Archive of Our Own

YAOI ON ICE: A Study On The Impacts of Male Queerness On Professional Hockey Skill

ABSTRACT 

With four of the best current NHL players officially out as gay or bisexual, fans and critics alike have speculated on a possible link between homosexuality and hockey skill. Using points per game data and a comparison of queer and straight players’ statistics, this study reimagines a popular P/PG prediction model to measure a potential “queer effect.” While limited by its small sample size, it found evidence that suggests queer players may be more naturally skilled at hockey and discusses popular theories regarding why. 
Keywords: Ice hockey, professional sports, homosexuality

INTRODUCTION

Ice hockey in the United States and Canada is a culturally conservative institution. Like most professional sports, hockey is traditionally male-dominated. This is true across all measured populations: players, coaches, support staff, sponsors, and fans¹ alike, men outnumber the women. In recent years, the success of the emerging PWHL has moved the needle slightly, but in the United States and Canada, professional male hockey players^{2, 3} still outnumber female players⁴ approximately 9.6 to 1. It is also—referring solely to the men’s leagues due to a lack of data available for most of the women’s—the whitest professional team sport. A 2018 Associated Press article on extreme racism and on-ice hate crimes faced by NHL players measured the league as 97% white. (The article has since been deleted, but remains accessible through the Internet Archive’s Wayback Machine.⁵ A more recent diversity report by the NHL claims to be 83.6% white, but this figure includes all employees, not just players. Other professional sports leagues reflect much more racial diversity, including the MLB (61.3% white)⁶, NFL (25-27% white)⁷, and NBA (19.1% white)⁸ 

Until quite recently, the league was also assumed to be 100% heterosexual. Then, within the last five years, four current players were revealed to be members of the 2SLGBTQA+ community, starting with Scott Hunter’s dramatic on-ice kiss with his boyfriend immediately after winning the Stanley Cup⁹ and ending (so far) with the scandalous outing of longtime rival captains Shane Hollander and Ilya Rozanov, outed both as queer and as being in a relationship with each other¹⁰. Though Hollander and Rozanov’s outing was the only one to result in professional consequences from the League, all four currently out players have received intense backlash both online and on the ice from people who believe that hockey should remain a “straight” space. 

There is no rational basis for these objections. Claims that gay players will make the locker room an unsafe space would hold more weight if there had been any issues, complaints, or allegations in the almost two decades¹¹ since the oldest confirmed queer (former) NHL player, retired goaltender Eric Bennet, was signed to the New York Admirals. If the players were not predators before coming out, there is no reason to believe that coming out would suddenly make them desire to behave inappropriately toward their teammates. Worries about behavior towards fans are similarly unfounded, and complaints that these men are supposed to be “role models” are both rooted in homophobia and applied with severe bias, as these players have faced far more public outcry than the dozens of confirmed adulterers, abusers, and rapists in the league¹² combined. 

However, a new kind of objection has recently begun to circulate in certain online spaces: the idea that being gay is a form of cheating, and that participation in homosexual acts has given these players an unfair competitive advantage. This paper will explore the reasoning behind this theory and examine whether the out queer players’ professional statistics differ from confirmed or assumed heterosexual players’s statistics in a way that demonstrates a statistically significant positive relationship between identifying as homosexual or bisexual and hockey performance. 

LITERATURE REVIEW

As in many male-dominated fields, participation in sports, whether as an athlete or a spectator, has historically been associated with masculinity. Women’s sports are becoming more popular for athletes and audiences alike¹³ and women’s viewership of men’s sports is increasing enough that professional men’s leagues are reconsidering their advertisement and audience engagement strategies¹⁴, but stereotypes and sexism still prevail. In an intersection of homophobia and misogyny, many sports fans are baffled by the idea that a gay man could play a professional sport. 

The transitive property occurring looks like this: man + gay = feminine, feminine = bad at sports, therefore, man + gay = bad at sports. A gay man can’t play hockey because he’s “girly,” and everyone knows girls can’t be hockey players. These stereotypes and assumptions present themselves oppositely depending on the perceived “gender” of the sport: hockey players must be straight because hockey is masculine, but male figure skaters or male gymnasts are probably gay because figure skating and gymnastics are feminine. It appears in our language regarding athletics as well: hockey is a sport, no one questions that. Football, baseball, and basketball are sports. Figure skating? If it’s in the Olympics, it probably has to be. Gymnastics? I guess. But competitive cheer? Intensely demanding, absolutely brutal on the body, and beyond the physical capabilities of your average person, whether they are male or female, a professional athlete or not, but go ask a man if cheerleading is a “real” sport. 

Additionally, female sports leagues do not have these same concerns about queerness. It’s well-known (and stereotyped, with its own negative effects) that in professional women’s sports, lesbians abound. Hollander and Rozanov may have broken the internet with the reveal of their relationship, but in the PWHL, players dating or marrying teammates (or rivals!) is the opposite of uncommon. It’s frequently celebrated, as shown by the iconic “wife line” that is Marie-Philip Poulin and Laura Stacey (Montreal Victoire)¹⁵ and the rivals-to-lovers Olympian captains Julie Chu and Caroline Ouellette¹⁶ Whether queer couples are playing against or alongside each other, there is already a precedent in professional hockey, which homophobic hockey fans (or those who profess to be “not homophobic, just concerned about fairness”) who accused the two rivals of throwing games for each other conveniently ignore. 

