[UPDATED] Weekly Stats, PvP Trends

[Whattteva 32]

One thing to keep in mind that I'd like to point out about respawning not being counted. I don't know about other players, but when I started playing, I didn't know that you could lose your stuff while swimming (found it hard way). I didn't know there was a respawn choice, so instead of doing that, I ran to the nearest zombie and got myself killed. Just something to keep in mind.

[ZedsDeadBaby 2287]

You did the statistical analysis incorrectly. The numbers you have gotten are a comparison between survival percentage ending in murder WITH a bunch of new players versus survival percentage ending in murder WITHOUT a bunch of new players.

Nah.

First of all, that's not even really true - the game has seen a constant influx of new players 10k+ per day every day for a long time. It is accelerating, but not so much that 7/7 was somehow a different demographic.

Second, how does that make my analysis "incorrect?" I clearly acknowledge the presence of new players and discuss how, if the game were in the state that people claim, this would actually increase the murder and banditry rates, not decrease them.

(some of which is not even counted for - players who have stopped playing).

Uhh, what? Why would they be counted? The analysis is determining what the rates of PvP are in the game. Players who are not playing aren't relevant to that analysis. You seem to just be kind of confused about what claims I've made.

I could easily present the counter argument that your numbers demonstrate that new players die to other sources (bugs, falling, starvation, zombies, suicide, etc) than PvP.

That wouldn't be a counter argument? That's exactly the argument I'm making. You really are confused...

This is further backed up by the fact that new players know less about the game so are more likely to die to things that more experienced players can deal with. Furthermore another reason you need more data is because there things that are hard to explain like the fact that the player base has increased by 38.896% but survival attempts have only increased by 26.386%

Yup. Life expectancy has gone up. So that's pretty much expected. Don't see how it's relevant to the analysis I was doing.

Another small tip: you should try to state your assumptions with regards to what the numbers in the data set represent. (e.g. do survival attempts include currently alive players? dead bandits? and other issues of that nature so that your statistical analysis can be verified)

Whatever it is, we can assume it is being calculated in the same way this week as it was last week, so I don't have to know precisely what that population is to calculate a downward trend in the number, which shows that no matter what the precise meaning, activity is decreasing, not increasing with respect to PvP related deaths.

At worst my "total deaths" value is off, but then it would be off in both tables and the trend would remain, which is the important point - not whether the numbers are 100% accurate but that they are trending in a particular direction.

Edited July 14, 2012 by ZedsDeadBaby

[Whattteva 32]

Doesn't surprise me. I've died maybe 15 times in Day Z, only 3 were at the hands of other players. I tend to avoid the "OMG PVP" debates, because I honestly never saw it as much of a problem. It exists, but it doesn't seem as ubiquitous as some angry threads make it out as.

Good work ZDB.

I think those people are not really complaining about the rampant PvP encounters. Rather, it's the fact that since PvP encounters are somewhat rare, usually, by the time you do encounter one, you'd already be geared pretty well. Obviously, you'd get much much more upset if you get killed with more stuff than if you were just a fresh respawn off the coast. It doesn't take many occurrences of getting killed for probably the majority of people to get mad.

As such, this creates an illusion that it happens often. People tend to give more "weight" to things that negatively affects them than things that affects them positively.

Also, think of it this way. You see obituaries in the newspaper everyday, but none of that affects you. If someone close to you dies, it will affect you much more. It's all about everyone's own subjective perspective really.

Edited July 14, 2012 by Whattteva

[Electrolyte 4]

Thanks for sharing Zed. I imagine significantly more new players die to infected than bandits, at least that's how I got started. There are a lot of ways to avoid dying even after being shot to account for, and I know I've let people die to infected fter knocking them out just for giggles. I can't speak for the latest wave of new players, but something I've noticed a lot of is how terrible most of coastal snipers actually are. It's really appalling to be honest.

[Outlawled (DayZ) 27]

The statistics are skewed by the influx of new players who don't understand the game yet. DayZ has a pretty big learning curve; from ArmA 2's control scheme and counter-intuitive inventory system to DayZ's game mechanics, there's a lot of shit to figure out. Anyone can tell you that they died to zombies/starvation/dehydration more often when they first started playing than they do now.

This isn't accounted for in the statistics.

I realize that this is just conjecture, but I feel like those statistics indicate that there is actually more PvP now. There's a 39% increase in players, yet there's only a 1.44% increase in deaths unrelated to PvP. If that 39% was made up of seasoned players, then that 1.44% increase might mean something, but it's not. That 39% is made up of brand new players who don't understand where to find guns, how to avoid zombies, or the fact that you don't actually need a gun to survive against zombies. As soon as those players understand how not to die to zombies, the percentage of deaths unrelated to PvP will decrease greatly. Of course, at that point, more players will have come in and skewed the statistics with their lack of understanding.

The only way to get any real trends out of the data is to get rid of some of the variables. Take a group of players, e.g. all the active players at this moment who are on 1.7.2.3, and track their statistics separately from all the new players that join every day.

Right now there are too many variables and not enough points of data.

You can't just look at a couple individual statistics over the course of a week and then claim that PvP is on the decline, ZedsDeadBaby.

