Jump to content


Exploring WG's database

datamining community status statistics analysis please be civil

  • Please log in to reply
153 replies to this topic

TocFanKe4 #141 Posted 28 July 2017 - 07:23 PM

    Major

  • Players
  • 20912 battles
  • 24,177
  • [GIRLS]
  • Member since:
    03-16-2014

View PostMr Crowley ll, on 28 July 2017 - 01:24 PM, said:

 

 logging in tests could throw the active vs inactive account numbers.  Logging in tests wont throw off the daily player counts because they actually need to play 5 games to be counted.

 

But shouldn't it skew them towards more active players? So things would be worse than they seem?


 

The E25 was built for Shenanigans


Kamakazi Rusher #142 Posted 28 July 2017 - 07:31 PM

    Sergeant

  • Players
  • 6623 battles
  • 109
  • [RDDT]
  • Member since:
    03-09-2014

View PostTocFanKe4, on 28 July 2017 - 01:23 PM, said:

 

But shouldn't it skew them towards more active players? So things would be worse than they seem?

 

Actually, if they're merely logging in but not playing any matches, neither Crowley's nor my own data set would be affected. Both only take into account the battles played.


Code Monkey, Network Engineer, and Ethical Systems Penetration Tester

Community Status Tracker - Official Thread

Take a look at my post where I datamined WG's servers! Account registration, inactive players, premium ownership, all that good stuff!


Mr Crowley ll #143 Posted 28 July 2017 - 07:33 PM

    Major

  • Players
  • 23621 battles
  • 3,155
  • [CRPT]
  • Member since:
    02-22-2014

View PostTocFanKe4, on 28 July 2017 - 01:23 PM, said:

 

But shouldn't it skew them towards more active players? So things would be worse than they seem?

 

depends, give it 6 months and compare the the active account data against daily player counts. 

 

 


Serpia WG #144 Posted 28 July 2017 - 07:37 PM

    Major

  • Players
  • 2346 battles
  • 5,600
  • [ONE]
  • Member since:
    06-26-2014

View PostHeroForADay, on 28 July 2017 - 07:49 AM, said:

^ what balance changes? There's been nothing significant in a year and a half. Mean while PC is making drastic changes to numerous lines to update nearly the entire game. 

 

Your veterans are burning out. 

Balance changes come with each server update. If PC is making some cool changes it means that more likely we will have some of them when we will be on the same server version. There is always some space for dispute on how much "significant" those changes are ;) 

 

View PostKamakazi Rusher, on 28 July 2017 - 06:08 PM, said:

"Login test sessions." Could you expand on this a little? What do they test for, exactly? Garage synchronization? And how often do you perform this? Would be pretty neat to get a look at some of the tests you guys perform!

 

Also, how would they "corrupt" new player numbers? I can see how it would impact the count of (in)active players if they were to log in (in)frequently, but just how many accounts does this entail?

Have not so much to tell about test sessions, but it looks like December 2014 spike you got is one of them. Believe kind of stress test or something like that. We do not put much attention on it as we are able to measure actual logins and never used the API data.

 

View PostKamakazi Rusher, on 28 July 2017 - 06:08 PM, said:

And that's completely understandable. But you should also realize that when the data isn't readily available, people are going to find ways of measuring it out themselves. This is why I'm working on a tracker that will record how many battles every player has participated in on a daily basis. It should help us record over time how many active players there are, as well as analyze the number of battles the servers are having to host each day.

And that is completely understandable too :) People are curious taking care of game future and this is great! I just do not want anyone to be upset with some data and there is actually even more feasible reason then incomplete data: f2p and p2p metrics are different. As an example, 5% of retention you`ve calculated looks scary, but what is the average metric for f2p titles is lower? It doesn't look scary anymore ;) Still, we do not calculate this metric the way you do, so even can`t compare. 

 

View PostKamakazi Rusher, on 28 July 2017 - 06:08 PM, said:

I'm sorry, but I have to file a complaint right here. Forgive me if I've had my head in the sand since I don't watch any of WG's livestreams, but here's where I feel that communication between WG and the community suffers.