Notably, the people making these accusations disproportionately make them against Hollander, one of the few Asian men in the league. There’s a lot to unpack there: racist stereotypes that Asian men are feminine + racist stereotypes that Asian women are submissive = racist assumption that Hollander is the “bottom” in his and Rozanov’s assumed sexual relationship and therefore the one who would allow his boyfriend to win. Racism in the NHL is referenced throughout this paper due to its pervasiveness in the League and its fans, but is not this paper’s focus. For a more thorough exploration of that issue, I refer you instead to Leveling The Ice: Confronting Racism In Hockey by Steven Sandor¹⁷ Otherwise, bear in mind that when considering any homophobia faced by any queer man in the League, it’s safe to assume Hollander experiences it more often, more severely, and entirely uniquely due to the intersection of his multiple marginalized identities. As he is the sole out queer person of color in the NHL at this time, there is not yet any scholarship on the effects of this intersectionality on his treatment by other players or by fans. 

Concerns about game fixing in the NHL are not new. Brian Tuohy, author of multiple books about sports gambling, game fixing, and other conspiracies and allegations across several professional sports leagues, claims to own a “(somewhat rare) 1969 first edition” of Hockey! The Story of the World's Fastest Game in which former NHL referees speak candidly about being instructed to “not only keep games close, but at times, dictate the outcome.” However, current editions of this book do not contain this chapter¹⁸. Mr. Tuohy doubts the integrity of the NHL over fifty years later, saying “once a cheat, always a cheat.” Various fan forums make similar claims, providing evidence of inconsistent or illogical calls made by referees. Hollander is not even the first NHL player to be accused of “throwing” a game. However, there is no evidence that this has ever occurred. The NHL thoroughly investigated San Jose’s Evander Kane after his wife alleged on social media that he had been conspiring with bookies, gambling on the outcomes of NHL games, and sabotaging his own team to win bets. Analysis of several years of footage of his gameplay and Sportsradar bet monitoring reports found no signs of a deliberate poor performance and Kane is still an active player¹⁹. Ottawa’s (ironically) Shane Pinto was suspended for sports-related gambling, the first NHL player suspended for gambling since the 1940s. While they did not release details of what Pinto actually gambled on, they did confirm that none of his betting was related to NHL games²⁰. These are the only semi-credible allegations of game fixing performed by a player in decades, both were investigated thoroughly, and both were disproven. And still, from the moment Hollander and Rozanov were outed, Metros fans have been calling it “obvious” that Hollander has been throwing games. 

To be clear: they are alleging that Shane “JHL MVP, JHL Playoff MVP, JHL Most Goals Scored, WJC Silver Medalist, WJC Gold Medalist, Second Draft Pick, 2011 All Star, 2011 Rookie Of The Year, 2012 All Star, 2013 All Star, 2014 All Star, 2014 Olympic Silver Medalist, 2015 All Star, 2015 Stanley Cup-Winning Captain, 2016 All Star, 2016 Stanley Cup-Winning Captain, 2020 All Star, 2020 Stanley Cup-Winning Captain, 2020 Conn Smythe Trophy Recipient, 2021 All Star, 2021 All Star Fastest Skater, I Would Need An Appendix To List All Of The Records He’s Broken"²¹ Hollander has been deliberately underperforming. 

First of all, hockey is an incredibly fast-paced and difficult sport. For Hollander and Rozanov to choose the outcome of a match, they would have to know that only Hollander’s actions would affect the Metros’ score. If the two were successfully conspiring, it would require that either the entirety of the Metros were knowingly participating (which no one else has accused them of, and no player has accused Hollander of asking them to do) or that the entirety of the Metros are useless dead weight that only exist to have enough bodies on the ice while Hollander single-handedly wins three Stanleys. Accusing Hollander of deciding games means accusing the rest of the Metros of total incompetence (which fans do have the right to do—however, considering the bulk of people demanding consequences for Hollander’s alleged cheating are Metros fans hoping to see him removed from the team, they may wish to strongly consider what a Hollander-less Metros would look like. There would be three fewer banners in their barn, for a start.)

Second, every move a player makes on the ice is monitored from several angles, measured, and recorded. Every single detail about Hollander’s participation in games is publicly available: average minutes played per game, average shots on goal per game, average goals, total penalty minutes per season, time spent in possession of the puck, and more. If Hollander played worse against Rozanov, it would be noticeable. Fans, sports analysts, commentators, teammates, and coaches would have noticed years ago if any of these statistics dropped or spiked in Boston, against Boston at home, when Rozanov was on the ice, or for any other reason. An NBA fan once analyzed four seasons’ worth of data to prove that, to a statistically significant degree, James Harden’s performance is inversely correlated with the ratings of local strip clubs: that is, in cities where the strip clubs are better, Harden plays worse, and in cities where strip clubs are worse, Harden plays better²². With the amount of scrutiny Hollander, Rozanov, and their famous rivalry are under at all times, on and off the ice, any correlation would have been noticed, publicized, and dissected years ago, and if Hollander were noticeably underperforming, it would be the responsibility of his coaches to remove him from the ice for the good of the team, if the fans didn’t eat him alive for it first. (Unless, of course, Hollander on a bad day is still better than all of the Metros on their best days. Again, if this is the case, the Metros have bigger issues.) 

Additionally, these allegations are disrespectful to Rozanov and the Boston Bears, an incredibly gifted player and a strong team in their own right. Claiming that Hollander threw games implies that Rozanov and the Bears could not have beaten him otherwise. Again, to be clear: this implies that Ilya “WJC Gold Medalist, WJC Silver Medalist, First Draft Pick, 2011 All Star, 2011 Rookie of the Year Nominee, 2014 Stanley Cup-Winning Captain, 2014 Hart Memorial Trophy Winner, 2014 Olympian, 2017 All-Star, 2017 Hart Memorial Trophy Nominee, 2019 All-Star (declined), 2021 All-Star, Neck And Neck With Hollander For Number Of Records Broken”²³ Rozanov cannot win a game against Hollander unless Hollander permits it. 