If you could, then I'd probably say that PvP is on the rise because of these statistics which I pulled from your original post:

July 7th -

Zombie kills / murder = 60.778

Zombie kills / (murder + bandit kill) = 51.948

Murders / bandit kill = 5.883

July 14th -

Zombie kills / murder = 59.485 (2.17% decrease)

Zombie kills / (murder + bandit kill) = 51.370 (1.13% decrease)

Murders / bandit kill = 6.330 (7.60% increase)

Edited July 14, 2012 by Outlawled

[funionz 7]

It's a hopeless cause ZDB, these are the same people who will look at Ander's statistics and say they are somehow incorrect

[grizzle 21]

"There are three kinds of lies: lies, damned lies and statistics." - Mark Twain

[FaxMonkey 10]

Nice to see somebody attempting some statistics but it's not a good enough analysis to draw any conclusions yet.

You can't draw trend conclusions from changes in approximately 1%. Combined with only two dates in the trend, it doesn't really mean anything and the changes could be due to random events.

You mentioned all the new players starting up each day. A good chunk of these new players are still figring out how to avoid Zombie's and get weapons let alone start PvPing.

You've created a trend analysis over a period that in includes major patch updates. The changes in the patches are probably enough to totally stuff your figures.

This would need more analysis over months before you can draw any conclusions.

How will you account for all the other external factors such as eg.

'number of deaths due to bugs'

'number of suicides by running into Zeds' (it took me a few weeks to figure out there is a respawn button)

'average time for new players to learn the system' (and start safely finding weapons and PvPing)

'number of players that are continuing to play'

etc

[funionz 7]

"There are three kinds of lies: lies, damned lies and statistics." - Mark Twain

Lol good quote, but seriously, the guy who manages all the data in the database can post facts and these people will tell him he's wrong or the numbers are skewed somehow, it's pointless.

Also, ITT : People who took a semester of statistics trying to sound smart

Edited July 14, 2012 by Aaron

[lev 39]

Nah.

First of all, that's not even really true - the game has seen a constant influx of new players 10k+ per day every day for a long time. It is accelerating, but not so much that 7/7 was somehow a different demographic.

Second, how does that make my analysis "incorrect?" I clearly acknowledge the presence of new players and discuss how, if the game were in the state that people claim, this would actually increase the murder and banditry rates, not decrease them.

The problem here is that your analysis was performed in such a way that the data assumes a massive influx. As I said earlier you need to perform a rate analysis at constant periods to accommodate for increasing number of players, not a 2 point analysis which is too general to claim a trend. I am also not arguing for or against whether banditry rates are in fact increasing or not because the figures are useless IMO since they don't present enough detail to discern any actual information.

Uhh, what? Why would they be counted? The analysis is determining what the rates of PvP are in the game. Players who are not playing aren't relevant to that analysis. You seem to just be kind of confused about what claims I've made.

They have been counted because the figures posted on the main page are cumulative and not figures for the currently playing population. Anyone who has ever tried DayZ and quit or are currently taking breaks from it still have their lives and deaths figured into the main page's stats. That is part of the reason why I think the main page's stats are not that useful unless you collect them over a period of time for rate analysis.

That wouldn't be a counter argument? That's exactly the argument I'm making. You really are confused...

I should have been more clear in indicating they might be dying too soon to experience PvP and/or PvP incidents might not be detailed by the statistics. This doesn't necessarily mean that the rate of PvP has decreased but could just indicate that new players are dying to things more experienced players would not die from and are not living long enough to experience PvP.

Yup. Life expectancy has gone up. So that's pretty much expected. Don't see how it's relevant to the analysis I was doing.

I made a mistake with these stats because I realized afterwards they reflected the problem of including the total population's numbers and not just the numbers of currently active players (by active I mean playing regularly and not taking a long break or having quit DayZ). I removed these in a later edit.

Whatever it is, we can assume it is being calculated in the same way this week as it was last week, so I don't have to know precisely what that population is to calculate a downward trend in the number, which shows that no matter what the precise meaning, activity is decreasing, not increasing with respect to PvP related deaths.

At worst my "total deaths" value is off, but then it would be off in both tables and the trend would remain, which is the important point - not whether the numbers are 100% accurate but that they are trending in a particular direction.

For verification purposes it doesn't matter that you don't know as long as you state the assumptions you were working with. That's how it is done in research so that others can follow along and either A) point out mistakes in assumptions or B) understand how you are getting to the numbers you are getting to.

As I said earlier, the only conclusion we can validly conclude from the numbers you presented are: "The numbers you have gotten are a comparison between survival percentage ending in murder WITH a bunch of new players versus survival percentage ending in murder WITHOUT a bunch of new players." AKA the data just isn't there to validly draw further conclusions. To accurately give an idea of PvP you would need a lot more detail that just instant present in the broad sweeping figures present on the main page. Not to mention I'm pretty sure bandit state is still as convoluted as it was before meaning people could lose bandit state over time which makes it even harder to figure out what is actually happening. With the generalized numbers on the main page, its pretty hard to obtain a figure representing the true occurrence of PvP incidents with a reasonable confidence interval. Something more useful would be, players active in the past 24h, murders in the past 24h, survival attempts in the past 24h, players taking (hit or not hit) fire in the past 24h. That kind of data collected over the period of at least a month or so would help determine the actual stats versus some back of the envelope calculations.