 

Your data shows that "new players prefer to try PvE first." I've never, ever heard this before. Whether or not it's a "no brainer," I don't think I've ever read an official WG statement that includes this. It makes sense now as to why WG had put quite a bit of effort into Proving Grounds for update 3.7. Yet, when it was released, we were only told that this was to enable Co-op, something that most veteran players wouldn't care for (since Proving Grounds puts you in the red for higher tiers when it comes to silver). Many of us were quite confused as to why this was made a priority, or rather shocked that it came out of nowhere. I think if WG came out and said, "We've implemented Co-op so that you can play Proving Grounds with your friends! We hope that this will help new player experience too as we find that they prefer to try PvE before PvP" it would've answered some questions and eased the concerns of players who are still anxiously waiting for tank re-balancing or new lines.

I feel you :( From time to time communication suffers. Dev team is not supposed to be presented on forums to answer the questions constantly, that is community team role usually. That is why I`m here now and the plan is to change the situation to better one.

Yes, PVE is popular through new players and it was developed to help new players to join the game in a less stressful way. CoOP is the next step - new player can join the game with friends and investigate game mechanics in a good company or can ask more experienced players to help. Still, there are many experienced players trying new tanks in PVE too :) 

 

View PostKamakazi Rusher, on 28 July 2017 - 06:08 PM, said:

While I understand how controversial it has become, Wargaming seems to be solely focused on the data/graphs that hint towards (in)active player counts. I haven't heard anything that confirms or denies data we've pulled/compiled on things like tank popularity and some assertions made from them, such as how the WT auf E 100 probably still exists because it's in the Top 5 most-player Tier 10 list. I get that WG is on damage control trying to assuage concerns of the playerbase dying off, but for the things that you can comment on, you've pretty much chosen not to.

This is cause most of the tanks data is correct. But I`m happy to help with those questions. If I remember the statement right, dev team didn`t want to change favorite players tanks (this will make almost all of its owners upset) if stats are still +/- balanced. Let`s gather more questions regarding tanks, I will pass them to the dev. team and turn back with answers. 

 

View Postdrewptwo, on 28 July 2017 - 06:30 PM, said:

 

I have to say I appreciate the detailed response and willingness to open conversation.  Although I don't understand a ton of what is being discussed (I'm not the super tech savvy analytical type) it is refreshing to see somebody from WG respond in a less ambiguous manner and offer some additional insight as well as a relatively open dialogue.  Just my two cents...  

Thank you! Appreciate it :)

 

View PostTocFanKe4, on 28 July 2017 - 06:37 PM, said:

Wouldn't login test sessions cause it to seem like more people are logging in? In that case a portion of the active player group could just be WG testing things, which means the problem is actually much worse than the numbers show?

My question would be how could login tests cause smaller numbers of active players?

Test sessions are just an example where API can not provide correct data. It doesn`t mean that other data is correct. September of last year shows more than 10,5 million of logins as an example (while only 9 million accounts API provides) and surely number of accounts is bigger than the number of logins (shame on me I do not remember what is the number of accounts in Kamakazi Rusher`s graphs). And what problem do you see? I believe you just trying to compare f2p stats with p2p expectations. 

 



VoltaicGrunt #145 Posted 28 July 2017 - 07:59 PM

    Captain

  • Players
  • 26330 battles
  • 1,559
  • Member since:
    04-18-2015

For those who don't know who I am, which is everyone, I run a business. I started off by inventing a tool to fit a machine to make a fastener which is used by the billions in all kinds of industries. I reduced production costs from 2.73 cents per unit to .98 cents per unit. Previously it took 3 men, 3 machines an 4 processes to make this clip. My tool reduced this to 1 machine, 2 men and 1 operation.

 

I make huge amounts of cash from licensing the tool to manufacturers. I use that cash buying other businesses. I can look at company accounts and tell you exactly what is going well, and what isn't.

