Tag: Fantasy Analytics

IHM Knowledge Center • Fantasy Hockey Intelligence Center

What Are the Most Common Projection Model Mistakes?

What Are the Most Common Projection Model Mistakes? Learn how projections, uncertainty, replacement value, scarcity, schedule, and risk should be translated into fantasy hockey decisions.

Editor: Coach Mark • Updated: August 23, 2026

Short Answer

Common model mistakes include overfitting small samples, ignoring role changes, treating games played as certain, misusing regression, and applying one valuation method to every league.

Full Explanation

the Most Common Projection Model Mistakes belongs to a broader valuation process. A projection estimates future production, while a fantasy decision asks what that production is worth in a specific league and against realistic alternatives.

Strong models combine role, expected games, historical rates, age, team context, regression, league scoring, position requirements, schedule, and uncertainty. The goal is not perfect prediction. The goal is a better decision than the available alternatives.

Main Factors

• Small-sample bias
• Role changes
• Games played
• Regression
• League settings

How It Changes Fantasy Value

Projected value rises when expected production is strong relative to replacement level and the player has secure usage, favourable schedule value, positional scarcity, or upside that the market is not fully pricing.

Value falls when a projection depends on fragile percentages, uncertain games played, temporary role, extreme assumptions, or a player profile that is easy to replace in the league.

How to Build the Decision

• Start with expected games and role.
• Project the statistics the league actually scores.
• Apply regression where current rates appear unsustainable.
• Compare projected output with position-specific replacement level.
• Adjust for uncertainty, schedule, roster need, and decision cost.

Decision & Controversy

Projection systems can disagree because assumptions differ. One model may value recent form, another long-term rates, and another usage changes more heavily. The strongest approach is to understand which assumptions drive the difference rather than blindly selecting the highest number.

Edge Case

A player with the higher raw projection can still have lower practical fantasy value when his position is deep, schedule is difficult to use, role is unstable, or a similar replacement is freely available.

IHM Signal System: the Most Common Projection Model Mistakes

• Projection signal: What is the most reasonable central estimate?
• Range signal: How wide are the realistic floor and ceiling?
• Replacement signal: What value is available without paying for this player?
• Scarcity signal: Is the position or category difficult to replace?
• Risk signal: Which assumptions could fail?

Trigger-level rule: If small-sample bias or another critical projection, uncertainty, scarcity, replacement, or schedule-value signal is unclear, do not convert the estimate into a precise ranking or transaction value.

IHM Insight: the Most Common Projection Model Mistakes

A projection is not the decision. It is one input into the decision.

The useful edge comes from translating uncertain future production into league-specific value more accurately than the market does.

Mini Q&A

What Are the Most Common Projection Model Mistakes?
Common model mistakes include overfitting small samples, ignoring role changes, treating games played as certain, misusing regression, and applying one valuation method to every league.

What should be checked first?
Small-sample bias.

Should a projection be treated as a prediction of one exact outcome?
No. A projection is an estimate with uncertainty, and strong decisions use ranges, context, and replacement value rather than false precision.

Can the same projection produce different fantasy value in two leagues?
Yes. Scoring, roster depth, position requirements, schedule use, and replacement level can materially change valuation.

What is the IHM trigger-level rule?
If small-sample bias or another critical projection, uncertainty, scarcity, replacement, or schedule-value signal is unclear, do not convert the estimate into a precise ranking or transaction value.

Why This Concept Exists

Fantasy hockey decisions are made before future results are known. Projection and valuation models create a structured way to compare uncertain outcomes, prices, roster needs, and alternatives.

Key Takeaways

• Common model mistakes include overfitting small samples, ignoring role changes, treating games played as certain, misusing regression, and applying one valuation method to every league.
• Small-sample bias is a primary modelling factor.
• Projection and valuation are not the same thing.
• Replacement level provides the practical baseline.
• Ranges are more useful than false precision.
• Scarcity and schedule can change raw value.
• Risk should be attached to the assumptions that create the projection.