Snapshot stats of the EU server

[ezymodo]

Well, it's been awhile since my last post. 

Lately I become aware of some stats collected on the warshipstats website (http://warshipstats.com/analytics/) , and decide to makesome plots of the EU data for other players.

While the data tracked only a fraction of the player base (4k out of around 30 K in EU server, most likely above average), I believe this will give some picture about the overall condition of the game.

 

1. Battles per tier

 

tier	battle %	cumulative %
1	1.1	1.1
2	4.2	5.3
3	13.1	18.4
4	21.0	39.4
5	24.4	63.8
6	16.4	80.2
7	9.5	89.7
8	6.6	96.3
9	2.4	98.8
10	1.2	100.0

 

As can be seen, the vast majority of the battles are concentrated in ships of tier 1 - 6, which collectively represents 80% of the battles going in the server. As the ship tier goes up, the number of battles goes down; players doing battle in his/her tier 10 ship in fact only represents 1.2% of the total battles. This number might go up as the game ages and more people obtain tier 10 ships, but it is likely that this number stays constant in the medium term. It also means that player in tier 10 ships are more likely to get unbalanced MM (one way or the other) outside of the peak hours. Players of high tier carriers might need to take special note, since the change in MM due next patch might lead to extremely long waiting time .

 

2.  General conclusions from ships of all tier

 

The figure above are plots of average xp and winrate of all ships vs damage and ship kills. It can be observed that:

1. Xp/ battle can be virtually considered as a linear function of damage done (r² = 0.95)

2. Winrate is influenced by both ship kills and damage done, but the influence of kills seems to be slightly higher. 

 

In short, go for the damage to get more xp during grinding, and go for kills to ensure victory.

This finding is similar to world of tanks; and what about supporting activities, such as capping and killing planes?

 

 

As can be seen, it is not clear that killing planes lead to more xp or victories. The majority of the ships kill too few planes to show any clear trends, except for few outliers that show a slight trend towards increasing xp and wr. On closer observation, these ships are high tier battleships and carriers, which aside from having strong AA also have enormous punch to kill other ships. It is also interesting to observe that the ship that kill the most planes have mediocre xp and win rate compared to the top performing ones; this ship is Lexington, widely considered as the most underpowered CV in the current lineup. In short, killing airplanes does not pay much in terms of xp and winrates.

 

The same can be said with capping and decapping, which do not lead to clear trends in terms of  wr. Observation of the outliers show that in fact the ship with most capping point on average have modest wr compared to its peers. This ship is Shimakaze, and it underscores that decent speed and camo do not lead to enough victory by capping. On the other hand, the best decapper in the game have a slightly higher wr; these ships are american cruisers (st. louis and des moines) with a high dpm guns that can shotgun everything in a cap. This discussion is somewhat academic, since the average decap point of the ships is 2 x higher than the average cap rate of the best capper (8 vs 4 points). In short, capping does not work well due to the ease b which decapping is done, even for the best destroyers.

 

3. Evaluation of ship lines between tier 1 - 10

 

 

The figure above plots the average xp, damage, and winrate of every ship divided per tier and line. Note that at high tiers, very few players actually own the ship (less than 50), and some anomalies might results (due to player skill above average, platooning, etc.). It can be observed that:

1. The plot of average xp and damage looks similar, which is not surprising considering that the correlation of the two is very high.

 

2. At tier 4 - 8, The IJN CV line clearly overperforms in terms of xp, damage, and winrate compared to its peers at the same tier. Thus, the planned nerf on the line seems to be justified.This performance is unmatched until tier 9 - 10, where the USN CV line takes the lead. Either way, the performance of both CVs at tier 9 - 10 surpasses other ship classes.