Rozanov’s team beat Hollander’s in World Juniors; has Hollander been throwing games for him since they were teenagers, or did Rozanov forget how to play hockey when they got together? And where is the sudden shift in both of their statistics that illustrates the exact date they got together, if they began playing differently due to being emotionally compromised?  They’ve been nominated for several of the same awards, how did they notify the awards committees which of them they’d decided would be the recipient? If Hollander only loses when he chooses to and only chooses to when emotionally compromised by his boyfriend, is he dating someone on every team that has beaten the Metros since the beginning of his captaincy? And if Rozanov can only win when his boyfriend lets him, is he dating someone on every team that ever lost to the Bears during his captaincy? 

No. The theory that Hollander has thrown games for Rozanov due to their romantic relationship is based entirely in homophobia and racism, not in logic or data. This analysis has not provided evidence of game tampering. Instead, it has provided evidence that Hollander and Rozanov are indisputably the best in the league. 

Which leads us back to our initial question. If none of the Metros or Bears’ combined four Stanley cups under Hollander/Rozanov captaincy were won by cheating, then they can only be attributed to skill. Since the beginning of the Hollander/Rozanov Era, four of the seven seasons’ cups have been won by teams captained by one of the two. One of the three non-Hollander/Rozanov cups was won by the New York Admirals, captained by Scott Hunter. This demonstrates an interesting pattern. Rozanov, captaining the Bears to one cup, is bisexual. Hollander, captaining the Metros to three cups, is gay. Hunter, captaining the Admirals to one cup, is gay. Five of the most recent cups, a staggering 71.43%, were won by teams with queer captains.²⁴ This is what sparked online speculation that being queer may somehow be providing teams with a competitive advantage. 

This is possible. While sexual diversity in professional hockey players only recently became a question, the impacts of sexuality on athletic performance have been studied in the past. A 2014 survey of over twelve thousand participants ages 12–22 found that individuals who self-identified as sexual minorities (members of the LGBTQ+ community, any sexuality other than entirely heterosexual) reported on average between one and three fewer hours engaged in moderate to vigorous physical activity per week than their heterosexual peers, and that they were 46–76% less likely to participate in team sports than same-gender heterosexuals.²⁵ Similar data from the University of British Columbia, studying coached sports participation of boys grades 7–12, found that not only do gay and bisexual teen boys participate in team sports at half the rate of their heterosexual peers, this disparity has been worsening over the last decade and a half. Participation by straight boys has remained mostly constant, participation by bisexual boys has decreased, and participation by gay boys has sharply declined and is still trending downward.²⁶ The small amount of participation in professional sports by queer men amplifies the perceived uniqueness of their skill levels. Of an entire league of men, only three* are publicly out as gay. Of an entire league of men, only one has the career records and awards of Hollander. When those small populations overlap, data can be very dramatically skewed. With such a small sample size, one man playing as well as Hollander singlehandedly boosts the statistics of the gay pro-hockey population. 

*(A reminder that Troy Barrett, Ottawa Centaurs, is also gay. As he came out by Instagram post, not by kissing his boyfriend on the ice after winning the Stanley Cup or being outed by a viral video making out with his rival, and he is not the captain of a Cup-winning team, he tends to fly under the radar when it comes to discussions of queer players. As an outspoken advocate for resources for survivors of sexual assault, he is also frequently deliberately removed from these conversations by those who would rather cheer for a rapist than a gay man. These people are disturbingly common in the NHL, but, as with the topic of racism, I must refer you to the work of others: Maizie Scwets for a cultural overview²⁷, Raul Daffunchio Picazo for real names and recent cases.²⁸ Please mind your mental health and read with care.) 

However, of five* currently out players in all of the NHL’s history, we are not looking at four mediocre players and one Hollander. We’re looking at one bisexual Stanley-Cup-winning goalie (Eric Bennet, New York Admirals, retired), two gay Stanley-Cup-winning Captains (Hollander, Montreal Metros; Hunter, New York Admirals), one bisexual Stanley-Cup-winning Captain (Rozanov, formerly Boston Bears, currently Ottawa Centaurs), and one gay player who has never won a Cup but is a skilled player on a team that experts predict will win their first cup in recent memory within the next three years, at most (and will definitely win next year of rumors of a Hollander acquisition come true) (Barrett, Ottawa Centaurs).  

*There are also rumors about the sexuality of retired player Ryan Price, best known for his role as an enforcer, through which he made players who bothered his teammates “pay the Price.” Social media posts made by fans of musician Fabian Salah seem to indicate a romantic relationship between the two. However, due to a lack of official confirmation from Price himself, Price has been excluded from this study.

It is entirely possible that the statistically anomalous skill of these players has nothing to do with their sexuality, and the facts that the top players in the league are all queer and that five of the past seven Stanleys were won by teams with queer captains is a coincidence. However, a 2022 study that took advantage of publicly known sexual diversity in the WNBA examined the performance statistics of players of various sexualities and found notable differences. The report claims that overall performance was similar regardless of sexuality, but documented the following: 

“Lesbian guards are more accurate shooters with a significantly higher field goal percentage than straight guards, and lesbian forwards recorded marginally more steals and assists than straight forwards. Straight females committed more personal fouls than lesbians, especially if they had a female coach in college (regardless of the sex of their WNBA coach).”²⁹

The study also references naturally occurring measured differences in testosterone levels between cisgender heterosexual and queer women. While it hesitates to ascribe any difference in performance to testosterone levels due to the worsening of transphobia in professional sports and increased threat of invasive medical testing if such differences were found to impact skill, it does note that a correlation is possible. This topic will be further explored in the Discussion section. Specifically studying women who play professional basketball, there is a measurable impact of sexuality on athletic performance. If these findings can be replicated in studies of other genders in other sports, it is reasonable to question if queer professional hockey playing men play differently, perhaps even better, than their straight peers as a direct result of their sexuality. 