 

I do not need numbers. I just need the graphs. The numbers are largely  irrelevant. It's the TREND which is important.

 

In this case, the trend is downwards, quite significantly, when it comes to 90+ day player retention. The other issue that stands out is the number of New/Inactive accounts. Without those numbers, a gloomy picture looks dire.

 

TEST ACCOUNTS

With 100K new players per month (60% inactive) and 100K avg players per day, HOW MANY test accounts must you create to skew the data?

 

 In the 1970's, American Express and the Dale Carniegie Institute did a study. They identified it is 11 times cheaper to retain an existing customer than gain a new one.  AMEX built their business on that.

It's no longer true. Mass marketing via interweb has made the cost of customer acquisition incredibly cheap, so WG's business model relies on "churn". Masses of new players coming into the game, a certain percentage of which will stay, and spend money.

When you cannot retain long term customers, what you're losing is actually far larger than the loss of a new player, because the majority of the income must come from the long term players.

A long term customer is incredibly valuable. They have gone through the entire acquisition process. Of 100 new customers, 9% play 90+ days, with 1% playing long term. So when ONE veteran player, a money spending customer leaves, WG must generate 100 new player accounts to replace him.

 

WHO ARE CONSOLE PLAYERS?

I have never seen a customer profile for Tanks. I have no idea of customer demographics or financial patterns for tanks. This is all confidential information, for good reason.

 

In general, console players are casual. They play more sessions per week than PC players, but for a shorter time.

Console players do not stick with a single game long term. AAA+ franchises release a new game on an annual cycle for that exact reason. Console players might stick with a franchise long term, but always want the newest latest shiniest version.

 

This creates a unique situation for WG, because it's a F2P game funded by micro-transactions. This is not the usual business model on Console, because it relies on long term loyal customers.

Which makes WGCB's inability to retain long term customers devastating for their cashflow.

 

WG's RESPONSE

Research shows new players prefer PvE. Really? Where exactly does this research come from, because it flies in the face of industry norms.

THE TOP 10 BEST SELLING CONSOLE GAMES ARE ALL PvP.

If that research comes from console tanks, it's a pretty damning indictment of the gaming experience new players are having. In fact, it is the most blatant admission of failure I have ever heard. Let's also point out that there is NO INCENTIVE to buy Premium time or Premium Tanks for a PvE scenario.

 

So, instead of following the example set by American Express, WG has decided to maximise New Player Acquisition at the cost of long term players. This is I believe, a huge, possibly fatal mistake.

 

Could I be Wrong?

YES, of course, I don't have enough info to lay things out completely. I may be wrong in thinking long term customers are the one's spending money, for example. It may be that WG makes more money from new players who buy a Premium tank the first week, then get rekt and quit. It may be that the biggest headache WG has is that it cannot get rid of it's long term customers. It may be that the thing holding the game back is Veteran players smashing new players so badly they all quit. It may be that Veteran players already own Premium Tanks, have finished grinding, and no longer need to buy new tanks or time, in which case, it's in WG's best interest to get rid of Veterans.

 

Whatever. This game is not dead, but I can smell the rotting flesh. I give them 2 years, tops.

 


RU 251 NEEDS MORE BULLETS


butsubutsu #146 Posted 28 July 2017 - 11:44 PM

    Major

  • Players
  • 21209 battles
  • 3,214
  • Member since:
    02-14-2016

View PostVoltaicGrunt, on 28 July 2017 - 07:59 PM, said:

For those who don't know who I am, which is everyone, I run a business. I started off by inventing a tool to fit a machine to make a fastener which is used by the billions in all kinds of industries. I reduced production costs from 2.73 cents per unit to .98 cents per unit. Previously it took 3 men, 3 machines an 4 processes to make this clip. My tool reduced this to 1 machine, 2 men and 1 operation.

 

I make huge amounts of cash from licensing the tool to manufacturers. I use that cash buying other businesses. I can look at company accounts and tell you exactly what is going well, and what isn't.