 

3. At tier 6 -8, The IJN and the USN destroyer lines enters the "valley of death", where little to no improvement can be expected until tier 9 is reached. In any case, the performance of the DDs is far below its contemporaries at all metrics. Compared to its CV peers, the DD get around 60 - 70% less average xp and damage; the winrate is equally abysmal, plunging down starting from tier 6 onwards. In fact , both Shimakaze and Gearing are top contenders for the worst performing ship in terms of winrate at tier 10, with winrate of 48.6% and 46.5% respectively.

 

4. Despite the complaints from some BB skippers, the BB is a very solid line, only second to CVs in terms of xp, damage, and winrates. The large HP pool, large alpha gun, and HP recovery ability still handles the CA endless HE spam, occasional torp from the DD, and the CV TBs pretty well. At tier 10, the Yamato and the Montana performs equally well in terms of xp, damage, and winrate, making both equally worthwhile to own. On the other hand, several lemons can be observed on both lines, such as Colorado at tier 7 and the Izumo at tier 9.

 

5. The cruiser line can best be described mediocre at most metrics. The IJN CA line is in general better in terms of damage and XP, quite possibly due to its lower gun arc and torpedoes. This performance is somewhat let down by Ibuki, but the Zao is again very solid in terms of damage and XP. The USN CA line on the other hand, flounders in terms of xp and damage between tier 6 - 9, only performing slightly better than its DD counterpart. In fact, the tier 9 Baltimore has lower average damage than the Fletcher. This is somewhat improved at tier 10 with Des Moines, however its average damage is 30% lower compared to Zao. On the other hand, the winrate of Des Moines is higher than Zao by 5.5%, possibly by decapping or AA coverage of its teammates.

 

6. The tier 7 can be considered as a "block" tier, with ships statistically underperforming or offering no improvement in terms of xp and damage compared to its predecessors. This include nagato,  pensacola, both DDs, and the infamous Colorado.

 

So the hierarchy of the ship lines for power grinders, statpadders, and team players is in general the same:

 

IJN CV = USN CV  >  IJN BB = USN BB > IJN CA > USN CA > IJN DD = USN DD

1. CV (both lines, depending on tier) 

2. BB (both lines, depending on tier) 

3. IJN CA 

4. USN CA 

5. DD (both tier, almost equally underpowered) 

Whether this is a bug or a feature, only Wargaming knows.  The coming patch most likely rebalance these ships, so we will see .

 

+1 if you think this is useful, and more detailed analysis by others would be welcome.

 

[Scilya]

stats me likey

[Takeda92]

"working as intended^TM"

 

[22cm]

Players of high tier carriers might need to take special note, since the change in MM due next patch might lead to extremely long waiting time .

 

Either way, the performance of both CVs at tier 9 - 10 surpasses other ship classes.

 

OMG, that s sooo bad, the seal clubbers must wait...

 

The tier 9 and 10 CVs need to be nerfed too.

 

And a funny thing, the infamous Furutaka is only the 14th worst ship in the game...

[Nechrom]

Thanks for the statistics rundown!

 

So what we see here is that the winners are the high tier "OP" ship classes, so high tier BBs and CVs, while the high tier DDs and CAs are the losers. Low and mid tier ships are fairly balanced even those that people like to complain about (Kawachi, South Carolina, Myogi etc). Some exceptions exist, like the Izumo, Colorado and Nagato. Furutaka is also leagues behind other same tier ships, but she's getting buffed.

[Thlurp]

thanks a lot for the stats, interesting read +1

[allufewig]

Absolute gem of a post, thank you very much!

 

Also confirms most of my obervations of CVs being the most influential class, BBs at T5+ actually pretty decent (but even T3-4 not as bad as I thought), US CAs extremely mediocre past T6 and DDs on average just extremely poor, again especially at T6 and above.

