3,100 out of 22,500 is.
Pretty lame trying to pick and choose one little instance to try vainly to prove a non-existent point.
what else are they making harder for new players? How many hours it is taking now?
- Thread starter Rockets
- Start date
- Status
Pretty lame trying to pick and choose one little instance to try vainly to prove a non-existent point.
No. a 14% sample is also huge for the population.
There can be no reasoning with someone who doesn't even acknowledge the basic facts of the topic they are discussing.
![[Diplomacy] Carnegie](https://forums-cdn.zaonce.net/data/avatars/m/123/123809.jpg?1553527149){kind=avatar}
You just need to wait for the Steam charts to start showing positive numbers - then they suddenly become significant and meaningful again
I'm not even sure how to respond to this.You just need to wait for the Steam charts to start showing positive numbers - then they suddenly become significant and meaningful again
I was under the impression that the steam charts show that the amount of on steam players has remained stable (with some ebb and flow) since the game started selling on steam.
That's what the numbers show.
But apparently you're never justified in drawing even the tiniest of conclusions from that because people want to think they're special because they use the Frontier launcher.
But apparently you're never justified in drawing even the tiniest of conclusions from that because people want to think they're special because they use the Frontier launcher.
But it's not adjusted for bias. If I survey 1000 people out of 10,000 and the 1,000 people I survey are, say, attending a church service I would get numbers that massively overestimate the proportion of religious people in the entire population if I project from that. For instance I "never", and I mean never buy games on Steam, because if I am going to be playing a game then the I think makers of the game deserve all the money from the sale of that game even if it costs me more. Steam is a "demographic" and what you are getting from Steam numbers, unless they are adjusted for bias, is a survey of Steam players and nothing more. If you deliberately select a subgroup for your survey and exclude everyone else, and then don't adjust for bias then all you get is data related to that subgroup and that's not projectible on to the general population.
I refer you to my previous posts. 3800 is 40% of 9500. The chances of that sample being biased are minimal. Even 14%, of 22000 is still very unlikely to be a biased sample because it's such a large proportion of the population.
Validity is an argument, not a fact.
Steam is a demographic; I have previously suggested that it's more likely that Steam is the majority demographic of ED players, not launcher users, and it behoves those who would suggest otherwise to support that claim with reasoning and evidence. Moreover, you have completely ignored the argument I was making (I refer you again to my previous posts) about the context that I used the data in:
OP said the game was dying. The evidence from what is (likely) the largest player pool suggests otherwise.
You have used no bias adjustments, therefore you can't argue the Steam user base isn't biased. For instance regular discounts on Steam may in fact attract a large number of players who want to play an ALT, either as a FC refueler or bubble mission runner or explorer. Unless you ask the question you can't eliminate that bias, this is why there are some fairly strong guidelines on polling, questions need to be asked to eliminate that bias and take it into account. You would ask whether this was an only account, an ALT of a primary or this a primary with other ALT's. It's very likely Steam has a high proportion of ALT's due to these discounts, so you can't count the steam charts as a total of individual players. You can't ask these question of query the Steam player base to eliminate these biases therefore you can't say the Steam users can be used as a generic sample.
I'll also point out that a person making the claim has the burden of proof, saying people disagreeing with you must then present evidence, while you have presented none in support of your position is called the reversal of burden of proof fallacy.
Bias adjustments aren't necessary here - I haven't used Steam charts as a count of total individual players, hence alt accounts and other artefacts in the data are irrelevant.
I have given evidence - the trend of conurrent logins since 2016 suggests that the numbers of concurrent players has changed very little in that time, outside of peaks after major updates. The response to this was "you can't say that; Steamcharts are meaningless" which is not only absurd, but hinges on a spurious (and unsupported) claim - that launcher users or Epic users are somehow different, along with the unspoken assumption that they are the population, when in fact it is more likely that Steam users are the majority. It is thus not a fallacy for me to suggest that evidence is required for that claim.