 

I do not need numbers. I just need the graphs. The numbers are largely  irrelevant. It's the TREND which is important.

 

In this case, the trend is downwards, quite significantly, when it comes to 90+ day player retention. The other issue that stands out is the number of New/Inactive accounts. Without those numbers, a gloomy picture looks dire.

 

TEST ACCOUNTS

With 100K new players per month (60% inactive) and 100K avg players per day, HOW MANY test accounts must you create to skew the data?

 

 In the 1970's, American Express and the Dale Carniegie Institute did a study. They identified it is 11 times cheaper to retain an existing customer than gain a new one.  AMEX built their business on that.

It's no longer true. Mass marketing via interweb has made the cost of customer acquisition incredibly cheap, so WG's business model relies on "churn". Masses of new players coming into the game, a certain percentage of which will stay, and spend money.

When you cannot retain long term customers, what you're losing is actually far larger than the loss of a new player, because the majority of the income must come from the long term players.

A long term customer is incredibly valuable. They have gone through the entire acquisition process. Of 100 new customers, 9% play 90+ days, with 1% playing long term. So when ONE veteran player, a money spending customer leaves, WG must generate 100 new player accounts to replace him.

 

WHO ARE CONSOLE PLAYERS?

I have never seen a customer profile for Tanks. I have no idea of customer demographics or financial patterns for tanks. This is all confidential information, for good reason.

 

In general, console players are casual. They play more sessions per week than PC players, but for a shorter time.

Console players do not stick with a single game long term. AAA+ franchises release a new game on an annual cycle for that exact reason. Console players might stick with a franchise long term, but always want the newest latest shiniest version.

 

This creates a unique situation for WG, because it's a F2P game funded by micro-transactions. This is not the usual business model on Console, because it relies on long term loyal customers.

Which makes WGCB's inability to retain long term customers devastating for their cashflow.

 

WG's RESPONSE

Research shows new players prefer PvE. Really? Where exactly does this research come from, because it flies in the face of industry norms.

THE TOP 10 BEST SELLING CONSOLE GAMES ARE ALL PvP.

If that research comes from console tanks, it's a pretty damning indictment of the gaming experience new players are having. In fact, it is the most blatant admission of failure I have ever heard. Let's also point out that there is NO INCENTIVE to buy Premium time or Premium Tanks for a PvE scenario.

 

So, instead of following the example set by American Express, WG has decided to maximise New Player Acquisition at the cost of long term players. This is I believe, a huge, possibly fatal mistake.

 

Could I be Wrong?

YES, of course, I don't have enough info to lay things out completely. I may be wrong in thinking long term customers are the one's spending money, for example. It may be that WG makes more money from new players who buy a Premium tank the first week, then get rekt and quit. It may be that the biggest headache WG has is that it cannot get rid of it's long term customers. It may be that the thing holding the game back is Veteran players smashing new players so badly they all quit. It may be that Veteran players already own Premium Tanks, have finished grinding, and no longer need to buy new tanks or time, in which case, it's in WG's best interest to get rid of Veterans.

 

Whatever. This game is not dead, but I can smell the rotting flesh. I give them 2 years, tops.

 

 

Sad to hear your educated guess about WoT console future. I believe that you may be correct unfortunately. The veterans are the ones spending money and they stay because they spent money. I have spent thousands of euros in less than two years on premium tanks mostly but also other things. I know that if I quit, those "investments" will be for naught. So I am inclined to continue the game and my spending for selfish reasons and the fact that I really love this game. 

 

At at the same time I do think we need more new players but we need to face the facts. Most of us playing are over 30 and makes quite good money and that makes us able to spend way too much on pixels and polygons. I think the solution here would be to lower the prices for premium tanks and other stuff so a whole new demographic will feel they are able to compete and play this game. Micro transactions that actually are micro would widen the demographics that will be able to participate and pay for services in this f2p game. More customers means that every customer needs to pay less for a functioning game and that is a win for all parts included. WG needs to decided if they just want to cash grab now or make a game that will make more money long term. It is also a way to determine how the customers will approach WG in the future. If this game dies on console and it is WGs fault in my opinion, I will never touch anything from WG on any system again. 

