Murder by Numbers

[Kalvarin 2]

I agree with Doktor' date=' this data is going to be really hard to interpret without controlling a few things better. An effect of population size is almost granted, as it increases density which in turn increases the probability of players encounters and therefore the raw number of murders (assuming that the probability to murder/encounter stays constant). Since we know that the players population has recently been increasing sharply, this is an issue in your stats.

[/quote']

No the density is remaining constant. Every server can have a maxium of 50 players, so increasing the number of players playing the game will have no effect on the density. You would be right if we had so few players that the average population of servers was very low, but if you observe the server list the average number of players is close to maximum capacity.

As far as understanding the data there are some problems yes, the main being that we only have the totals from the beginning of the mod to calculate from. So any change caused by a patch may be impossible to see in the data, even if it exists.

[combatcomm1 234]

As these numbers go up, and they will, more mechanics are needed to circumvent this.

http://dayzmod.com/forum/showthread.php?tid=6986 - Information exchange

http://dayzmod.com/forum/showthread.php?tid=7434 - Realistic loot Alternative to class system

[tails13 0]

Tails' date=' you are doing it wrong.

To find a meaningful difference you must first use longitudinal data, from as far back as you can. Only information for two days is not enough.

Maybe Rocket can let people use an API to look at player data for these purposes

To find a statistically significant increase, you need the data for the bandit-survivor ratio and murders per player at the start and end of each major patch. Then you need to do some analyses to see if the Effect Size is significant at all which may indicate a meaningful difference and not just a difference statistically speaking.

This is to ensure that the increase is not just a natural increase in the population as the mod gets more popular, but in people actively trying to kill rather than survive. Further, that the number of people who only want to kill other players is not increasing beyond what was expected as the mod got popular and attracted people with very little inhibition (who just shoot everyone on sight).

If the rate at which % of bandits in the population increases significantly faster than the No. of murders per player, then that may suggest more people are just killing players rather than surviving. However, you have to take into account as time moves on, people just get better at surviving AND killing players.

More information would help though, such as where deaths occur and with what weapons, and if they were by a survivor first or already bandits

[/quote']

I can only use information that I have access to.

Obviously I only really thought of this yesterday in light of bandit skin removal, but it's something I wish to continue to record so that in say, a months time we will have deeper data and can look at comparisons between different builds from this point forward.

As I mentioned before, when rocket implements an effective way to use the humanity score we'll be able to see if that makes a difference in the data.

This isn't just about the bandit skin removal, that's simply what gave me the idea. I suppose it's now more about a mix of that plus observing how the popularity increases will influence the amount of player killing.

For all we know now, we could see it decline again by itself as those playing just to PK get bored over time.

[tails13 0]

*update 29/5 9:53pm*

[tails13 0]

I lost the data I had grabbed from the last ten days when the forum reverted back so I'm done with this. Should have been recording it somewhere else as well as here!

[dallas 5195]

Yea sucks, but you did show murders and bandits increased.

Thanks for the service while it laster.

[zak757 146]

I only have 1 murder, 45 zombie kills, and have been alive for 3 days with a Glock and AK-74 Kobra. I need to stop bushwacking and head to the city :/

[The Killing Joke 43]

I lost the data I had grabbed from the last ten days when the forum reverted back so I'm done with this. Should have been recording it somewhere else as well as here!

That sucks dude. Even though I called some of the data into question, I'm sorry this shit had to happen to you - you were doing pretty good work here.

[tails13 0]

Thanks guys.

The bandit level was increasing, but the amount it was increasing by had decreased by a fair bit.

When I began recording it, it was increasing by about 1.5% a day. Two weeks later it was taking three days to increase by 1%. So it had began to plateau at about 17.5%.

[FragMonkey 1]

post moar data