Data Analysis #3: The arrival of Statshark answers some old questions

[Beeschurger]

@Bruce_R1

The effect of individual events appears to be variable and within a fairly small range. The March ground event (Fiat 6614) did not pull people from air, over and above other factors. Similarly the Meteor event in April did not pull people from ground.

Isnt it possible that this is affected by how “good” the event reward is?

For example I main ground, getting ground vehicle even if its not good is no brainer for me because it does not require me to change anything about how I normally play Warthunder.

However getting air vehicle requires me to switch over to ARB and play it regularly, a significant change to how I normally play. And, in order to do so, the reward needs to be appealing.

Since the start of the year, that wasnt the case with air events.

[Kweedko]

@Bruce_R1

AB numbers would also be up with ground RB, but they’re not

Are you blind? Or just ignoring that GAB and GRB totally correlating month to month.

[Bruce_R1]

Oh absolutely that’s going to be a factor. The argument put forward was that the growth in ground RB was entirely tied to people switching modes en masse due to single vehicle events, not that they switch more for the good events.

So far we’ve shown available data shows the event swing on average is only 2-5% of games played, not enough to account for the ground RB gains. Probably the “good vehicles” get a higher swing within that range, sure. Two things: the Fiat 6614 event (no comment on the actual vehicle quality) was actually quite popular if you count vehicles placed on the market and market cost: more than the T86, and as much as the T34. So if people only switched between modes for events they would have switched just as much for the Fiat, but as the stats show GRB actually was played much less that month than in May (T-34) or June (T86). And May, the month where ground RB broke all records, 15 million more sessions or 500,000 more actual games (500 more every single HOUR) than the previous month was… just the T-34-85 event, nothing else? Did anyone feel that was the hugest, most amazing War Thunder event they’ve ever seen, bigger than anything else this year? Did we all miss the hype?

The simple fact is more games are being played in ground RB in 2025 than presumably ever before, both in absolute and relative terms. I’ve suggested a couple reasons why over the last six months it has apparently continued to cement itself in as the dominant game mode. There’s a worthwhile debate to be had as to those reasons WHY that is; but to pretend it’s not against all the Statshark data is basically flat-earthism levels of denial.

We should get a good new data point in another month by the way, assuming the remaining August event vehicle is an air vehicle and a good one. We didn’t see any big swings back to air for the previous events this year (granting NF-5A did move the needle a bit), but hey, maybe the next one will do it. Or maybe the stats are right and 95% of War Thunder players’ mode choice night-to-night has nothing to do with chasing events. We’ll see. My take would be, as we’ve seen in the past, it can be very easy to be on a forum, where community member engagement and awareness levels are very high compared to the average player, and mistake our informed, event-chasing experience as a subgroup for the regular player’s experience. Happens in every game, all the time.

[Beeschurger]

@Bruce_R1

The argument put forward was that the growth in ground RB was entirely tied to people switching modes en masse due to single vehicle events, not that they switch more for the good events.

Ah, I see what Ive got wrong.

[TheDudesRug]

My theory that I’m basing on a hunch, so take with a grain of salt, is that the continued instability and dramas occurring with World of Tanks has led to an exodus of players to War Thunder. Now a player from that game could settle with Thunder’s arcade mode… but why? why would you come to play some dodgy arcade mode, when you can embrace the bigger point of difference between these two games: RB mode.

[Bruce_R1]

So a couple interesting graphs (I think) showing what happens when a mode undergoes major decompression, as Naval did in June:

Here is the “match fairness” graph for Naval for April, before decompression. Note the significant likelihood that 5.7 you would be top tier (which is what high means… 3 on the vertical scale is 100% top tier probability. This was because the matchmaking for the 7.0 battleship games were sweeping up all the 6.0-6.7 players. I showed earlier this “bow wave” 4 steps below the top tier BR was consistent across all BRs and modes where there hasn’t been recent decompression:

Spoiler

Decompression in July saw the existing Naval base split over the addition of 4 new BRs. For context, the player population for both naval modes is split relatively similarly across BRs currently, very different from the other modes in this regard, where AB tends to have more low tier and RB tends to have more high tier players (note these are relative, proportional numbers, not absolute), very much belled with the bulk of players still playing the mid-tiers.