With regard to alts, it's possible, but I'd suggest that the vast majority of players don't use them at the same time, and if they do, it won't show in Steamcharts as artificially inflating the numbers anyway, because it's actually quite difficult to get Steam to let you play two games concurrently.
Heads you win, tails everybody else loses.
Your entire argument is based on the idea that the few mean more than the many. It's why as I said before polls are almost always wrong because they assume a minority sample is representative of a majority. Almost always that is wrong.
Opinion polls mostly suffer from the fact that they involve humans answering questions. Humans are notoriously bad at honesty, accuracy and consistency which tends to skew the results. Statistics based on their behavior rather than than their words are generally much more accurate. This is how the National Grid knows to have more power available for making cups of tea when the Corrie ad break comes on.
I thought people were uncertain of the relevance of the Steam data as they don’t know what percentage of the total player base it is and if that percentage is changing.
Precisely.
I've personally known people that lied on an exit poll just to mess with the poll takers.
The fact of the matter is that regardless of if you're on Steam (which I am) or you're direct from Frontier, they have the data on who is connected to the server. They also have any complaints mailed them and they know what the ENTIRE player base thinks, how many players their are and any and all trends thereof.
It's a pointless and more than a bit stupid an argument to make that for some strange reason the minority of people playing from Steam mean more, show more or anything for that matter, than anybody else when the fact is no matter where you're playing from you're connecting to Frontier's servers, using Frontier's systems and filing any tech support or complaints about the game to Frontier.
The only numbers that truly have meaning are:
As long as the first number is greater than the second, we’ll continue to see ED being developed. And only Frontier is privy to those numbers. Steam player concurrency doesn’t give us a window into those numbers. It doesn’t tell us whether the Steam player base, as a cohort, is more or less likely to buy cosmetics or expansions compared to other cohorts. It doesn’t tell us whether players who regularly log in 40 hours a week is more or less likely to buy extras, compared to those who play three hours a week. It doesn’t even give us the weekly average and mean time spent playing per week, or the length of said sessions.
- how much money Elite Dangerous bringing in for Frontier
- how much money continued development of ED costs Frontier
As long as the first number is greater than the second, we’ll continue to see ED being developed. And only Frontier is privy to those numbers. Steam player concurrency doesn’t give us a window into those numbers. It doesn’t tell us whether the Steam player base, as a cohort, is more or less likely to buy cosmetics or expansions compared to other cohorts. It doesn’t tell us whether players who regularly log in 40 hours a week is more or less likely to buy extras, compared to those who play three hours a week. It doesn’t even give us the weekly average and mean time spent playing per week, or the length of said sessions.
Didn't the Epic giveaway hand out 8 million copies for free (according to a comment by Frontier on this very forum) which pushed to total bought, on all platforms, to around 12 million?
So, if Epic players are double the previous existing copies sold, all by itself, perhaps the steam figures should be tripled, if they are assumed to be a 'trend'?
(as we know, Epic do not release player figures, so using the 'tried & trusted' steam demographic should suffice to extrapolate)
So, if Epic players are double the previous existing copies sold, all by itself, perhaps the steam figures should be tripled, if they are assumed to be a 'trend'?
(as we know, Epic do not release player figures, so using the 'tried & trusted' steam demographic should suffice to extrapolate)
Fortunately nobody is making the argument that Steam users 'mean more' - quite the opposite in fact.
Steam users can be considered to be reasonably representative of the playerbase as a whole - with the caveat regarding the very long-term players which I detailed earlier in this thread.
Update 12 live steam announces real piracy aka Not smuggling, not trading in illegal goods, but taking goods by force from NPC cargo rack of ship will now pay more, So easier to earn way more from stolen LTD and platinum in your starter cobra mk3 as the 25% cut the black market take will be reduced...
You are still required to research Aka use " youtube " to find the correct systems to pirate in... but after you have done the research you will now get paid more for taking a NPC cargo...