METHODOLOGY

The skill measurements used in this study is average points scored per game, with “points” including both goals and assists. These numbers were averaged across the past ten seasons, encompassing the entirety of Hollander, Rozanov, and Barrett’s careers and the last decade of Hunter’s. Bennett has been removed from the data pool both due to his retirement and difficulties comparing the skills of goalies to players in other positions, resulting in a sample size of n=4. Currently, out queer players make up less than one percent of the league (0.38%). I have chosen to proceed with this study due to the ability to include 100% of the current out queer players but note that for the results to be strongly statistically significant instead of simply providing some evidence of a possible correlation, this study would need to be repeated at a later time when out queer players make up a higher percentage of the league. My decision to conduct this study at this time instead of waiting for that day was motivated in part by the desire to examine the “concerns” of homophobic fans that the inclusion of queer players would somehow worsen the quality of the League, and partially by the desire to, should my results disprove these “concerns,” increase the odds of that day someday occurring. 

Data for our queer players was compiled from averaging their personal career statistics. This data is publicly available through the NHL.³⁰ Data for the heterosexual League was calculated by removing all queer players’ statistics and calculating new League averages in the same manner.³¹ See Appendix A for more detail on the process of data compilation for both demographics. 

The Younggren and Younggren model of determining statistic probability³² determines a player’s probable points per game based on experience level, over/under average retirement age, top speed, position played, and line played: 

E[P/PG] = β₀β₁Rookie + β₂Over28 + β₃[Speed] + β₄[Position] + β₅[Line] +  ϵ_I

Variables in [brackets] are subdivided into binary subvariables based on the standard classifications for each category. The most recent calculated values for β are as follows:

β₀ = 0.35366, β₁ = -0.12732, β₂ = -0.08094, β₃ = [1: 0.10413, 2: 0.20529, 3: 0.30922, 4: 0.42094, 5: 0.54327 ], β₄ = [L: 0.12672, R: 0.14843, C: 0.34205, D: -0.00428, G: -0.4067,  β₅ = [1: 0.411083, 2: 0.08376, 3: -0.02931, 4: -0.04582]

Applied, the average points scored per game of a non-Rookie, non-28+, speed category 3, second-line left wing should be approximately 

E[P/PG] = 0.35366 - 0.12732(Rookie) -  0.08094(Over28) + 0.10413(Speed1) + 0.20529(Speed2) + 0.30922(Speed3) + 0.42094(Speed4) + 0.54327(Speed5) + 0.12672(PositionL) + 0.14843(PositionR) + 0.34205(PositionC) - 0.00428(PositionD) - 0.4067(PositionG) + 0.411083(Line1) + 0.08376(Line2) - 0.02931(Line3) - 0.04582(Line4) +  ϵ_i

Or:

E[P/PG] = 0.35366 - 0.12732(0) -  0.08094(0) + 0.10413(0) + 0.20529(0) + 0.30922(1) + 0.42094(0) + 0.54327(0) + 0.12672(1) + 0.14843(0) + 0.34205(0) - 0.00428(0) - 0.4067(0) + 0.411083(0) + 0.08376(1) - 0.02931(0) - 0.04582(0) +  ϵ_i

Or: 

E[P/PG] = 0.35366 + 0.30922 + 0.12672a + 0.08376 +  ϵ_i

So the expected points per game for a non-Rookie, non-28+, speed category 3, second-line left wing is 0.87336, plus or minus a fifteen percent margin of error.

This study tested the theory that this margin of error can be decreased when controlling for an additional factor: sexuality. The baseline expected number of points per game by a player of any sexuality, before considering age and position, is 0.35366. However, when separating players by sexuality, the expected number of points per game before considering age and position is 1.4825 points for queer players and 0.34803 points for straight players. Looking at this gap, it is entirely reasonable that people would begin to suspect a correlation, but this is also a sample almost entirely made of first-line Centers not in an age group that impacts their likelihood of scoring, giving them an automatic advantage over an average of the entire league. A group of all-Centers will score more goals per person than a group that includes goalies, regardless of any other factor, so sexuality has to be fairly weighted against all other controls. By removing the data from our four queer players and recalculating the β values based on actual P/PG scores over the last decade, taking only the data from our four queer players and recalculating the β values based on actual P/PG scores, then comparing the differences in the “queer” and “straight” β values, I attempted to isolate the measurable “queer effect” on hockey skills, if such a thing exists. 

RESULTS

This method of study indicates a potential correlation between sexuality and hockey performance is possible. In the new equation Ê[P/PG], the value of β₀ is .0056335 lower. All other positive values are decreased similarly. Negative values are largely unaffected. The new equation is as follows:

Ê[P/PG] = β₀ + β₁Rookie + β₂Over28 + β₃[Speed] + β₄[Position] + β₅[Line] +  β₆Queer] + ϵ_i

with values 

Ê[P/PG] = 0.34803 - 0.08732(Rookie) -  0.06005(Over28) + 0.09401(Speed1) + 0.18492(Speed2) + 0.27687(Speed3) + 0.40992(Speed4) + 0.51184(Speed5) + 0.10592(PositionL) + 0.10779(PositionR) + 0.22092(PositionC) - 0.00425(PositionD) - 0.4067(PositionG) + 0.35772(Line1) + 0.07264(Line2) - 0.02921(Line3) - 0.04581(Line4) +  β₆[Queer]  + ϵ_i

Tested with the data of three randomly selected heterosexual players per team, this formula, with no value assigned to the “queer factor” to simulate Queer = 0, found accurate P/PG with an 11% margin of error, an improvement from the previous 15%. Plugging in the data of our four queer players (see Appendix C), averaged and weighted with the new 11% margin of error, found a β₆, or “queer effect,” of 0.27149. This implies that being queer may improve a player’s likelihood of scoring approximately as much as having a level three top speed. Being queer may have more of an impact on scoring chances than playing Center. This is no guarantee of success—a queer player with a level one top speed is still substantially less likely to score than a straight player with a level five top speed—but if these results can be replicated with a larger sample size, they could have significant ramifications on sports culture worldwide. 