 

And WG should definitely stop focusing on the PVE part of the game and start focusing on what really matters, PVP!



RogueMerc69 #147 Posted 29 July 2017 - 09:44 PM

    First lieutenant

  • Players
  • 12088 battles
  • 506
  • [RDDT]
  • Member since:
    10-19-2015

View PostSerpia WG, on 28 July 2017 - 07:37 PM, said:

Balance changes come with each server update. If PC is making some cool changes it means that more likely we will have some of them when we will be on the same server version. There is always some space for dispute on 

 

This is cause most of the tanks data is correct. But I`m happy to help with those questions. If I remember the statement right, dev team didn`t want to change favorite players tanks (this will make almost all of its owners upset) if stats are still +/- balanced. Let`s gather more questions regarding tanks, I will pass them to the dev. team and turn back with answers.

 

I think most people want actual balance, not buffing and nerfing a few tanks once every 2 years. Also the second statement is a bold faced lie. If they didn't wan't to change peoples favorite tanks, how come the Foch was nerfed so much that it became a meme, how come they removed the FV4202 from tier 10? Also the lying aside, that's a terrible reason to leave imbalanced/broken tanks in the game, some people liking a broken mechanic doesn't mean you leave it in and punish the rest of the people.



HeroForADay #148 Posted 29 July 2017 - 10:08 PM

    Staff sergeant

  • Players
  • 13718 battles
  • 330
  • Member since:
    02-07-2014

View PostRogueMerc69, on 29 July 2017 - 03:44 PM, said:

 

I think most people want actual balance, not buffing and nerfing a few tanks once every 2 years. Also the second statement is a bold faced lie. If they didn't wan't to change peoples favorite tanks, how come the Foch was nerfed so much that it became a meme, how come they removed the FV4202 from tier 10? Also the lying aside, that's a terrible reason to leave imbalanced/broken tanks in the game, some people liking a broken mechanic doesn't mean you leave it in and punish the rest of the people.

 

Exactly.

 

The term significant is defined as "sufficiently great or important to be worthy of attention; noteworth" 

 

Our balance changes happen so infrequently, and even when they have, they've been done ridiculously. Some of us don't forget the fiasco that was the German heavy tank nerf then buff. Then yeah, the WTE100 was changed years ago now on PC, still got the OP as hell version... Still have the ridiculous Deathstar that's now buffed more. 

 

So don't patronize the community. The rarest TD in the game is a meme, the most OP still OP, but don't worry our latest balance change was only half of what was changed on PC. Poor Lowe didn't get his armor buff. 

 

One or two tank changes isn't significant, so don't make me laugh. Ya'll cherry pick the hell out of the PC changes because you think you know better about how this game should be developed. But yeah, I'm sure the 4202 is going to be super competitive for awhile since that server change happened a long time ago... 


#ConsoleTax

Why do the Defender and Patriot have higher costs on Console than they do PC?

Well, that is simple. The answer is because World of Tanks PC is a different game than World of Tanks Console. And their tanks are not always the same as our tanks even though they might look the same, the stats are sometimes different, the costs are sometimes different. So… uh.. You can’t ever, for anything whether it’s… uh… the maps, the gameplay, the costs of the tank, the look of the tank, the stats of the tank, anything like that. You can never compare the two. The only thing that they really have in common is that they are both made by Wargaming, they both have the name World of Tanks, and they both have tanks in them. So that’s uh… the reason.


Kamakazi Rusher #149 Posted 31 July 2017 - 11:30 AM

    Sergeant

  • Players
  • 6623 battles
  • 109
  • [RDDT]
  • Member since:
    03-09-2014

View PostSerpia WG, on 27 July 2017 - 07:25 PM, said:

In case you or any other forum member will have additional questions or ideas, I will be here! Personal messages are open for everyone as well! 