You are still required to research Aka use " youtube " to find the correct systems to pirate in... but after you have done the research you will now get paid more for taking a NPC cargo...
Last edited:
Oh please.
Opinion polls are not "almost always wrong" - with most opinion polls that fail to predict the actual result, the results are within the margin of error (directly related to the confidence interval). Uncertainty is built into statistical modelling. For instance, the predictions for the Brexit vote were "wrong", but within the margin of error. It's signal of how well populations can be modelled that 3% margins can be so magnificently wrong.
The few do not mean more than the many. I didn't say that. Please refer back to my point about qualitative versus quantitative data earlier - numbers only tell you so much; talking to people reveals a lot more stuff that also happens to be a lot more interesting, at least to me. It's very difficult to generalise from qualitative data, although those that invest the (significant!) time into doing qualitative research alongside the quantitative generally benefit from it.
Please, go to the supermarket and buy a big bag of Skittles (or M&Ms, if you prefer chocolate), and count out 100 of them, then divide them into two piles of 47, separating six. Consider what more you'd need to know about them on a divisive issue such as Brexit, or Scottish independence (coming soon. Thanks Brexit!) in order to be able to accurately predict their responses to questions on that topic. That's three people out of fifty, and you're trying to tell me that I'm wrong about the few meaning more than the many?
My point is that you can make predictions about surprisingly large numbers of people from surprisingly few people if you ask the right questions of the data. I'm not saying that the opinions of three people are more relevant than the opinions of 3,000. What I am saying is that the behaviours of 300 people can tell you a lot about the behaviours of 3,000,000, provided (again) that you are asking the right questions and are aware that there is a margin of error built into the model.
If you're trying to say that "update 6 was bad" because only 5,000 people play the game according to Steamcharts," then yeah - that's just factually incorrect and an unwarranted mixing of qualitative ("bad") and quantitative ("5000") data. If you want to suggest that it's not logical to conclude that the game is "dying" (a charged assertion attempting to make a quantitative statement less amenable to analysis by layering it with a loaded qualitative term) because the pattern of the number of concurrent players on Steam has changed little in six years, then I don't think that's an irrelevant point. An absolute statement like "the game is dying" is easily falsified, in my opinion, with numbers like that.
See my point above about error margins and confidence intervals.
No one has ever said that Frontier don't know how many people play their game. The point has been (and will continue to be) that in the absence of any official data from the developers (and no one would seriously expect FD to just release that data for no reason - and internet arguments are not a good reason), Steam numbers are a pretty good proxy for some minimal inferences. For instance, given that people have been prognosticating dooooom for Elite: Dangerous since 2014, the concurrent player stats (available since 2016) show that to be just untrue - the data about concurrent player numbers (note that I'm not trying to infer total player numbers or unique logins, or what people feel about the game) has been remarkably stable. The game might be dying, but for a dying game it's really sticky. At the very least, these stats are inconvenient for the forum's Cassandras. Maximally interpreted (never recommended, although it's common to talk in absolutes about the "death" of a game. It's almost like people who use the word in that sense don't know what they're talking about), they absolutely refute the claims.
I haven't seen that post - do you think you'd be able to find it again?
Finally! Some bloody nuance!
Last edited:
I'm going to take issue with this, on reflection.
The university I work for took years to work out a template for their VLE that both enjoyed a sane* workflow for the teaching staff and didn't confuse the hell out of the students. This is a serious institution, one that takes UX (and the research behind it) seriously, and has user numbers in the hundreds of thousands.
That the data is there doesn't mean that the owner of that data is even aware of what stories it might tell; it just means that that data is there to be interrogated.
In all seriousness, FD probably don't know what you think they know; they don't even have a fraction of the staff they'd need in order to be able to uncover all but the most pressing issues. I'm sure the staff they have in the UX role are very good at their jobs, but there are likely only a handful of them working on Elite: Dangerous.
* Well, semi-sane, I guess. You're always dependent on the sanity of others.
Last edited:
- Status