[LilJumpa]

 

Thx for the analysis (+1)

 

One thing I have to add though:

 

Your written conclusion for killing planes and Capping/Decapping is a little "dangerous" for teamplay in its current wording. The majority of players will just understand/read : "Capping and killing planes is useless, doesn't give XP or wins"

Ofc you are right when you look at the pure numbers and yes, killing planes is not rewarding in terms of XP. But:

 

capping/decapping:

Decap points are rewarded everytime you shoot a ship in a capcircle, regardless of the game mode, and you can decap the same ship multiple times. Cap points are only rewarded if you successfully win a random standard battle by capping or you are in the cap circle when the last ship is killed. AFAIK capping in domination mode doesn't give you anything in terms of cap-points (or only a fraction, I would have to look at game summaries here to confirm)

 

That explains the difference in the numbers. Still: capping is a very important aspect for every gamemode and should stay one of the main goals for DDs/fast ships. It is far more important than in WoT.

Telling everyone that capping is not rewarding and you will not win more is , as i said, dangerous.

 

killing planes:

First of all, this will actually change XP-wise in one of the upcoming patches. Secondly the CV-Balance in 4.1 will change some of the numbes anyway (mirror MM and Loadouts).

If you spread the word, that killing planes is useless, that could actually lead to US-CAs not even thinking of escorting or screening( or skilling AA) or CVs going for damage only. Those  also are important aspects of teamplay in the current meta.

 

 

Looking at the raw numbers (of the above average players that is) your conclusion is right. But please consider the aspects, which can not be seen.

 

 

[Figment]

Can you compare base cap points and defense to win and draw rates?

 

Cap / defense per class would be nice too. I'm expecting that caps would be less relevant than defense.

[dumbo11]

When doing this analysis have you excluded premium ships?  (those are balanced below normal ships and will tend to drag down averages).

 

 

 

 

[Trigger_Happy_Dad]

Excellent work, thank you very much @ ezymodo!

 

Yesterday I made this:

 

 

Just to show that it is indeed a problem that tier VI, VII and VIII DDs make so little damage against so many HPs.

 

Schaden pro Runde der verschiedenen Schiffsklassen von tier V - VIII:

 

V Nicholas:  20k damage >>> VI Farragut:  20k damage >>> VII Mahan: 19k damage >>> VIII Benson: 20k damage          =  79k damage overall (tier V - VIII)

 

V Minekaze:  27k damage >>> VI Mutsuki:  19k damage >>> VII Hatsuharu: 20k damage >>> VIII Fubuki: 25k damage    =  91k damage overall (tier V - VIII)

 

V Omaha:  29k damage >>> VI Cleveland  31k damage >>> VII Pensacola  25k damage  >>>  VIII New Orleans  30k damage     =  115k damage overall (tier V - VIII)

 

V Furutaka:  16k damage >>> VI Aoba:  28k damage >>> VII Myoko: 37k damage >>> VIII Mogami: 44k damage    =  125k damage overall (tier V - VIII)

 

V Kongo: 27k damage >>> VI Fuso: 37k damage >>> VII Nagato: 38k damage >>> VIII Amagi: 47 damage                         =  149k damage overall (tier V - VIII)

 

V New York:  30k damage >>> VI New Mexiko: 39k damage >>>  VII Colorado: 35k damage >>> VIII North Carolina: 49k damage    =  153k damage overall (tier V - VIII)

 

V Bogue: 25k damage  >>> VI Independence:  43k damage >>> VII Ranger: 48k damage >>> VIII Lexington 49k damage           =  165k damage overall (tier V - VIII)

 

V Zuiho: 49k damage >>> VI Ryujo: 59k damage >>>  VII Hiryu  58k damage >>> VIII Shokaku 67k damage                                  =  233k damage overall (tier V - VIII)

 

But the idiots still don't get it that it is a problem if a tier VI / VII / VIII DD only manages to cause the same damage like a tier III / IV DD.

 

http://forum.worldofwarships.eu/index.php?/topic/27244-dds-ab-t8-dauerparkticket-im-hafen/page__st__200__pid__488333#entry488333

 

When doing this analysis have you excluded premium ships?  (those are balanced below normal ships and will tend to drag down averages).