Spoiler

This belling combined with the instability of the first month of decompression, has let to a BR spread without a bow wave, or, really, any “bad BRs” (other than 2.3… friends don’t let friends play 2.3):

Spoiler

So you might ask how that’s even possible, how can every mid to high BR have a better than average chance of an uptier position? We have to remember that this line always is going to start as a basic slope up (the top BR will always be top tier, and 1.0 will always be bottom tier), and it changes from that straight upwards line to a wavy upwards line because players start to cluster en masse to avoid bottom-tier situations, which hasn’t had time to solidify in naval yet post-decompression. But also that bell in the middle (see the previous graph) means in general there’s always more players just below whatever BR you’re at, so you will tend to always have a higher-than-average tier so long as you’re on that right-side “downslope”.

Basically what you’re looking at is the accordion stretched out due to decompression. It will be interesting to see in future months how long it takes for that standard War Thunder bow-wave up-tiering pattern to reassert itself, with 7.7 in particular cleaning out, either due to player population changes or possibly additional BR changes.

[TIGER_TANK_1]

It is no doubt that high tier has similar amount of players considering the lack of rangefinders on old battleships. Anyway 6.0 lines up with my experience. Lots of cruisers to club with the WW1 dreadnoughts.

[Josephs_Piano]

@Kweedko

Puts a kybosh on the outraged cries of “naval is dead” - the drop in April and March has been made up in spades by July.

[Josephs_Piano]

You don’t get to tell people that NRB is somehow objectively better than NAB - for yourself by all means, but otherwise it is nonsense judgemental bulldust.

[_Poul]

@Josephs_Piano

Puts a kybosh on the outraged cries of “naval is dead” - the drop in April and March has been made up in spades by July.

I don’t think most people who say “Naval is dead” actually mean that it has no players. It’s more about the fact that nowadays you’re basically facing AI bots in every battle.

Some players thought that the major update, which mostly focused on Naval and added iconic battleships to the game, would somehow make the mode popular, but it didn’t.

It looks especially bad when you compare Naval to Air or Ground modes. I think this graph is more interesting, as it shows the difference in popularity more clearly:

@_Poul

Arcade and Realistic stats combined:

In July, Naval made up only about 2.66% of total Ground sessions and just 6.62% of total Air sessions. So there’s no doubt that Naval’s popularity is marginal compared to other game modes.

The popularity of a game mode directly translates into revenue - the more popular the mode, the more money it generates.

The problem is, developing the Naval mode doesn’t cost only 2.66% compared to developing the Ground mode. So there’s a valid question: how long will Naval continue to be developed?
We saw a similar situation with Helicopter PvE and nowadays it’s completely abandoned by the devs. They don’t even care about major bugs, like broken objectives, that make the mode frustrating to play.

I don’t think Naval has a bright future in War Thunder. If you really think about it, how many updates could still significantly boost its player base? Submarines, maybe aircraft carriers, or modern missile ships and that’s about it. But how much effort, time, and money would it take to implement those into the game? At this point, it’s more a question of whether it will ever happen at all.

From my perspective, the Leviathans update was a disaster. If the Naval player base had grown five to ten times after that update, it would have been a success, but it never happened. From the devs perspective, it doesn’t really matter whether Naval’s popularity compared to Ground is 1.5% or 2.6%. Those are still marginal numbers, effectively meaningless. I think that’s what people mean when they say “Naval is dead”.

[Kweedko]

Do you understand that naval battles had way longer sessions than ground and even longer than air, especially on 5.3+. And with much distant respawns and nerf to accuracy on all guns 100+mm, which came with update, they become even longer even on low tiers. Where avarage player get 5 sessions on airbattles, naval player get 1. And with single respawn tendency in ground, their battles shortened too.

So it’s spherical cow in a vacuum you comparing there.

And even with all downsides of WT naval and it’s problems and bugs, it is still more popular than WOWs in terms of numbers of players.