DISCUSSION

If further testing indicates a natural predisposition to hockey skill for queer men, it’s worth discussing why we haven’t noticed it before. The answer is simple: homophobia, and the tendency to take one step forward, then two steps back. A small gesture is done, then immediately undone and worsened. Some teams created Pride nights (to some extent—see the Columbus Blue Jackets’ “Hockey Is For Everyone” night that attempted to be a Queer Pride and Disability Pride game at the same time, subsequently not quite accomplishing either³³), just for the League to promptly ban Pride jerseys (which were only worn during warmups, not during play) and Pride tape to minimize “distraction.”³⁴ Players throw slurs and other homophobic names at other players who are playing frustratingly well, other players who are playing poorly, and at referees who make calls they don’t like, and based on their willingness to do so in front of their teams, their coaches, cameras, and the world, they have not been taught to worry about repercussions.³⁵ It’s no wonder no one came out before Scott Hunter—and considering the years he and the others spent playing while closeted, we will never know how many queer players came before him. So what will happen if further testing reveals that queer men are simply naturally better than hockey? I see several options: first, homophobia in hockey decreases, resulting in a decrease in homophobia in professional sports, and subsequently, in countries that value hockey. If hockey is a measure of masculinity that straight men respect, maybe it can make them respect queer men, too. This is the least likely option. 

More likely than that: as hockey becomes synonymous with queerness, the popularity of the sport will decrease. Similar to figure skating or cheer, it will become another sport that straight men acknowledge takes skill but isn’t really a real sport. Ticket sales, TV viewership, and sponsorship will decline as straight men with fragile masculinity refuse to accept that queer men can be athletic. 

Most likely: increased volatility of homophobia until hockey is safely heterosexual again, regardless of the skill levels of the players involved. As we saw when a trans woman had the audacity to tie for fifth place in a swimming competition, insecure athletes with bruised egos would rather attempt to ban an entire marginalized population from sports participation across the board than admit that maybe someone they don’t respect might still be good at things they do.³⁶ Unfortunately, this has already started. The netizens calling for League consequences for queer players and claiming a skill difference between gay and straight hockey players were not citing research done on lesbians in the WNBA, or doing any of the math I’ve done here. They’re pointing fingers and making guesses. According to several off-putting conversations on Reddit and similar forums, this skill difference is very real, and its cause is, allegedly, testosterone. The positive impact of testosterone on athletic performance is well-documented, though the exact causes and effects are still being studied, to the extent that medically supplementing testosterone is considered a form of doping and banned from most professional competitive sports.³⁷ The theorists speculate that queer hockey players, most notably Hollander and Rozanov due to their relationship with each other, are violating League anti-doping policy by introducing exogenous testosterone to their systems anally and orally. 

There are so many problems with this assertion. First, the League mandates extensive random testing of players as part of their anti-doping policies. Similarly to a theoretical spike in their statistics after getting together, if Hollander and Rozanov were supplementing their own natural testosterone to improve their gameplay as a direct result of their relationship, a difference between their pre-relationship established baseline and post-relationship new testosterone levels would have been noticed and called out immediately. Rozanov especially, as a Russian national, is under heightened scrutiny due to Russia’s documented history of doping at the Olympic level.³⁸ Rozanov has emphatically denied any involvement, backed by character testimony from all of his NHL teammates and coaches, and voluntarily submitted himself for additional screenings, all of which have come back clean of any evidence of wrongdoing. 

Second, it doesn’t actually work that way. While testosterone can be taken orally³⁹ and can theoretically, experimentally, medically inadvisably be taken rectally,⁴⁰ the amount of testosterone actually present in ejaculation is both widely variable and fairly negligible.⁴¹ The amount needed to impact athletic performance in this manner would be physically impossible from a bodily production standpoint and extremely difficult to the point of medical distress from a recipient’s standpoint. Sexually transmitted testosterone is not impacting any players’ abilities on the ice. 

Additionally: we do not know the details of the private sexual lives of these athletes, nor should we. Sexuality and romantic orientation are a far more nuanced topic than most cisgender heterosexual people assume, with near-infinite possible categories, identities, and intersections. Through personal statements or interviews, two of the out players studied have labeled themselves as homosexual/gay and two have labeled themselves as bisexual. A preference for these labels, especially for use to a broad audience likely to have a very narrow understanding of the intricacies of gender, romantic orientation, and sexuality, does not preclude the possibility that any of these men also experience other identities. Sexuality and romantic orientation do not always align, especially regarding asexual or aromantic identities, and both make labels more complex. For example, an asexual homoromantic man and an aromantic homosexual man may both publicly identify as “gay” for the sake of avoiding intrusive questions about their personal lives or intimate relationships and activities. The posts that sparked this line of inquiry also made crude references to sexual acts the posters assume have been performed by these players or these players have had performed on them. Confirmation of a romantic relationship is not confirmation of a sexual relationship, and that confirmation is not something the public is owed. At no point did I, or anyone involved in conducting this study or writing this paper, reach out to any of the named players to request clarification on what sexual acts they participate in, nor do I intend to, and nor should you. This is not an exploration of the impact of sexual activity of any kind on hockey performance.

