On the ice, we see the result. We don’t see the data warehouses, video tagging, quality control, models, scouting notes, and the discussion that preceded the decision. That’s exactly where the NHL is changing today.

The result is visible. The system remains hidden.

The 2025–26 season featured two distinct case studies. Carolina finished the regular season with a 53–22–7 record and 113 points, then won the Stanley Cup with a 16–3 record in the playoffs. Anaheim earned 92 points, returned to the playoffs for the first time since the 2017–18 season, and advanced to the second round.

These results are not conclusive proof that statistics win games. A hockey outcome is determined by the roster, the training process, player health, goaltender performance, the work of the coaches, salary cap management, the timing of trades, and chance.

The key takeaway is more specific: both clubs have begun treating information as infrastructure, not as an afterthought.

When I lifted the Stanley Cup in 2004, decision-making relied on video, scouting, coaches, and experience. Today, those same pillars are supplemented by data teams, sports science, and systematic feedback. Hockey hasn’t become a math problem. Decision-making has become a more precise craft.

Data doesn’t eliminate uncertainty. It gives an organization a better chance to figure out where it might be wrong.

Carolina: Analytics Is No Longer a Separate Department

Eric Tulsky became the permanent general manager of the Hurricanes on June 18, 2024. He joined the organization as a consultant in 2014, later leading analytics, hockey management, and strategy, becoming assistant general manager, and eventually taking over all hockey operations. He holds a Ph.D. in chemistry from the University of California, Berkeley, and, according to the club’s profile, holds 27 U.S. patents.

However, the most important information is not his academic background. What matters is that analytics at the Carolina Hurricanes is not a peripheral function within the organization. It is integrated with personnel decisions, professional scouting, contracts, the salary cap, and hockey intelligence. In the current structure, Assistant GM Tyler Dellow leads the data team alongside the salary cap and contracts group.

The 2025–26 season was extraordinary: 113 points, first place in the Metropolitan Division, and the Stanley Cup after winning four series. Carolina defeated Ottawa, Philadelphia, Montreal, and Vegas in the finals.

Carolina 2025–26 Verified data
Regular Season 53-22-7, 113 points
Playoffs 16-3, Stanley Cup
General Manager Eric Tulsky
Management Model Integration of data, scouting, personnel, the salary cap, and contracts
Methodological Conclusion Analytics functions as a decision-making infrastructure, not as a standalone report

The Stanley Cup doesn’t prove that an algorithm has beaten the hockey experts. It shows something more precise: an analytics-driven front office can take full responsibility for the roster, scouting, contracts, and results. A model is only valuable when a scout can challenge it, a coach can translate it into a game plan, and a player can execute it under pressure.

The biggest change isn’t that the analyst sits at the table. In Carolina, he sits at the head of the table, but he makes decisions together with the hockey people.

Anaheim: Off-Cap Investments in Development

Pat Verbeek has been coaching the Ducks since February 2022. His career path is the opposite of Tulsky’s: two decades in the NHL as a player, followed by roles in scouting, player personnel, and management. Under his leadership, Anaheim has expanded its scouting and player development programs, built a performance department based on data and sports science, and strengthened its analytical capabilities in video scouting, research, development, data science, and engineering.

This model is particularly important during a rebuilding phase. A young player’s development isn’t linear, and a short streak of good games doesn’t necessarily indicate consistent progress. The club needs to take into account video analysis, playing role, physical conditioning, the quality of opponents, and long-term trends. The younger the core of the team, the more costly erroneous conclusions become.

Season Points Result and Organizational Context
2023/24 59 The Start of a Major Rebuild
2024/25 80 Year-over-year increase of 21 points
2025/26 92 Return to the playoffs, advancement to the second round

In two seasons, Anaheim went from 59 to 92 points. Verbeek was named a finalist for the Jim Gregory Award for General Manager of the Year. The official announcement also highlighted the expansion of the staff in scouting, player development, performance, and analytics as part of the rebuild.

Here, too, it’s impossible to pinpoint a single cause. Anaheim changed coaches, added experienced players, and gained another year of development for its young core. Therefore, the correct interpretation is not that analytics generated 33 points. The correct interpretation is that the club built its roster and decision-making system simultaneously.

Anaheim is not proof that data accelerates every rebuild. It is proof that a rebuild without a high-quality information system increases the cost of every mistake.

