7-Factor Analysis: Simple Analytics for Any Level of Hockey

For decades, hockey has been a statistically simple sport.  Players who score lots of goals and get lots of points are really good.  Goalies who let in lots of goals and have low save percentages are usually not very good.  Hockey isn’t baseball.  It’s not a series of one-on-one matchups with isolated incidences.  It’s a free-flowing game with a wide variety of variables.

Beyond the obvious black and white statistics, everything has always been up for debate.  Hockey fans for years have debated Gretzky vs. Lemieux and Crosby vs. Ovechkin or Toews.  Words like character, toughness and leadership get thrown around.  There is banter about two-way play and clutch performances.  These debates are what make being a hockey fan fun.  It’s the endless comparison and argument over differing situations and variables.

For the scientific-minded fans, enough is enough.  No more “ya, buts” and “in my opinions”.  They want to know once and for all how to truly define a player’s worth.  Baseball has its WAR (Wins Above Replacement), why can’t hockey have its all-in-one determiner?

From the perspective of the franchises, who invest millions a year into extensive scouting blankets and video analysis efforts, why not try and find a way to gain an edge.  Maybe there is a formula or two out there that can more closely measure the true, overall value of a hockey player.  Everyone laughed at Bill James when he dabbled in sabremetrics, producing otherworldly statistical concoctions in his annual Baseball Abstracts.  Decades later, James is the undisputable golden boy of statistical analysis in baseball, revolutionizing the way everyone looks at player value.

Whether you like it or not, analytics in hockey is here and it doesn’t appear to be going anywhere anytime soon.  From Corsi to Fenwick, new ways of measuring player value beyond goals and assists are emerging and building momentum.

Since it appears these new statistics are becoming common speak and a big part of today’s game, how does this translate to junior and minor hockey?  Without expensive tools and video analysis, how can we find new ways to measure worth at low-tech levels of the game?

Here is a simple, low-tech statistical approach to capturing overall value in individual hockey players, entitled, “7-Factor Analysis”:


The way 7-Factor Analysis works is totalling up the following basic statistics (5 with positive effects on the game, and 2 with negative effects) to produce an overall “7-Factor Score”.


The Positive Categories (Worth +1 for each tick on the sheet):


Shots on Net

All shot attempts that reach the net, resulting in a shot on goal (goal or a save) or post hit. 

You have to shoot to score and not all goals are of the pretty variety.  Teams that attempt more shots that hit the net tend to be more successful.  This category is only capturing shots that get to the net and not shots that are blocked or miss the net.


Blocked Shots

Opposing team shot attempts that are negated via a blocked shot by a player.

Nothing is more frustrating than having your shot blocked.  Players who block a lot of shots tend to be the players who are in good positions defensively.  These types of players are worth their weight in gold.


Finished Checks

A player who delivers a check (any type of body contact—a bump a hard hit) to an opposing player.

One of the most tiring things in hockey is receiving a check (big or small) and playing through contact.  It absolutely saps the energy out of you.  Over the course of a 60 minute game, this can really wear down an opposing team.



Anytime a player creates a turnover for the opposition (This could come from stripping someone of the puck,  finishing a check and coming away with possession, or beating a forechecker to a dump-in and making a successful defensive zone exit via a pass or skating it out.)

The point of the game is to score more goals than the other team and you can’t do that when you don’t have the puck.  Valuable defensive players are able to create a lot of turnovers.


Completed Passes

A successful pass completed from Player A to Player B, maintaining possession.

One of the most important attributes of successful teams is puck control and puck movement.  Teams that control the puck through quick, successful passes, tend to maintain possession for longer periods of time, resulting in better opportunities to score.


The Negative Categories (Worth -1 for each tick on the sheet):



Anytime a player losses possession of the puck, to the other team, after being in control, other than from a successful shot attempt (Shot made it through to the net).  This includes: being stripped, a finished check resulting in loss of possession to the other team, a pass attempt that misses its intended target, and a dump-in where the other team gains possession.   

Giving up possession of the puck means you can’t attack and are forced to defend.


Missed/Blocked Shot Attempts

Anytime a player attempts a shot that doesn’t result in either a shot-on-goal or post hit. 

When players attempt shots at the net that are blocked or miss the net, they are risking a turnover and limiting their chance to score a goal.  With possession being a major key to success, you want to ensure that your hard fought efforts to gain possession at least result in a shot on goal.


So how does it work?  Easy.  It is easily completed as a one or two person job.  All you need is a template that lays out all the columns for the different categories for each player.  During the game, you add a quick tick to the appropriate box for every relevant event.  At the end of the game you tabulate all of the scores for each player by subtracting the amount of ticks in the negative boxes from the amount of ticks in the positive boxes to produce an overall “7-Factor Score”.


