Here's some cold hard math for you, as of March 2020, using the most commonly accepted criteria for an outlier (i.e., an outlier must exceed the mean plus or minus twice the standard deviation): K/D by Subclass: Arcstrider: 1.3658 Nightstalker: 1.3152 Stormcaller: 1.313 Striker: 1.2656 Gunslinger: 1.2787 Voidwalker: 1.2472 Sentinel: 1.2444 Dawnblade: 1.2419 Sunbreaker: 1.2032 Average: 1.275 Standard Devation: 0.049202 Criteria for Upper Outlier: >1.373404 Criteria for Lower Outlier: <1.176596 Outliers: NONE Win Rate by Subclass: Arcstrider: 55.2341 Striker: 53.3734 Nightstalker: 53.105 Stormcaller: 52.8728 Sentinel: 51.1421 Voidwalker: 50.8981 Dawnblade: 50.7407 Sunbreaker: 49.3275 Gunslinger (What? A Hunter class at the bottom?): 48.4137 Average: 51.6786 Standard Deviation: 2.148795 Criteria for Upper Outlier: >55.97619 Criteria for Lower Outlier: <47.38101 Outliers: NONE. Note that Gunslinger, the MOST POPULAR SUBCLASS, had the LOWEST win rate and is consistently ranked among the bottom 3 subclasses in K/D (The above data is the most recent data point. Previous data points show Gunslinger placement alternating between the bottom 3.). So the trends show Arcstrider at the top, but second and third place fluctuating between Warlock and Titan classes. The gaps between the classes on average is small, with no statistical outlier. This is the case for all data points. (I've just shown the most recent one here.) So what does this mean? All the classes are operating close to the average. It means that the trends formed by each subclass are a result of inherent strengths and weaknesses within the subclass, rather than the class itself. If the Hunter class was an issue, Gunslinger wouldn't have been in the bottom ranked classes between Decemeber and Feburary in terms of K/D, and it certainly wouldn't be at the very bottom in terms of win rate. Heck. Right now Gunslinger is right in the middle in terms of K/D. There are also zero statistical outliers. So, if like everyone else on here, you decide to use statistics to determine class adjustments, Hunters are fine. I could do an average across all data points since Decemeber, but considering when there were peaks in Hunter performance there were peaks in the performance of other subclasses, the result would be the same: no outliers. Gunslinger near the bottom or in the middle. And as a Hunter main, I know very well what happens next on these forums. So have fun. At least be nice to eachother. Edit: [spoiler]Guardian.gg has the data. I believe my data point is from today on console. Also, this data is more focused on discrediting the "The stats say Hunter is OP" arguement. Say it's OP all you want. The stats say it fits within the mean. [/spoiler] Edit: [spoiler]I'm well aware of the effects of sample size and inter-subclass variances. (For those of you that don't, a larger sample size means the "measured" mean is more representative of the "real" mean. As a quick example, that means that the placement of Gunslinger is fairly concrete, but Arcstrider could very well be over-represented, as it is played much less. Taking into account sample sizes will likely bring down Arcstrider, while cementing the places of Stormcaller and Nightstalker.) However, I do not have the raw data for these points. I don't think it's publically available. In a perfect world, I could do a proper ANOVA method. But we don't have that luxury. This post is more to discredit the "Stats say Hunters are OP" argument, and it suits its purpose. [/spoiler] Edit: Removed comp stats because they're glitched on guardian.gg. Edit for Clarification: [spoiler]As far as the stats, they're only good for the purpose of discrediting the "The stats say Hunter are OP argument." If we wanted the actually statistically rank the effectiveness of each subclass in comparison to one another, we'd need to do an ANOVA table, as sample size plays a large role in inter-subclass variances. So what that means is that while the data of classes with high populations like Gunslinger and Voidwalker are probably fairly representative of the class, the other classes may move up or down once inter-subclass variances are taken into account with an ANOVA method. Since we can't take the inter-subclass variances into account, the only conclusion we can draw is "The data presented by guardian.gg does not indicate a significant outlier." Unfortunately, we can't statistically rank the effectiveness of each class apples-to-apples. Sure wish we could though.[/spoiler]