LIMITATIONS

This paper exclusively examined cisgender males due to the assumption that all current players in the professional men’s hockey league are cisgender. This is a reflection of current NHL policy, not an endorsement of trans or nonbinary exclusion. While this paper focuses solely on gay and bisexual men, it does not intend to exclude any other queer sexual orientations. It does reflect the public labels of current out players. The applicability of the findings of this study is also lessened due to the extraordinarily small sample size (n=4). This was due to a lack of availability of additional subjects. Should the NHL someday be more diverse, further study will be beneficial in reinforcing or correcting these findings. 

CONCLUSION

Further study into this subject is warranted, but current findings suggest a positive relationship between homosexuality and/or bisexuality in men and hockey skill. This seems to imply that online speculation that participation in same-sex male sexual activity increases hockey skill, but no sexual activity here has been confirmed. That said: further testing of the accuracy the formula determined in this paper is entirely possible. A difference-in-difference methodology could be easily used to study the effects on hockey performance of an increase in same-sex sexual activity, as any discontinuity occurring at the cutoff between zero same-sex sexual activity and any positive quantity of same-sex sexual activity would likely be statistically significant and reasonably assumed to be a direct effect. However, this hypothetical testing would face ethical concerns regarding pressuring straight men to participate in actions that do not align with their sexual preferences for the sake of scientific advancement or professional success, ethical concerns regarding pressuring closeted queer hockey players to out themselves to maintain data integrity, and logistical concerns involved in guaranteeing that any players who enthusiastically consent to participation in further study are actually heterosexual, regardless of what they may indicate or believe. As such, I will be pursuing no further study at this time. 

APPENDIX A:

Someone on Reddit calculated the “most average” player in the NHL during the ‘24/’25 season and determined it was Drew O’Connor. Their work is here: https://www.reddit.com/r/hockey/comments/1mzr21h/slightly_unserious_stats_episode_5_who_is_the/ 

All “heterosexual League averages” are O’Connor’s statistics from this year, found here: https://www.nhl.com/canucks/player/drew-o-connor-8482055 

They are compared to the top three players and the 30th best player of the ‘25/’26 season, found here: https://www.nhl.com/stats/skaters?reportType=season&seasonFrom=20252026&seasonTo=20252026&gameType=2&sort=points,goals,assists&page=0&pageSize=50

I am not RPFing or speculating on the sexualities of any of these real players, but I needed some sort of reference point to assign real numbers to the described skill levels of canon characters. (That’s what got me into this mess; this whole thing was going to have no real research but I needed a number for the men outnumbering women stat in the first paragraph, then I was like “why am I doing research to get a realistic fake number when I can just do research to get a real number,” then if I already had one real number why should everything else be fake, and here we are.) Shane, Ilya, and Scott are commonly accepted to be some of the best, so I gave Shane MacKinnon’s numbers (most goals, fewest assists because he’s carrying the Metros on his back, highest plus/minus because his team literally doesn't do shit without him so he’s always on the ice), Ilya got Kucherov’s (more assists than Shane because his team can actually play, second-highest plus/minus because he does spend a lot of time on the ice, highest penalty minutes), and Scott got McDavid’s (most games played because the other two keep getting knocked out for injuries, most assists because he trusts his team, lowest plus/minus because he’s older and spends less time on the ice and his team has a much more balanced division of play, not a “get the puck to our specialest boy and he’ll take it from there” star-dependent style.)

Troy got #30 in the rankings, Hutson, because he’s a good player but he’s no Hollanov. Fewer goals by a lot, fewer assists but not by as far, plus/minus lower then Hollanov’s but higher than Scott’s. Still significantly better than most of the league but not going to make history. (Sorry, Troy.)

If you can’t tell, I put entirely too much time, effort, and legitimate research into what was originally a joke. This is crack taken FAR too seriously. I have a graduate-level education in data sciences and have written hundreds of research papers and this is what I am choosing to do with those skills and also my summer “break” from academia. Recreational research papers. 

APPENDIX B

This is bullshit. There is no formula to calculate the odds of an individual player making a shot. The referenced formula is real ( https://evolving-hockey.com/blog/a-new-expected-goals-model-for-predicting-goals-in-the-nhl/ ) but can predict goals based on where they’re shot from, type of shot, etc, not whether a specific player will score. The theory to make this look somewhat believable is that rookies are less likely to get points, players over average retirement age (28) are less likely to get points, faster players are more likely to get points, position and line impact scoring probability (this is bad practice; most of these factors impact line placement and are therefore being counted sort of twice). I was going to control for team but there are like 32 of them and I was not doing that for the whole beta value section. Also, the “top speed” categories 1–5 are based on exactly Nothing. I just wasn’t going to fuss with figuring out a real, consistent, and logically sound way to do that. This works if you squint and maybe tilt your head a little and that’s all that matters.

Don’t try to do real math here and apply this to a real player. It’s mostly fake. (Unless I’ve somehow pulled something groundbreaking out of my ass, in which case, let me know so we can both make bank betting.) However: the average points per game for the four players I picked to sacrifice their stats to the gays = 1.4825, which weighted by their being .38% of the league, means they make up .0056335 of the average player’s points per game (0.35366, using this one deeply average man as an approximation for a total average), so the math is that without the gays included in their math, the average player scores 0.34803 points per game and “the gay effect” is boosting League-wide stats because the gays are just that powerful. Everything else is baseless, though. Because this is #fiction. This is how regression equations work, though. This is a regression I could put in my software and make a graph out of. You know, if the data were real. 

APPENDIX C

But the thing is, since I have already written numbers, it would bug me to just make something up here. SO!

