Deal or No Deal? Revolutionizing NHL Trade Analysis.

Would you trade Jake Guentzel?? What if you had access to AI-driven insights into his predictive future performance and durability?? Would that help your decision?? NHL GM’s are feeling the pressure as the trade deadline looms just a few days away.? Do I trade for a player to help us win now?? Do I trade away a good player to build a future?? Do I gamble our future by moving assets in the hope of winning now?? The next few days will see NHL GM’s make critical decisions.? Championships, millions of dollars, and careers are all at stake.? ?

Probility AI has developed a revolutionary method of assessing and comparing players that can help GM’s make those tough decisions.? The Probility ARC (Availability-Reliability-Capability) uses AI models to understand players’ predictive futures and can provide insights into a player’s future performance and durability with high degrees of accuracy.?

We employed these tools to examine a few players currently bandied about as likely trade targets.? Player comparables are a tried-and-true method of evaluating potential roster moves. ?Probability AI has adopted that methodology but compares players based on our proprietary, predictive player durability and performance ratings.

Using those metrics to analyze potential trades yields some interesting results.? Starting with the defensemen, and beginning with traditional performance comps, Sean Walker and Noah Hanifin jump off the page as attractive targets.? Walker has met or exceeded his predicted games played, TOI, and total points for the 2023-24 season and is trending towards a career year.? Hanifin is trending to meet his predicted games played, TOI and total points.? Hanifin is out-producing Walker this season and carries a higher price tag to go along with that performance (Hanifin’s current AAV is $4.95M compared to Walker’s $2.65M).? Both players are proven NHL defensemen who will benefit a team looking to make a playoff run this season.?

But we can go further than that.? Probility’s unique metrics provide a deeper look into how the predictive futures of those players compare.? Hanifin has a higher ARC score measuring overall predicted durability and performance, marking him as the player with the higher potential upside.? Both players have nearly identical Availability scores indicating both have a substantial portion of their career ahead of them.? Hanifin scores higher in Reliability and Capability – he’s more consistent with more to give.? It gets interesting when you compare their Injury Index score, however.? Our models show Walker is significantly less likely to miss future games to injury than Hanifin.?

A GM looking to sign either of these players can hardly make a mistake – they are both good choices.? But which player better fits your organization?? Is it the player with higher potential upside in terms of productivity, but with a higher potential injury risk, or the player who may not produce as much but is likely to be a steady part of your lineup?? The answer depends on your team’s needs now, in the future, and differs by team.? And these tools won’t make the decision for you.? But a GM armed with this information is better positioned to make the right call.

Turning to the forwards, Frank Vatrano leads our list.? The winger has played every game for Anaheim this season and has already posted career highs in goals and total points with 21 games remaining in the season.? He already met or exceeded his predicted goals, assists, total points, and TOI for the 2023-24 season, and will exceed his predicted games played.? Of all forwards we considered for this study (Guentzel, Vatrano, Novak, Smith, Henrique, and Tarasenko) Vatrano leads his contemporaries in only one Probility durability and performance metric, but it's a biggie – Injury Index.? Vatrano has the lowest predicted future games missed to injury by a large margin.? He also places in the top 1/3 of the league in Availability, Reliability, and Capability scores as well as his overall ARC score.? And with the lowest AAV by $1M+, he’s a bargain.

Others on the forward list shine for various reasons.? Choosing between them is a complex matter and each team employs their own methodology to determine which player is the right fit for their team.? At Probility, we believe a player’s predictive future, particularly his predicted future availability, is a critical piece of information for GM’s looking to add players for the long term.? If you find yourself needing to make one of those decisions and want to learn more, give us a call.?

Predictive Futures – How Does it Work?

Probility assembled a unique data set with 20+ years of player performance, contract, and injury data for every player in the NHL.? If you've played a single shift in the league, you are in our data set. Our Data Team then analyzed that data using AI models, subjected to rigorous validation testing, to derive three key player’s predictive future metrics:

• Availability: The “Fuel Gauge” – a comparative ranking of the percentage remainder of a player’s expected total career.? Players near the end of their careers show low Availability scores as one might expect.? Younger players with strong predictive availability score very high here.?
• Reliability: The “Consistency” metric - a comparative ranking of a player’s season-by-season performance volatility.? Does the player regularly perform to expectations (and will they continue to do so), or do we see large performance fluctuations across his career?
• Capability: The “Output” metric – a comparative ranking of a player’s raw production expected for their career.

Each metric is ranked by position (forwards against forwards, defense against defense) to provide a focused view of how each player compares to his peers in each category.?

These metrics are then combined in the ARC Score – a comparative, overall measure of a player’s expected career durability and performance.? The ARC score supplies a snapshot of a player’s predictive future measured against every other active player at his position, providing a player comp like no other.

NHL GM’s can use these metrics, along with their institutional knowledge and wisdom, to make better decisions regarding roster optimization, for example, and assemble the best possible team.? It’s not a panacea, nor is it infallible. It’s a statistical model that predicts probable outcomes.? But for all that it is remarkably accurate across the league, with r2 values ranging from .62 for next season predictions to .90 for career.? It is designed to produce unique insights that can be combined with a team’s existing decision-making process to help make more-informed decisions.

We’ve developed additional tools to help improve organizational decision-making as well.? Our Scenario Builder allows GM’s and players to see how a player’s availability and performance metrics change when that player moves from one team to another.? A player’s predictive future is not always the same moving from team to team, and the Probility ARC platform allows the user to view those changes in real time.? For GM’s this tool provides insights into unique player values and potential arbitrage situations.? For the player, this tool helps them understand where their best possible predictive future lies.

The traditional way of comparing players is proven and has value. ?That process can be augmented, however, with new metrics on which to compare players. ?GM's employing these new metrics will make better informed decisions and produce superior outcomes.? AI is bringing more and better tools to the forefront to help.? Probility AI has harnessed that power and built a complete tool set to enhance decision-making in the NHL (and NFL) and is helping to change the game.