 

Premium ships are not in the database yet.

 

They are not relevant for this discussion anyway.....

 

[Hannibalurg]

People need to get a job.

[dumbo11]

Premium ships are not in the database yet.

Not sure what you mean - the Atlanta/Aurora etc. are certainly on that stats page?

[Trigger_Happy_Dad]

Not sure what you mean - the Atlanta/Aurora etc. are certainly on that stats page?

 

Last time I looked they were not.

 

And again - premium ships are not relevant here.

[Broevaharo]

To be honest, none of these stats surprise me, it's what I experience in the games I play every day.

I play both DD lines, both CV lines, both CL lines + the Kongo and Fuso and I can see on first hand that tier 7 is a show stopper for most players (my highest tier ships are CV's and the Fubuki at tier 8)

Grinding the Mahan and the Fubuki is a nightmare, grinding the Pensacola and the Myoko is okay-ish, the Lexington is a ship that can't do anything good in its tier and all the IJN CV's are over performing compared to the other ships.

 

Looks like the stats prove what I was saying all along, the DD lines are massively under performing  in the higher tiers and really need some TLC from WG.

I'm also not really surprised about the better performance of the IJN CL's, they have far superior guns (flat trajectory = easier to score hits) to their US counterparts and even though they have bigger citadels than the US, the abilitly to hit targets reliably clearly wins.

[allufewig]

People need to get a job.

 

 

Someone puts effort into analysing and visualising statistics and all you do is pick on him?

 

For someone who doesnt care for the game you play a damn lot and write a damn lot. Get a job yourself, Mr. #1 bullshitposter.

[dumbo11]

Doing some eyeball checking, it seems the average stats do seem to exclude premiums, so ignore that question .

My own comments:

-  when people analyse stats, tiers 9 & 10 are usually excluded.  It's unclear whether those tiers lack players, are 'seal-clubbing' and/or WG haven't even attempted to balance them. [probably all of them]

- after T5 the roles of some ships change.  The DD silent assassin is lost due to air spam and beaten by torpedoes dropped from point-blank range.  The USN CA switches from the best hunter/killer into a more 'support' role.

 

There's certainly work to be done post T5...

- the USN support CAs seem effective enough in winning matches, but they aren't rewarded properly for it.  (rather than AA rewards being too low, I suspect a major problem is that some matches simply have no aircraft).

- IJN CAs are slightly too effective around T7-8.

- DDs... not a shock .

- T9+ lol.

- a few hopeless ships (furutaka/langley/colorado).

 

In terms of BBs, the win-ratio of BBs is actually not good/bad until you reach the 'wonky tiers'.  If DDs get buffed then the USN BB line would likely have all of the lowest win-ratios from T5-7 (and within hair's breadth of tier 8).

[ilhilh]

Thanks for this.  I love glimpses into global stats.

[XueLong]

Best forum post I've seen for a long time. Keep up the good work and many thanks mate!

[22cm]

I m waiting for the CV fan-club boys to start flooding this topic with their usual "CVs are not OP, it s just skill", "number of battles on top-tiers is too small", "CVs are high -risk high-reward ships", etc. You know, their usual crap.

[Seafort]

I m waiting for the CV fan-club boys to start flooding this topic with their usual "CVs are not OP, it s just skill", "number of battles on top-tiers is too small", "CVs are high -risk high-reward ships", etc. You know, their usual crap.

 

We'll also see how much they [edited] when they can't play their favourite money making ships as often as they want due to the changes in MM

[SISHIN]

Very interesting reading.

 

One thing that bug me most is the economy they made such way that to prevent players to play T10 ships exclusively after owning them.

 

However what they actually have done is preventing players from sailing T10 ships. Yeah, it's really sad to see that tier 10 ships only represents 1.2% of the total battles.