Template Example:


Positives Negatives
Player Shots on Net Blocked Shots Finished Checks Takeaways Completed Passes Turnovers Missed/

Blocked Shot Attempts

Overall Score


Player 1
Player 2
Player 3
Player 4
Jamie McKinven
Author / Blogger at glassandout.com
Jamie McKinven, author of “So You Want Your Kid to Play Pro Hockey?” and “Tales from the Bus Leagues,” is a former professional hockey player who played in the NCAA, ECHL, CHL and Europe.

About Jamie McKinven

Jamie McKinven, author of “So You Want Your Kid to Play Pro Hockey?” and “Tales from the Bus Leagues,” is a former professional hockey player who played in the NCAA, ECHL, CHL and Europe.

View all posts by Jamie McKinven →

8 Comments on “7-Factor Analysis: Simple Analytics for Any Level of Hockey”

    1. Good question. I hadn’t really thought of that. Since receiving a pass is a skill as much as delivering a pass, you could count it as a plus for both players. At higher levels, I would only count the plus for the person making the pass.

  1. I like where you’re going with this, but if you don’t somehow account for actual puck possession time / opportunities, you’re going to end up with a lot of mud in your 7 factor score.

    First of all, it’s easy to think up scenarios in which both of your “negative” metrics could actually be deemed positive metrics, if you don’t account for possession time / opportunities. Suppose a Player A has 3 times the turnovers of Player B. Should that reflect poorly on Player A? Maybe, but what if Player A has the puck 5 times more often than Player B? Similarly, if Player A gets more of his shots blocked (or more of his shots miss the net), he might be less efficient than Player B, but what if Player B has less blocked shots / missed nets because he never has the puck, or when he does have the puck he’s in the defensive zone instead of the offensive zone? Players who have the puck a lot are going to turn it over more, they’re going to miss more nets, and they’re going to take more shots that end up blocked.

    Likewise, you could end up counting events that are actually detrimental to your team with your “positive” metrics. If you count finished checks as positives, you’re going to end up giving a lot of positive points to players who don’t have the puck very much. By rule, you can’t hit a guy who doesn’t have the puck, and since there’s only one puck on the ice, you can’t really throw a hit if you (or your team has the puck). Therefore, any finished check is going to register at a moment when the hitting player’s team doesn’t have the puck. It’s fairly consistently true that teams who frequently outhit their opponents also frequently lose.

    The same logic applies to takeaways (if your team had the puck all game, you would register zero takeaways) and to blocked shots (you can’t block a shot unless the other team has the puck and it’s direct toward your net).

    I’d suggest you’d be far better off in tracking shot attempts. Count shot attempts for (which equals shots on goal for, shots for that are blocked, shots for that miss the net, and shots for that go in the net) as positives, and count shot attempts against (shots on goal against, shots against that you block, shots against that miss the net, and shots against that go in your net) as negatives. This stat (otherwise known as Corsi) will far better reflect who is controlling possession.

    1. Thanks for commenting Danton. I think any way of statistically trying to quantify value in hockey comes with plenty of room for criticism. The 7-Factor Score was something we developed about 5 years ago when I was coaching Jr. A hockey in Ontario. We wanted a way to score players based on what we felt were the most important aspects of the game, regardless of what type of player you are (forward, defense, 4th line grinder or top scorer). We found that the scores were very accurate in determining quality of play in players across the board (no matter if they played 7 minutes a night or 30). Your arguments that players who possess the puck more and play more minutes are going to have more turnovers, missed shots, etc., is absolutely correct. However, these players will also have more completed passes, successful shots on net, and turnovers created, so really, it ends up balancing out. Conversely, players who aren’t relied on for offense, can have valuable stats like blocked shots, finished checks and turnovers created tracked. These players, even if they play minimal minutes, will also miss passes, turnover pucks and get shots blocked. The 7-Factor Score allows you to capture all aspects of the game from blocking shots and creating turnovers to creating offense through successful passing, shooting and puck possession. It isn’t perfect (I don’t think any stat in hockey is, the game of hockey is too fluid with too many skewed variables), but it was very helpful for what we were looking to do at the low-tech Tier II Jr. A level. It also did another very important thing for us. It helped to develop a benchmark for each player; something to build off of and measure themselves against, while striving for personal improvement.

      As for Corsi and Fenwick and counting shot attempts for and against, I’m not a big fan. There is too much more to consider in a hockey game while evaluating players game-in and game-out than shot attempts for and against. Teams become too focused on just throwing everything at the net from everywhere. This creates more turnovers and ultimately you see lower quality scoring chances.

  2. I see where you are coming from, I just think counting some of these things is an over-complication without any benefit.

    I think that we should count things that we know to be indicators of how the team / player is performing in relation to things that impact the outcomes of games. Underlying my approach to analytics is one important assumption: having the puck leads to winning, so having the puck is better than not having it. That may be debatable in certain situations, but I think the analytics community has produced enough evidence to prove it’s almost always true, or at least to convince me that it’s almost always true. That said, all of what I’m about to say rests on that assumption, so if anyone disagrees with that assumption, they’re not likely to relate to what I have to say.