[FaxMonkey 10]

Obviously not since the "last players per 24h" stat has been going up, not down.

Maybe I'm confused but 'Players in last 24 hours' is lower on the 14 July (107,370) than the 7th of July (112,337).

This ads weight to z-layrex's theory that people have stopped playing the game properly for any length of time because of the game ruining update.

It also adds weight to the argument that this isn't enough analysis to draw any trends/conclusions from.

[lev 39]

Lol good quote, but seriously, the guy who manages all the data in the database can post facts and these people will tell him he's wrong or the numbers are skewed somehow, it's pointless.

Also, ITT : People who took a semester of statistics trying to sound smart

Don't really need a lot of statistics to realize that 2 data points are not enough to draw a generalized conclusion. If we really wanted to get into detail, we could probably even question how the database entries are obtained and which SQL queries are actually generating these stats.

Also, ITT and forum in general: too many people with an anti-intellectual mindset.

[dallas 5195]

Does this mean I can expect not to get KOS by everyone I meet?

[QKTunak 0]

Betting it also is not counting people who die after ALT-F4ing and trying to log into a different server for help

[sandorski 17]

I suspect there's something wrong with how Murders stats are being collected. Of my 8 or 9 lives, only 4(I Hack, 2 Zs, 1 unknown spawned on beach as new without dying the previous session)were not PKs. That said, I suppose Noobs might be tipping the stats through simple mistakes and multiple repeats of those mistakes. Might be more informative to see further breakdowns, based upon the amount of playtime the player has, where these deaths occur, how long after Spawn, etc.

[orb917@hotmail.com 117]

Maybe I'm confused but 'Players in last 24 hours' is lower on the 14 July (107,370) than the 7th of July (112,337).

This ads weight to z-layrex's theory that people have stopped playing the game properly for any length of time because of the game ruining update.

It also adds weight to the argument that this isn't enough analysis to draw any trends/conclusions from.

this was a good argument, so i checked the front page just now and it's at 134,668 in the past 24, so it has increased.

[idunc 62]

These stats actually reflect real life. As the population goes up, the crime rate goes down.

[FaxMonkey 10]

this was a good argument, so i checked the front page just now and it's at 134,668 in the past 24, so it has increased.

The last 24 hours is probably up because of the last patch that was released a day ago. This patch also seems to have gone some way to fixing a few issues.

More evidence that one week worth of data is no where near enough.

[LIMITID 28]

Interesting though not unexpected. Alos of the perception of PVP playing a huge part in the game comes from people rushing to the forums to rage and scream about how they where killed by another player. Also, after playing the mod for over two months now, I am yet to see these mythcal bandits who camp the coastlines and cities waiting for noobs. Less then a quarter of my deaths have been due to PVP, and I have spent a fair amount of time in the cities, most people who see you just try to avoid you.

[GodOfGrain 191]

Don't forget the DC from PvP. At least 30% of (potential) murders are not in the stats due to DC.

[ZedsDeadBaby 2287]

Maybe I'm confused but 'Players in last 24 hours' is lower on the 14 July (107,370) than the 7th of July (112,337).

You're not confused, I just took the stats at 10AM on the 14th whereas I took them at 3PM on the 7th. That's why I didn't make any mention of it in my analysis (as it wasn't really relevant at the time).

Go look at the 24h stat now. See?. 136k and climbing. And this 24h cycle isn't over yet.

Neat, huh?

The last 24 hours is probably up because of the last patch that was released a day ago. This patch also seems to have gone some way to fixing a few issues.

More evidence that one week worth of data is no where near enough.

It has gone up pretty much every day between the 7th and today, so yeah. I mean attribute it to whatever you want if it makes you feel better. It's not the patch.

30% of (potential) murders are not in the stats due to DC.

30%? Stats you pull out of your ass are so much stinkier than ones that are based on actual numbers.

Edited July 15, 2012 by ZedsDeadBaby

[Yeung-Jin 28]

Don't forget the DC from PvP. At least 30% of (potential) murders are not in the stats due to DC.

That's a good point. I rarely run into DCer's for whatever reason, the only time I'm reminded of them (or "friendlies") is when I come to the forum

[The Poacher 31]

The very low percentage of deaths due to PVP is very surprising to me. I'd say 9 times out of 10 when I die it's from being shot. And on average when I die I usually have between 5 and 15 murders.

Basically this. Though I usually die to my own stupidity (fighting zombies for too long, no morphine/bandages, obvious avoidable glitches), rather than players, unless I was being stupid and got spotted by a player.

Edited July 15, 2012 by Lincolnshire Poacher

[thorgold 91]

hmmm,.... I don't believe this. I still think PvP is what causes most deaths

And I think the sun is square, but that doesn't change the empirical data.

[Phatd00d 1]

Is this counting the respawns trying to get back closet to your dead body or favorite spot on the map?