Ê[P/PG] = 0.34803 - 0.08732(Rookie) -  0.06005(Over28) + 0.09401(Speed1) + 0.18492(Speed2) + 0.27687(Speed3) + 0.40992(Speed4) + 0.51184(Speed5) + 0.10592(PositionL) + 0.10779(PositionR) + 0.22092(PositionC) - 0.00425(PositionD) - 0.4067(PositionG) + 0.35772(Line1) + 0.07264(Line2) - 0.02921(Line3) - 0.04581(Line4) +  β₆[Queer]  + ϵ_i

Plugging in Ilya/Kucherov because his P/PG was highest at 1.71 (and also Scott’s canonical age is a little ambiguous, books vs show). Putting him in speed category 5 since he and Shane are arguing over who is fastest of the two of them but are definitely some of the fastest in the league. He’s a first line Center. If you’ve never done regressions, these are all binary variables! So the positions are (value)(category, yes/no). 0.00592(PositionL) means +0.00592 odds of scoring if you play left wing, so it’s 0.00592(0) if not or 0.00592(1) if yes. The whole equation removes categories the person isn’t (different lines, different positions) by multiplying them by zero and includes the one they are by multiplying by one. β₆[Queer]  is (score value of being queer) x (Queer: yes or no?).  ϵ_i is just the standard error. 

Ê[P/PG] = 0.34803 - 0.08732(0) -  0.06005(0) + 0.09401(0) + 0.18492(0) + 0.27687(0) + 0.40992(0) + 0.51184(1) + 0.10592(0) + 0.10779(0) + 0.22092(1) - 0.00425(0) - 0.4067(0) + 0.35772(1) + 0.07264(0) - 0.12921(0) - 0.04581(0) +  β₆[1]  + ϵ_i 

So:

1.71 = 0.34803 + 0.51184 + 0.22092 + 0.35772 +  β₆[Queer]  + ϵ_i

Gives us: 

1.71 = 1.43851 + Queer Effect

Therefore: 

Queer Effect = 0.27149

And yes, I did initially end up with a number way off from where I wanted it to be and have to go through this whole thing and change a bunch of numbers by one decimal place. If you see a math error, no you don’t. 

REFERENCES

1. Statista. “Share of NHL Fans by Gender US 2024.” May 2024. https://www.statista.com/statistics/1478103/nhl-game-viewers-gender-us/?srsltid=AfmBOortlFlsx-s7P7ElEb_rXjoPHsJ5rCEMDoCqw_HjA9Abz5C4txnv. 
2. Curt Ebejer. “How Many Players Are In The NHL (2023 Update) | Cleats Hub.” Cleats Hub, June 16, 2023. https://cleatshub.com/learn/how-many-players-are-there-in-the-nhl. 
3. Sean McCann. “AHL & ECHL Prove to Be Vital Development Ground for Players Aspiring to Make the NHL.” PHPA.Com. Accessed July 23, 2026. https://phpa.com/news/2081. 
4. thePWHL.Com. “Beginner’s Guide to the PWHL.” Accessed July 23, 2026. https://www.thepwhl.com/en/beginners-guide. 
5. Associated Press. “Racism Lingers for NHL Players 60 Years after O’Ree Landmark.” Wayback Machine, Internet Archive, November 9, 2018. https://web.archive.org/web/20190721211142/https://www.usatoday.com/story/sports/nhl/2018/11/09/racism-lingers-for-nhl-players-60-years-after-oree-landmark/38451681/. 

6, 8. Andrew Gaetano, Krishin Shivdasani, Andrew Chen, Nickolas Garbis, and Dane Salazar. “Racial Concordance Between NBA and MLB Players and Their Team Physicians.” Orthopaedic Journal of Sports Medicine 13, no. 2 (2025): 23259671241310472. https://doi.org/10.1177/23259671241310472. 

7. Jaimes, Richard. “NFL Statistics And Demographics 2026.” Quantumrun Foresight, October 22, 2025. https://www.quantumrun.com/consulting/nfl-demographics/. 
8. Rachel Reid. “Game Changer.” Accessed July 23, 2026. https://www.rachelreidwrites.com/gamechanger. 
9. Rachel Reid. “The Long Game.” Accessed July 23, 2026. https://www.rachelreidwrites.com/the-long-game. 
10. Game Changers Series - Rachel Reid Wiki. “Timeline of Events.” July 22, 2026. https://game-changers-series.fandom.com/wiki/Timeline_of_Events. 
11. Rachel Reid. “Role Model.” Accessed July 23, 2026. https://www.rachelreidwrites.com/role-model. 
12. McKinsey & Company. “Women’s Sports Market Growth: A $2.5 Billion Opportunity.” Accessed July 23, 2026. https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/closing-the-monetization-gap-in-womens-sports-a-2-point-5-billion-dollar-opportunity. 
13. Alyce Brown. “NFL Turns to Outlets Betches, The Gist to Broaden Its Appeal to Women.” NBC Bay Area, November 4, 2025. https://www.nbcbayarea.com/news/sports/nfl/nfl-aiming-broaden-appeal-women/3975128/. 
14. Devin Heroux. “Linemates to Lifemates: A Hockey Love Story.” CBC Sports, November 21, 2024. https://www.cbc.ca/sports/longform/features/Hockey-love-story-Poulin-Stacey. 
15. Mathew Rodriguez. “‘Heated Rivalry’ Has Nothing on the Real-Life Rivals-to-Wives Tale of These Olympians.” Them, December 11, 2025. https://www.them.us/story/julie-chu-caroline-ouellette-heated-rivalry-hockey-wives-couple-gay. 
16. Steven Sandor. “Leveling the Ice.” Beacon Press. Accessed July 23, 2026. https://www.beacon.org/Leveling-the-Ice-P2219.aspx. 
17. Brian Tuohy. “NHL Referees Admitted the League Was Fixed.” The Fix Is In. Accessed July 24, 2026. https://www.thefixisin.net/refs-admit-nhl-is-fixed. 
18. NHL Public Relations. “NHL Investigation Regarding Evander Kane Completed.” NHL.Com, September 22, 2021. https://www.nhl.com/news/sharks-kane-gambling-investigation-completed-326267762. 
19. NBC Sports. “Ottawa’s Shane Pinto Suspended 41 Games, Becomes the 1st Modern NHL Player Banned for Gambling.” October 26, 2023. https://www.nbcsports.com/nhl/news/ottawas-shane-pinto-suspended-41-games-becomes-the-1st-modern-nhl-player-banned-for-gambling. 
20. Game Changers Series - Rachel Reid Wiki. “Shane Hollander.” July 24, 2026. https://game-changers-series.fandom.com/wiki/Shane_Hollander. 
21. AngryCentrist on Reddit. “I Analyzed James Harden’s Performance in Every NBA City to See If There Is a Correlation between His Box Score and the City’s Average Strip Club Rating. : R/Nba.” Reddit.Com. Accessed July 24, 2026. https://www.reddit.com/r/nba/comments/dt0ucg/i_analyzed_james_hardens_performance_in_every_nba/. 
22. Game Changers Series - Rachel Reid Wiki. “Ilya Rozanov.” July 24, 2026. https://game-changers-series.fandom.com/wiki/Ilya_Rozanov. 
23. Game Changers Series - Rachel Reid Wiki. “Major League Hockey.” July 24, 2026. https://game-changers-series.fandom.com/wiki/Major_League_Hockey. 
24. Jerel P. Calzo, Andrea L. Roberts, Heather L. Corliss, Emily A. Blood, Emily Kroshus, and S. Bryn Austin. “Physical Activity Disparities in Heterosexual and Sexual Minority Youth Ages 12-22 Years Old: Roles of Childhood Gender Nonconformity and Athletic Self-Esteem.” Annals of Behavioral Medicine : A Publication of the Society of Behavioral Medicine 47, no. 1 (2014): 17–27. https://doi.org/10.1007/s12160-013-9570-y. 

