Twitch game stats tell a story...

In what's quickly becoming a theme of this blog, I'm going to procrastinate still further on the D3 design discussion (I swear on my mother's eyes that part 2 is coming) to instead look at Path of Exile's Twitch viewership over the past year:

Viewer stats and trend lines for PoE from Harbinger to June 25th.

Looking at this graph, it's pretty easy to see that the Abyss and Bestiary content releases under-performed. People just weren't interested; the "Path of Exile/Pokemon" experience of Bestiary grabbed more attention than Abyss, at least initially, but it didn't hold them; and both releases' overall numbers were poor compared with the big spikes on either side: Harbinger and Incursion.

Harbinger and Incursion, at launch, had very similar numbers: 15,029 average viewers (peaking at 85,576) for Harbinger, and 16,607 average viewers (peaking at 84,219) for Incursion. This tracks very nicely with PoE's stated peak concurrency numbers, too, which had both leagues peaking at over 100K players on their opening days.

Two weeks into each league, though, the differences are already apparent. Harbinger's Aug. 19th number was 11,786 average viewers, with a peak of 22,127, a falloff of -21.6% avg. (-74.1% from peak). Incursion's June 16th number, by comparison, is 9,505 average viewers, with a peak of 13,684, a falloff of -42.8% avg. (-83.8% from peak).

This is a stark difference. Peak viewership probably means less than average viewership, but either way, Incursion is doing significantly worse in terms of holding viewers' interest than Harbinger did. And it probably shouldn't surprise anyone to learn that PoE's forums have turned into a significantly more divided and partisan place than they were just two weeks ago. One of the most active threads, "They actually managed to do worse than bestiary," is now up to 31 pages of back-and-forth between people who hate the game's current meta (and the challenge league that seems built entirely around it), and people who love both of those things.

As a survivor of Diablo III's official forums, this is bringing back all kinds of ugly memories for me.

PoE, however, is not fundamentally flawed the way D3 was (and still is); PoE's fundamental progression/RPG and combat mechanics are all basically sound, and game's core experience can still be enjoyable. But fundamentals are only your starting point; your core experience must be solid, but if you bury that core under too much other crap, those solid fundamentals can become harder and harder to see. And a metagame which makes playing an Area of Effect skill-based build feel mandatory, and playing anything else feel wrong, is clearly hurting PoE's ability to hold onto viewers' (and, most likely, players') interest.

PoE has been free to play since 2013, and Grinding Gear Games have been adding to it ever since. And, as a long-time supporter of the game, I've loved most of those changes; some of PoE's features, like the massive passive skill system and Fall of Oriath's 10-act structure, are already appearing in other ARPGs, which means that GGG hasn't just made a good ARPG, they've redefined the genre, pushing ARPGs as a whole forwards, for the first time since Diablo II. 

(D3, remember, came out before PoE; but to date, no other ARPG has adopted any of its design changes.)

So, yes, PoE has been a great game, for most of its five-year live-service life. But its current balance now works against its original core design concept: that you should be able to build any type of character you want, to suit whatever playing style you prefer, and succeed (or fail) on an equal footing with other players, regardless of their build types. The specifics would matter; they'd matter a whole helluva lot, in fact. But not the broad strokes; summoners should feel very different from archers, who should feel different from direct-damage spell-users, who should feel very different again from melee characters... and, in PoE's current meta, they just don't.

GGG are still a very talented group, and their game's fundamentals are still sound, so it's possible for them to re-balance their game, and thus recapture the qualities that made it great in the first place. But the longer they wait, the harder that task will become, and the more of their current players they'll risk losing in the process. For GGG, risking players who are still playing (and paying) to chase players who have moved on to other games may simply not worth doing. Failing to correct, though, will put PoE into an irrevocable decline, just after they achieved some of their greatest success.

To end is the fate of all empires. But for such a great game to end with such a whimper? That would be real shame.

Updated June 26th

Now over three weeks into Incursion, the pattern has continued. 8,616 avg./13,316 peak for Harbinger as of Aug. 28th (-42.7% avg./-84.4% peak) and 6,910 avg./11,971 peak for Incursion (-58.4% avg./-85.8% peak). PoE's Steam player numbers on show a similar trend:

From SteamCharts.com

It will be another month before we can compare average player counts between the two expansions, since SteamCharts only track averages monthly, not weekly, but Harbinger's peak player counts were 98,110 for the week of July 31/2017, and 60,597 for the week of Aug.28/2017, a fall-off of 38.2%, while Incursion's peak player counts were 95,824 for the week of May 28/2018 and 45,566 for the week of June 25/2018, a fall-off of 52.4%. PoE isn't only playable on Steam, of course, but GGG are claiming 140,870 peak players for Incursion, so SteamCharts' numbers track, and they put Incursion's peak lower than Harbinger's peak, with a significantly sharper fall-off.

GGG have more accurate, detailed, and comprehensive stats than this, of course, but I'll be very surprised if they tell anything other than this same story. Hopefully Chris Wilson's team is paying attention to these early warning signs, and not just seeing what they want to see.