 

We know that the percentage will increase as more player got they hand on that tier. But if no economy adjustment have been made, I would expect to see any major increase there.

 

 

[Seafort]

Very interesting reading.

 

One thing that bug me most is the economy they made such way that to prevent players to play T10 ships exclusively after owning them.

 

However what they actually have done is preventing players from sailing T10 ships. Yeah, it's really sad to see that tier 10 ships only represents 1.2% of the total battles.

 

We know that the percentage will increase as more player got they hand on that tier. But if no economy adjustment have been made, I would expect to see any major increase their.

 

They don't want people to play their Tier 10 ships so they have ludicrous ammo and repair costs to stop people from doing just that. I'm personally struggling with affording tier 7 ships as I don't play as often as some people so don't make enough credits to afford the upgrades now nevermind new ships.

 

I don't see myself lasting much longer playing this game as the grind is ridiculous. I'll probably stop playing when my premium account is up in 20 days or so.

[Luckyio]

It's fairly obvious that this is working as intended. CVs are supposed to be an elite club of true pros. BBs are slightly bugged in sometimes being able to take CV's kills from him/her. This buggy capability needs quite a bit of  fixing to prevent this kind of abuse.

 

Rest are exactly where they need to be.

 

What next, someone will suggest that seals need to beat a club? That's like suggesting that scissors need to beat rock.

[ezymodo]

I see that there are some emotional reactions in this thread, especially those connected with CV (either pro - or anti-).

If you think I'm an anti - CV player, I'm sorry to say that I'm primarily a CV player.

50% of my games are in CVs, and my most played ship is Hakuryu (122 games), with around 142K average damage.

I learn a lot from watching the videos and tutorials of the better players, and getting sunk by them,   and I'm still trying to improve on that ship to beat the masters.

So I don't get any benefit by disclosing the analysis; in fact, I seem to hurt my own interest by doing so. 

 

But you know what? Wargaming doesn't care.

They don't care that people are screaming that this or that is OP or whatever, or this and that is unfair or UP.

This is because most people are emotional, as can be seen by the numerous threads in the forum.

What they listen to, is numbers. This is the case for WOT, and this is the case for WOWS as well.

 

So let us take Wargaming's perspective. What do the numbers tell them?

1. Most customers play at mid - tier, and this tier is the one that needs to be taken care of first (80% of customers at tier 1 - 6... pareto principle).

2. That at mid tier, the IJN CVs are overperforming in most metrics.

3. That at high tier, the ship performance vary widely. They also know that there is not enough player in that tier yet, and they are waiting for the fluctuations to settle down.

4. That DDs at all tiers are underperforming, irrespective of nation.

5. That most of the ships people complain about is, in fact, underperforming.

 

And you can see what they do in the next patch: they give some nerf to ijn Cvs, buff to some dds, and so on. This is consistent with the cold, hard numbers available to them.

Judging from the numbers, we might also infer that a future major rebalance would be done on higher tiers, when they think that enough customers are already there.

Do I think IJN CVs are balanced? absolutely, it's totally balanced. Does the number agree with me? Nope, and Wg is on the side of the numbers.

 

So, what can I do about it? Seeing the meta development ahead of the patch, and grind and develop skills accordingly.

In this case, it would be rational to grind both CV and BB line to hedge the bets, since both are strong contenders and it is unlikely that both will be nerfed at the same time.

In my case, this would meant grinding the USN or IJN BB line, since I already got a CV line. It is a very nice line, and I will enjoy it while it lasts.

By the time the nerf hits the Haku, I hope I already got a decent BB with the skills to use it well. 

And I hope everyone would take the same approach towards grinding, since it takes some amount of investment in time and/or money .

I would leave this figures over here to those that are still confused about which line is worthwhile. Look at the numbers, and judge for yourself.

 

P.S. As for the comment of Hannibalung, I'm now in the process of writing my Ph.D thesis and still trying to find a job. If anyone have one, that would be awesome .