Lorem ipsum dolor sit amet, elit eget consectetuer adipiscing aenean dolor

Are we burnt out? What do the server logs say?

26 games per day? Because playing other games for 8hrs a day is takes up everyones time! :rofl:
I got 52k+ in GW B1 tho, so I’m happy with that.^^

2 Likes

I just play GoW while waiting for my other games to load or find opponents …

2 Likes

You’re making a lot of assumptions about which data they do or do not keep, which may be true or may not. I don’t know either, so I’m not going to debate your assumptions with my own, but I will say this. To keep track of final score result in every guild event going back a few years, for every single player from every guild, would not be a large amount of data by modern database standards.

It is at least possible they would off-load and store that high level data in an archive database separate from the main game server (which of course has no need for past history). It’s just a matter of whether they deemed it useful to keep or not. I guess the whole thing is moot anyway since they would never be allowed to share the info with us, but it is fun to speculate.

To the best of my understanding and comprehension of the matter, then simultaneous running of World Events and Guild Wars is intentional. I’ll disclaim that statement with the usual “I’m not a dev/designer” statement, but I’m extremely sure it was said somewhere (on-stream, I think?) that the change was made for economic reasons.

My opinion here, but I very strongly believe that the devs have zero intentions of ever running Guild Wars as a seperate week ever again.

Of course, they are looking at the data. Live services live and die by reviewing the data and then taking action.

All in all, I think the issue here is that the “right questions” you want the devs to review are not the “right questions” being asked from the devs’ perspective on the situation.

This is a sound example of a question the devs are asking from their perspective on the matter.

4 Likes

I agree completely about redoing the schedule to avoid this type of overload. Having three significant events running at the same time was too much.

I think they should run a 13-week schedule to basically have quarterly campaigns. 9 weeks of regular campaign world events, with 3 guild wars mixed in before each 3 of the 9, followed by week 10 of campaign where tower of doom is the final event of the campaign! They could even make a special special ultimate doom boss room that includes the campaign villain or just have the lore tie to it in some way. Either that or have ToD run concurrently with the 10th week of campaign but make sure no other events happen that week.

1 Like

In my case, I can say that I didn’t play less, but that more of my play was directed by the overlapping events.

So, I had less time to just play the parts of the game that I wanted to play, and more time devoted to “homework”.

4 Likes

Of course it would not be large, but what would be reasoning to store such data ? :wink:
And why would you want to keep it? If you want to analyse players activity, you dont need to keep every single record. You can gather only cumulative statistics and that can give you some
data like: 1% of guilds reached reward12, 5% of players bought at least tierI from the shop etc.

Main thing is - you need to plan in advance to collect such data and do such summaries.
If you dont plan it, you wont have the data stored.

worked few years with huge and complicated system which logged some basic data about user activity (time of login, time of logout, tables/raports accesed by user, any errors they might receive)
it was just such small portion of what could be logged, but it generated so much data, that it became a problem, which we had to manage.

My assumption is that they collect only minimum part of data, which is nessesary for running the game.
Other than that, they might collect some extra data, when they have a thesis they want to test.

You ask why they would want to retain it and I ask why they would not? I’m just saying that storing that weekly event results data is not some arduous thing to do, nor would it have great expense tied to it. The player activity in those events (in particular where players spend gems) is directly tied to their revenue stream, so they may find having that type of data available worthwhile.

You make the point about needing to plan ahead, and that is true, but in my experience that’s exactly why companies like to retain raw data. If you have the raw data you can always parse it out any way you want later, mitigating the need to foresee every future information requirement.

I had a feeling your response might include an indication that you had some experience working with volume data. I was figuring possibly as a DBA or similar position based on your responses, but not someone that works with analytics or BI reporting. I’ve been working with data for large corporations for about 15 years myself, so I also have some experience with data retention and reporting policies. People would be surprised how much data that most would find completely inconsequential gets retained for years in that setting, especially anything to do with financials.

Ofc, companies are collecting a lot of data, which they aren’t using atm. Thing is, even if they do collect data, which could be used to answer Starlite questions, they would need to process it, in order to have that info. Even if we ignore computing time, which is needed, you also need to pick up a raw criteria, fe. what does:
“finishing a faction” mean ?
getting it’s reown to 2500? or maybe just getting it’s level to 500? or maybe just getting the last reward from the event ?
etc.

But than again, i highly doubt data about players performance in events is kept for longer than the event and let’s say extra week or two… Easiest example would be GW results, which are shown to players for a week after GW, and after that week, there are no results. Simple data-retention. And seeing as people are reporting wrong calculations of GW scores, and Kaffka was asking for more data from players -> it might be they dont have historical data anymore.

Another asumption would be, that unless you have some timestamped data store -> nobody is keeping track of previous player gamestate (quests done, challenges done, faction reown gained, troops owned, weapons owned etc.) - this data i probably stored and updated in realtime/semi realtime and only way to restore previous values would be to do a database restore from a backup.

Some of these questions require also that data stored would have exact time of action done (answering “how many people rushed Guild Wars” would require a timestamp for each GW battle). Which might be true for data in logs, but than again -> logs aren’t things that live long (unless you want to make a hadoop-like data lake and analyse that data in future- but that requires additional data-mining skills which probably aren’t popular amongst Android/IOS developers :wink: )

To sum up my assumptions, as you pointed out:

  1. I doubt some of data required to answer Starlite questions is not kept long enough to answer question.
  2. Even, if data is stored it would require additional data-mining skills from developers, which I doubt many have.
  3. Even, if there is a person with knowledge and data, he/she is probably too busy doing analyses for his boss and doesn’t have time to answer these questions. Especially they aren’t that much relevant to company finances.
    When we talk about big numbers, a group of old players quiting isn’t a problem as long as they are replaced with new ones. Heck, probably new players are more eager to spend money on this game than older ones, who have huge reserves of gems/keys/etc. So it’s more “healthy” for the game (from financial perspective) to see old players quiting and new players joining.
1 Like

I agree with you on most points, but this is the part I have doubts about. If I was running this business I would definitely want that information retained. Not in the active game server of course, that would be inefficient, but exported to a data warehouse server for future analysis. A service to hold that type of data with minimal connections regularly hitting it would be pretty cheap. Of course I’m not running this business though, and they may have not had the foresight to to do that, but I wouldn’t rule it out. I wouldn’t expect them to be answering these questions to us on the forums either, but if they were competently running their business this is exactly the type of information that they should be looking at regularly and analyzing as part of their decision making process (but of course that’s a big if).

Note: You’re right that data mining isn’t the same skillset as the mobile developers, but it also is not a difficult skillset to find or develop either. It’s actually pretty easy to learn.

My guess would be they keep track of:
How many offers were bought last month compared to previous months?
What was the average money spent by a player vs previous months?
How many players bought anything vs prevous months?
Which offers were most sucessfull and which not ? = which were bought most and which were not at all.
on top of that, statistics about gem spending:
Total gems spent by players in last month vs prevoius months
Average gems spend/player vs previous months.
Distrubution of gems spent: chests vs events shops vs daily deals and how it changes over months.
Number of players that logged in at least once in last month vs same metric in previous months = active players.
Number of gems/money spent by players/active players vs previous months
etc.

These are business related statistics that MATTER for company. Collecting data for those, will use up some storage. But hey, they might fit in more data into data werehouse (if they have one)- depends what they want to measure :wink: