Assembly Hall

Hell why not, its a game right. Besides if it works with random pickup trucks and station wagons in the middle east with 50cals and anti-air cannons welded to them it wouldn't be much different. If the flat bed trailers can haul a tank they can have a tank chassis built on to it with a working turret and other guns. The first armored vehicles to be used in war were basically civilian vehicles with armor and guns welded onto them. Hell there are current Black ops vehicles that are civilian SUVs with armor and guns attached. Why not a big rig? Russia and most of Europe did it with rail cars and trains during World War I, and II. Pirates would take Merchant Vessels and put cannons, flame throwers on them to attack other merchant Vessels. Ahh.....there's something to think about....

All those "Pirate" faction ships would be total crap if we were to use Real world Pirate examples lol. Using movies as basis for game balance is far better than real life.....

small arms fire like from light small turrets on a destroyer....hrmm....interesting....if pirates would run why is it that pirates namely NPCs don't do that? most of the time they chill out and let you blow them up. The AI is pretty random on having them run away to fight another day. Same goes for Players, again if we wanted to balance ships based on real life than there would be very big changes that people would like or "THINK" looked cool. Again its a game, why not.

Okay, so you want your station wagon with a .50 cal to be as combat effective as a vehicle designed specifically for combat.

No.

And. This. Is. A. Video. Game. Not. The. Real. World.

And an armed merchant vessel is fine for attacking other merchant vessels...not so much for an actual...you know...naval vessel.

And yeah, it is a game, but it should still be balanced. You are advocating for imbalance.

The thing that would most help CCP get and retain people is to get someone in from Statistics Iceland to teach them how to do a survey and how to objectively examine the results.

The survey of 8000 people, how did they determine these 8000. Was it a random sample? Chosen via some other means etc..

The output data, less than 1% answered ganking. How many didn't answer? How many said things that might incorporate this such as game mechanics or your game sux? What was the primary answer these 8000 gave for leaving?

The better non-NPE questions CCP would already have the data or should have the data for are:

How many people leave within a few days of a gank? Do people know how to tank a ship? How many are solo. How many in small corps, how many in Large?

What is the profile of the customer likely to spend the most? What do these people want? Do they PvP/PvE? Are they casual players or hardcore? Never assume anything.

However given the levels of bias in some areas of CCP they are probably better handing the data and a wad of cash to Statistics Iceland and asking for an analysis.

The other things they need to do is become part of the 21st century where overt forum moderation, like deletion of social media comments is considered a black mark against a company and the other is the same problem they always had and that was world class horrible customer service.

And remember Statisticians do it by the numbers

Actually I did Play Dust514 and I believe it is still installed on my PS3. You are talking about two different games, also two different forms of combat. I was using a .50cal as an example to illustrate that it is possible to tack weapons on to a vehicle that would normally not have the "hard" points for those weapons.

small arms when it comes to a ship like naval ships would be a 5in deck gun like you see on the front of a destroyer which is about the same size as as your 125mm rail guns, or the same size as a modern day M1's 125mm smooth bore. The larger guns like the 8 and 10in main guns on cruisers or the big bad 16s on the mighty mo. are your larger weapons. Due to the armor thickness on Cruisers, battle cruisers, and battleships. The Guns on frigates and destroyers wouldn't do much except kill the people that were exposed on the decks of the large vessels. Most of the time the Frigates and destroyers were used as small escort vessels for convoys, or as over the horizon scouts for the big boys. This would give the large guns a visual spotting aid. Again this is WWII era naval vessels if you want to talk colonial we can, or modern. Every naval vessel has its pros and cons that could be analyzed and help balance this game with its futuristic weapons.

In regards to Q-ships they were extremely effective against the Nazi U boats. A torpedo cost a lot of money and they only had a few on board. When attacking a merchant vessel they would first surface and use the deck gun on the U-boat. This is why a lot of WWII subs had a deck gun. The Q-ship would mount frigate/Destroyer size weapons like a 5 inch gun usually 1 to 3 turrets worth that the would cover with mock truck cut outs. So that they would appear to be a cargo vessel. When the U-boat would surface they would move the covers out of the way and blow the pour bastards out of the water. This is why modern subs no longer have surface weaponry. When the subs stop coming up the use for a Q-ship went away. If you are ever in the Chicago land area hit up the Museum of Science and Industry They have a captured U-boat U-505 which you can walk through and tour as well as lots more information about naval combat. If you are on the east cost there is the USS intrepid which has more naval air history than naval combat engagement history. I believe there is something similar on the west coast as well.

Weather we use Real world to balance, or movies we can all agree that the balancing does need fixing when it comes to ships, overall combat in the game etc. The Industrial vessels need some serious work, as do many other types and classes of vessels in this game.

Look a person who does not understand statistics complaining about others doing statistics.

No politicians and media organisations cherry pick the statistics that they want and ignore the reasons they occured or ignore the more important statistics that don't tell the story that they want to. They used to complain that people didn't answer like they were supposed to now they just assume most people are idiots and they are right.

National accounts are a great example of this. Where politicians and the media tell people how well things are going by pointing to growth in GDP figures. Even though that growth has only occured because the economy is in such a crappy state that no one can afford to buy imported goods. So the GDP goes up because even though exports have fallen, imports have fallen even more and that imbalance causes a rise in GDP even though the internal economy is flat or going backwards.

Yeah, well I don't think so. You see, in your hubris, you fail to realize that what you are talking about is not science, but scientism (and in case you don't know it, it is a bad thing).

It is hard to know what effect ganking has on players, the player base and new players because we do not have the counter factual. That is we cannot do the following:

Have player Joe log in and get suicide ganked on his 11th day in game.
Rewind the clock and have Joe not get suicide ganked in his first 15 days.

(The 15 days was what was used in that presentation by CCP at fanfest).

In essence we cannot to a controlled experiment. At best, we can go for a "natural experiment". Which is kind of what CCP tried to do.

They took 80,000 individuals--i.e. not accounts, but subscribers so as to avoid problems with alt accounts. Then they looked to see how many suffered a ship lost in their first 15 days. Of the 80,000

68,400 were not killed in their first 15 days.
10,800 were killed legally--i.e. their killer was not in turn scorched by CONCORD
800 were ganked.

Of those three groups the ones who stayed with the game the longest were (in order):

1. Those ganked,
2. Those killed legally,
3. Those we did not die in their first 15 days.

In addition, the "fun fact" of less than 1% of those leaving the game cited ship loss as the reason. In listening to the presentation it was not less than 1% of the 80,000, but less than 1% of everyone leaving the game. So your 8,000 players is really...well nothing.

Now, is it possible that there was some confounding factor at play here? Sure. Maybe those 800 players who were ganked were all in really good player corps....which probably helps retention. And maybe the same is true for those killed legally too. But given the numbers that strikes me as unlikely, especially if they picked a random sample of 80,000 subscribers.

And the reason you pick such a large sample is so that when you do start breaking that sample down into sub-groups you will typically have enough so that you won't have a problem with a small sample. For example, if CCP had sampled just 1,000 subscribers they might end up with 10 or 11 guys who were ganked. With such a small number then the "good corp" problem (for example) could ruin your results. But when looking at 80,000 with almost 12,000 of them having been killed in their first 15 days you probably don't have to worry about that kind of a problem. Of course, CCP could have used a stratified sampling approach, where they go in and look for subscribers in each category and randomly pick say 5,000 each.

And yeah, there are additional factors that can come into play as well that could influence these results. Age, marital status, job, etc. could all be factors. But again, with such a large number of players who were killed and ganked those effects are likely to be washed out. Also, CCP does not have much of this data. They might be able to get age, but job and marital status? Nope, probably not...at least not from what they get when a person signs up from the game.

As for the rest of you questions like know how to tank a ship, that would require an extensive bit of surveying done when people quite....which would likely be quite costly and you always have the issue of those who simply refuse, is there something about those players that are biasing your results.

Again, that <1% was for all subscribers who have left the game. Not the 80,000 players CCP used for that presentation.

So you going on about 8,000 highlights, that indeed, you do need to watch it again and refresh your memory.

As for the 80,000...well I'm pretty sure CCP knows that picking subscribers randomly would be the best way. That is really basic stuff. Considering they probably have had well over 1,000,000 cumulative subscribers.....

And collecting demographic data is not costless. Would it be awesome to have it? Sure. But it comes with costs. First of all you have to risk losing subs if you make such information mandatory. If you don't then that whole issue you were worried about with regards to "how the 8,000 were selected" pops back up--that is responders vs. non-responders vs. outright liars. And if you hire a firm to do it, that is an obvious and direct cost too.

I work for a company that has lots of customers and we don't go out collecting that kind of data. And we are a utility...some people might think we'd have that data. How many people in the house, what kind of appliances, etc. But we typically don't have that data. If we want it we have to go to a vendor who has collected some of that data.

Maybe you should consider thinking like a Bayesian.....