    The reason I’m not a big fan of counting, say, blocked shots, takeaways and finished checks as positives is because I think they are actually more indicative of a player whose team DOESN’T have the puck when he is on the ice. I’m not saying that players shouldn’t block shots, attempt to take pucks away from opponents, or finish their checks when the other team has the puck. I’m saying that if a player is having to do a lot of that, it’s because the other team has the puck a lot when he is on the ice, which should reflect negatively on that player’s contribution to his team – not positively.

    So in the 7-factor score, you’re counting some things as negative even though those things actually might be positive indicators (raw number of turnovers, raw number of shots blocked by opponents) and you’re counting other things as positive that might actually be negative (raw number of shots blocked, raw number of checks finished, raw number of takeaways). I agree that some of these positive (but actually negative) and negative (but actually positive) things would cancel each other out, but then all you are left with is a convoluted data set where the events captured aren’t actually reflecting the players’ net-positive contribution to the team’s performance.

    I think if you could come up with some way of weighting these events by opportunity, you’d have a better argument. For example, if you could express finished checks in terms of the number of opportunities to finish a check, then it’s useful. If Player A has 25 finished checks in a game, and Player B has 5, under your current system Player A is going to have 20 more positive points in his score than Player B (other things equal). But what if Player A has 50 opportunities to finish checks, and Player B only had 7 opportunities because your team had the puck most of the time Player B was on the ice? The same logic could apply to takeaways (successful takeaways weighted by opportunities for a takeaway) and shot blocks (successful shot blocks weighted by opportunities to block a shot). Adding up all of these different types of events without any sort of weighting just muddies the picture, in my opinion.

    I think it’s also notable that studies have shown these “real time stats” (hits, blocked shots, turnovers) actually have very little impact on the outcome of the game, mainly because of what I’ve explained above – they’re inversely related to puck possession. Here is a good article that provides more detail: http://blogs.thescore.com/nhl/2013/02/25/breaking-news-puck-possession-is-important-and-nobody-told-the-cbc/ .

    With respect to counting Corsi and Fenwick (shot attempts), I’ve yet to hear of doing so resulting in teams starting to throw every puck at the net from everywhere just to pump their Corsi / Fenwick numbers. It’s a commonly cited argument that I don’t think has any grounding in reality. I think net shot attempts, especially when used as a team stat, are very useful in providing a general sense of your team’s performance. Your goal is to win on the scoreboard, not on the net shot-attempt count, but it’s widely proven that teams winning the net-shot attempt count win more games on the scoreboard over a longer period of time.

    With our midget team, we track every shot on goal, missed net, blocked shot and goal (for and against) and every player who is on the ice when each of those events occurs. It’s been extremely useful in helping us to confirm or refute what we think we are seeing on the ice. It helps us avoid making snap decisions on line shuffling, systems changes, etc. and helps us to determine whether a player and team scoreboard success (or lackethereof) is really due to the way the team is performing when they’re on the ice, or more so a result of luck (good or bad).

    1. Thanks again Danton! This is a well put together view on analytics and strategies for evaluating performance. You’ve provided a lot of good, verifiable information on ways teams can accumulate and assess different statistics. I have to admit, I was dead against the “Analytics Revolution” in hockey in the beginning. Since I got over my “old school” mentality and allowed myself to question every assumption I grew up with about the game, I have learned a great deal and am eager to continue doing so. You’ve given me (and anyone who reads your comments) lots to delve into and learn about. It’s all fascinating stuff. Thanks again for weighing in and helping me to step back and look at things from a different perspective!

  3. I would like to point out the 6/7 Factors mirror the game stats kept in conjunction with the Intelligym. The exception is “takeaways”.

  4. The thing I like about McKinven’s 7 factors is, they can be used for teaching and coaching players to get better. Corsi, and these things are great when evaluating a player; to show if he is good, or bad. Good for deciding who to trade for, or who to sign, or who to play more as a coach. But as McKinven points out in his reply, to show a player a bad corsi rating, and to explain to him he didn’t get enough shots on net, often what the player takes away from that is, “shoot the puck more.” “Get more shots on net.”

    But getting shots on net is a result of the cumulative effort of doing all the little things right; creating turnovers, not turning the puck over, passing, position, etc.

    So, despite the 7-factor analysis not being that great on actually quantifying a player’s worth. I think it has value in quantifying the little things necessary in order to boost one’s corsi at the end of the day. Afterall, if keeping these stats encourage your players to turn the puck over less, work harder on the forecheck to earn a take-away, or block shots, ultimately your team’s corsi score will improve, and your team will be better because of it.

    Although I do agree that the 7-factor analysis should be weighted for opportunities, although that’s a bit time consuming, and subjective. Because without it, a player might believe he’s not hitting enough, or not blocking enough shots, when in fact, he’s got the puck all the time through his skill, and therefore relatively doesn’t get as much opportunities for that then say, a grinder might.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.