 

Hey Serpia, as I was structuring a database I came to find a listing of players who look like their account IDs were altered for some reason? Their usernames would have "_old_######" appended to them. Could you or your team explain why these are the way they are? There are some duplicates too, such as Dit31 and Dmixen, and oddly enough a lot of XGN players.

 

Here's a link to a list of accounts.


Code Monkey, Network Engineer, and Ethical Systems Penetration Tester

Community Status Tracker - Official Thread

Take a look at my post where I datamined WG's servers! Account registration, inactive players, premium ownership, all that good stuff!


I SierraHotel I #150 Posted 31 July 2017 - 02:06 PM

    Captain

  • Players
  • 31195 battles
  • 1,981
  • Member since:
    09-09-2016

View PostKamakazi Rusher, on 31 July 2017 - 11:30 AM, said:

 

Hey Serpia, as I was structuring a database I came to find a listing of players who look like their account IDs were altered for some reason? Their usernames would have "_old_######" appended to them. Could you or your team explain why these are the way they are? There are some duplicates too, such as Dit31 and Dmixen, and oddly enough a lot of XGN players.

 

Here's a link to a list of accounts.

 

:trollface: 

good one!


If you play against the best, you learn the most difficult elements of the game, if you play against arrogant and empty shells, you lie to yourself that you are good.

This forum is a direct reflection of the quality of the product - WoT


SkookumChunk #151 Posted 31 July 2017 - 03:15 PM

    Corporal

  • Players
  • 17005 battles
  • 78
  • Member since:
    04-20-2016

I don't really understand the "December 2014 spike" explanation of "we were doing stress tests". 

 

If WG was doing stress tests wouldn't they be doing them on test servers? Does that API really return data from test servers as well? Is there a way to filter that data out?

 

I guess it is possible to stress test production servers while they are made unavailable to the public.

 

https://en.wikipedia...sting_(software)

 



VoltaicGrunt #152 Posted 31 July 2017 - 03:31 PM

    Captain

  • Players
  • 26330 battles
  • 1,559
  • Member since:
    04-18-2015

View PostKamakazi Rusher, on 31 July 2017 - 11:30 AM, said:

 

Hey Serpia, as I was structuring a database I came to find a listing of players who look like their account IDs were altered for some reason? Their usernames would have "_old_######" appended to them. Could you or your team explain why these are the way they are? There are some duplicates too, such as Dit31 and Dmixen, and oddly enough a lot of XGN players.

 

Here's a link to a list of accounts.

 

Could these be renamed accounts? XGN insist their members change their GT to include XGN in a similar way we insist lepers carry bells.

 


RU 251 NEEDS MORE BULLETS


Kamakazi Rusher #153 Posted 31 July 2017 - 04:21 PM

    Sergeant

  • Players
  • 6623 battles
  • 109
  • [RDDT]
  • Member since:
    03-09-2014

View PostVoltaicGrunt, on 31 July 2017 - 09:31 AM, said:

 

Could these be renamed accounts? XGN insist their members change their GT to include XGN in a similar way we insist lepers carry bells.

 

 

I could see that. Question is, why create a new database entry instead of renaming/changing the player nickname?

 

Edit: It doesn't fit for all accounts. Some are the exact same names.


Code Monkey, Network Engineer, and Ethical Systems Penetration Tester

Community Status Tracker - Official Thread

Take a look at my post where I datamined WG's servers! Account registration, inactive players, premium ownership, all that good stuff!


SkookumChunk #154 Posted 01 August 2017 - 12:25 AM

    Corporal

  • Players
  • 17005 battles
  • 78
  • Member since:
    04-20-2016

I think the renamed accounts are the result of a database datafix (like the one that was done when people couldn't move their crew back into their premium tanks). They created new accounts with new IDs for these people. WG left the old vestigial accounts lying around after the datafix was run (which is pretty sloppy I think)

 

For example: PizzaKraecker17 has a new account id of 10679079 vs their old ID of 10601843.

 

http://www.wotinfo.n...079&server=XBOX






1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users