Two Paths, Same Principle

Organization Starting Point Common Principle
Carolina An analytically oriented GM on a team built to win Information is factored into personnel, game strategy, and contract decisions
Anaheim A former player and scout during the rebuilding of a young team Data, video, performance, and development grow alongside the team

Decision-makers don’t need more metrics

Neither the general manager, the coach, nor the scout needs another dashboard full of acronyms. They need an answer to a specific question: Is a player’s performance repeatable? In what role? Against opponents of what caliber? What does the video confirm? Where do the numbers differ from what’s observed? And what risk does the decision entail?

This is the line between analytics and hockey intelligence. A metric describes a phenomenon. Hockey intelligence contextualizes it within a decision, a timeline, a budget, and accountability.

Humility in the Face of Numbers: What Analytics Knows and Doesn't Know

Hockey is a low-scoring, fast-paced, and highly context-dependent sport. A single rebound, deflection, or brief lapse in concentration can change the course of a game or even a series. Therefore, the model does not predict the future. It estimates probabilities based on the information available to it.

The human perspective is just as incomplete. A scout doesn’t view every game without bias; a coach under pressure may overestimate the most recent result; and a player is unable to objectively evaluate all of his own decisions. Statistics and experience shouldn’t cancel each other out. They should reveal each other’s blind spots.

In the locker room, I learned that incorrect information presented with absolute certainty is more dangerous than admitting uncertainty. A good model, therefore, should not just show the answer. It must also show the quality of the input, the sample size, the sensitivity of the result, and the variables missing from the data.

Data knows Data cannot
Compare recurring patterns in a large sample. Predict the outcome of a single shift or a single series.
Isolate the influence of role, teammates, opponents, and score. Fully measure character, physical condition, confidence, or adaptation.
Identify anomalies that need to be reviewed on video. Make decisions without considering the hockey context or human factors.
Work with scenarios and identify uncertainty. Eliminate randomness and stochastic noise from sports.

A good model doesn’t just say what it thinks. It also says how certain it is and what it doesn’t know yet.

The infrastructure you can’t see on the ice

A quiet revolution doesn’t start with a model. It starts with whether the club can establish a reliable chain of events, from what happens on the ice all the way to the decision. If even one link in the chain breaks down, an output that looks exactly right can lead to the wrong conclusion.

Layer What Must Work Typical Failure
1. Data Collection Consistent video, event data, tracking, and accurate timestamps. Missing matches, inconsistent event definitions, inaccurate tagging.
2. Quality Correct player identity, error checking, data versioning. The model analyzes incorrect input or mixes incompatible leagues.
3. Context Role, teammates, opponents, score, age, and sample size. The raw number is mistaken for the player’s true value.
4. Translation A simple hockey question, video evidence, and a clear conclusion. The analytical output remains outside the coach’s or scout’s workflow.
5. Feedback A record of the decision, the result, and subsequent model calibration. The organization repeats mistakes because it doesn’t know what it has learned.

The NHL salary cap was set at $95.5 million for the 2025–26 season. For the 2026–27 season, it increased to $104 million, and the projection for 2027–28 is $113.5 million. The cap regulates player payroll. Scouting, video analysis, player development, performance metrics, and data analytics remain areas where clubs build their own operational advantage.

Why We’re Building CibakHockeyIQ

We are not building CibakHockeyIQ to replace scouts, coaches, or general managers. The goal is to create a validated interpretive layer between a complex data model and a club’s practical needs.

Every module must pass three tests: input quality, hockey-specific interpretability, and usability in real-world decision-making. If even one fails, the product isn’t ready. That’s why today we’re focusing our energy on development, testing, and calibration—not on grand announcements.

My experience from the Stanley Cup gives me one advantage: I know which information will withstand the pressure of the locker room and which will remain just a pretty graph. Technology, in turn, allows us to question first impressions before the club pays for a mistake with a contract, a draft pick, or a lost year of development.

The future doesn’t belong to the teams with the most dashboards. It belongs to the teams that know how to turn uncertainty into better decisions.

Carolina has shown that an analytics-driven approach can lead all the way to the Stanley Cup. Anaheim has shown that during a rebuild, it pays to build both the roster and the decision-making infrastructure at the same time. Neither story proves that data guarantees success. Both demonstrate that a high-quality organization does not want to make decisions blindly.

Martin Cibák
2004 Stanley Cup Champion, Tampa Bay Lightning
CibakHockeyIQ, co-founder & Chief Hockey Officer

Data is not a substitute for hockey judgment. It is a check on that judgment.

Sources and Verification

The numerical and personnel data were updated and verified as of August 27, 2026. The clubs’ results are interpreted as organizational case studies, not as evidence of direct causality. The article does not contain internal pricing, business targets, or unverified market claims.