.