[_Poul]

@Kweedko

Do you understand that naval battles had way longer sessions than ground and even longer than air, especially on 5.3+.

Yes, I do understand that. At the same time, at lower BRs, Naval Arcade battles can be very short. I even have some 5 minutes battles with coastal vessels on my channel. Naval isn’t only about slow-paced battleships at 6.3 BR and above.

@Kweedko

Where avarage player get 5 sessions on airbattles, naval player get 4-5.

So even you admit here that the difference at average isn’t huge. Even if we count 4 sessions for Naval players at average (instead of your estimated 4.5), that’s just 20% difference. It doesn’t change that much.

EDIT: I see you edited your message and changed 4-5 to 1. Now it doesn’t make any sense at all - it’s clearly trolling. Even if every Naval battle lasted the maximum possible 25 minutes (which is obviously not true), that would mean other game modes have battles lasting only 5 minutes at average. At this point, you might as well say that other game modes have average battles lasting 5 seconds, and then do the math to “prove” that Naval is more popular than Air and Ground.

[Kweedko]

@_Poul

So even you admit here that the difference at average isn’t huge. Even if we count 4 sessions for Naval players at average (instead of your estimated 4.5), that’s just 20% difference. It doesn’t change that much.

It was a typo, i corrected it, it’s one instead of 4-5.

@_Poul

Yes, I do understand that. At the same time, at lower BRs, Naval Arcade battles can be very short. I even have some 5 minutes battles with coastal vessels on my channel. Naval isn’t only about slow-paced battleships at 6.3 BR and above.

Your all “costal” games 110k vs 450k of 3.3-5.0 in numbers of players.

and 300k “sessions” versus 1.2m “sessions”

[Kweedko]

@_Poul

and then do the math to “prove” that Naval is more popular than Air and Ground.

Never stated that. Only that your method is wrong, and your conclusions is far from accurate.

[MotorolaCRO]

@_Poul

The problem is, developing the Naval mode doesn’t cost only 2.66% compared to developing the Ground mode. So there’s a valid question: how long will Naval continue to be developed?

People don’t realize their dead mode uses as much development time as the game’s bread and butter.

@_Poul

EDIT: I see you edited your message and changed 4-5 to 1. Now it doesn’t make any sense at all - it’s clearly trolling. Even if every Naval battle lasted the maximum possible 25 minutes (which is obviously not true), that would mean other game modes have battles lasting only 5 minutes at average.

Naval shills clearly had become so desperate that they resort to using such bold claims.

[Kweedko]

@MotorolaCRO

People don’t realize their dead mode uses as much development time as the game’s bread and butter.

And then you prove it. That modes that were abandoned for two years used same dev time as “popular” modes. And that Naval is a money recipient from “popular” modes, not donor to them.

[MotorolaCRO]

@Kweedko

That modes that were abandoned for two years used same dev time as “popular” modes.

Ships don’t just appear out of thin air, same with mechanics as well. Naval got way too much attention in the Leviathans update relative to it’s popularity.

@Kweedko

And that Naval is a money recipient from “popular” modes, not donor to them.

Naval is definitely living off of the success of other modes in this game.

[Josephs_Piano]

@_Poul

In July, Naval made up only about 2.66% of total Ground sessions and just 6.62% of total Air sessions. So there’s no doubt that Naval’s popularity is marginal compared to other game modes.

Absolutely - as I’ve said before. I’ve been wargaming over 50 years, and naval has never been as popular as land-based games in any format - figure, board or electronic.

and I doubt it ever will be - comparing the numbers is a pointless exercise - naval has its aficionados, and there will always be fewer of them than for tanks and planes.

That doesn’t stop it being a valid form of gaming, or mean the people who do it should give it up or some such nonsense.

If you want to go down that route what are you playing WT at all since it is so much less popular than various other games?

[_Poul]

Updated graph with August 2025 data:

Combined Arcade and Realistic data:

[SpectreXXI]

Is there any way statshark can differentiate player performance based on server? That would be great if they can do so, I’m interested is there any disparity between the server and if better, timezone.