“26. Gay Boys Play Team Sports at Half the Rate of Their Peers.” Out On The Fields, December 3, 2020. https://outonthefields.com/male-sport-participation/. 

27. Maizie Schwets. “The Hidden Presence of Rape Culture in Canadian Hockey.” Journal of Undergraduate Studies at Trent (JUST) 9, no. 1 (2025). https://ojs.trentu.ca/index.php/just/article/view/1147. 
28. Raul Daffunchio Picazo. “Inside The Games • Canadian Hockey Players on Trial.” July 23, 2026. https://www.insidethegames.biz/articles/former-canadian-ice-hockey-trial-rape. 
29. Jarryd Willis, Tugral Bek Awrang Zeb, Joie Haydel, et al. The WNBA and Athletic Performance Based on Sexual Orientation. April 15, 2020. https://osf.io/ykqzc_v1. 
30. NHL.Com. “NHL Stats.” Accessed July 25, 2026. https://www.nhl.com/stats. 
31. aschwan41 on Reddit. “Slightly Un-Serious Stats Episode 5: Who Is the *most Average* NHL Player? : R/Hockey.” Reddit. Accessed July 25, 2026. https://www.reddit.com/r/hockey/comments/1mzr21h/slightly_unserious_stats_episode_5_who_is_the/. 
32. Josh Younggren and Luke Younggren. “A New Expected Goals Model for Predicting Goals in the NHL.” Evolving-Hockey.Com, May 1, 2021. https://evolving-hockey.com/blog/a-new-expected-goals-model-for-predicting-goals-in-the-nhl/. 
33. Columbus Blue Jackets. “Hockey Is for Everyone.” Accessed July 25, 2026. https://www.nhl.com/bluejackets/community/hockey-is-for-everyone. 
34. Bill Chappell. “The NHL Bans Pride Tape, Setting off a Backlash from Players and Fans.” Sports. NPR, October 12, 2023. https://www.npr.org/2023/10/12/1205476006/nhl-bans-pride-tape-backlash. 
35. Jack Walker. “NHL Player Hurls Antigay Slur during Olympics Hockey Match.” Yahoo Sports, February 23, 2026. https://sports.yahoo.com/articles/nhl-player-hurls-antigay-slur-194000603.html. 
36. Ruth Graham. “Riley Gaines Finished 5th. Now She Believes Victory Is in Her Grasp.” U.S. The New York Times, August 27, 2025. https://www.nytimes.com/2025/08/27/us/riley-gaines-trans-athletes.html. 
37. Wood, Ruth I., and Steven J. Stanton. “TESTOSTERONE AND SPORT: CURRENT PERSPECTIVES.” Hormones and Behavior 61, no. 1 (2012): 147–55. https://doi.org/10.1016/j.yhbeh.2011.09.010. 
38. Issitt, Micah L. “Russian Sports Doping Scandal.” EBSCO, 2023. https://www.ebsco.com. 
39. Mayo Clinic. “Testosterone (Oral Route) - Side Effects & Dosage.” Accessed July 25, 2026. https://www.mayoclinic.org/drugs-supplements/testosterone-oral-route/description/drg-20461351. 
40. Dr. Bridget, Clinical Pharmacist. “Expert Answers on Testosterone Troche, Hormone Troches, and More.” JustAnswer Pharmacy, January 23, 2021. https://www.justanswer.com/pharmacy/fyask-odd-question-testosterone-troche.html. 
41. Vitku, Jana, Lucie Kolatorova, and Richard Hampl. “Occurrence and Reproductive Roles of Hormones in Seminal Plasma.” Basic and Clinical Andrology 27 (September 2017): 19. https://doi.org/10.1186/s12610-017-0062-y.