World of Warships

Statistics can mislead: a reflection on CV and DD stats from yesterday’s “2+ months of data” post

WorldOfWarships5 - Statistics can mislead: a reflection on CV and DD stats from yesterday's "2+ months of data" post

Yesterday, a post called "2+ months of data: where we were, where we are, and what happens next" mentioned that DD damage has been unaffected by CV rework, and also that CV damage has not increased wildly, except at T6. Before anything, let me say that I don't think it was the author's intention to mislead — he (and I, at the time) made what I suspect is a natural error in the interpretation of the numbers.

At the time, I didn't make much of this statement. It seemed a little counter-intuitive, considering how much more effective CVs feel against DDs since the rework, but I chalked it up to feeling sometimes not mirroring reality. But over the night I realized something important that was missed in that post. Since the post is now long off the front page, I felt it was best to start a new discussion about this.

The most important fact that snuck under the radar in that post is that the CV population has increased by 200% since the rework. This is a huge deal when combined with the observation that damage didn't significantly decrease for CVs in tiers 6-10, and in some cases increased. Here's why.

Generally speaking, you can expect that whoever is best at a class will be playing that class preferentially over other classes they are worse at. So in a steady state, the best DD players will be playing DD, CV players, CV, etc. Of course, this is not a universal. There are some people who love swapping around, and others who are wrong about what they're best at. But by and large, someone who is skilled at a class is going to prefer to play that class.


This has a few implications. One is that the marginal player of a given class will tend to be worse than average at playing that particular class. So if a class' population decreases, it will tend to be relatively less class-skilled players leaving the class. And if the population increases, it will tend to be relatively less class-skilled players joining.

So, if we suppose that from pre- to post-rework, CV strength were flat, but we also observed a 200% population increase, we'd expect average CV performance to go down, since the influx would consist of people who are relatively less skilled at playing CV. But in most cases, CV damage actually went up! This at least suggests that CV power must have increased significantly to compensate for an influx of players who were less skilled at CV.

Correspondingly, the number of DD players has been cut in two, yet performance has only remained steady. That means that it likely now takes much more skill as a DD to have the same performance as before, since the people who quit DD were probably on average less enthusiastic and/or skilled at the class than the average.

So if anyone thought the results of that post around DD and CV performance were counter-intuitive, this might go some way toward explaining the discrepancy between lived experience and the numbers shared there. There are a bunch of ways we could verify this information if we were on WG's data science team, but unfortunately we are not, so the best I can offer is speculation.

Source: Original link

© Post "Statistics can mislead: a reflection on CV and DD stats from yesterday’s “2+ months of data” post" for game World of Warships.

Top 10 Most Anticipated Video Games of 2020

2020 will have something to satisfy classic and modern gamers alike. To be eligible for the list, the game must be confirmed for 2020, or there should be good reason to expect its release in that year. Therefore, upcoming games with a mere announcement and no discernible release date will not be included.

Top 15 NEW Games of 2020 [FIRST HALF]

2020 has a ton to look forward the video gaming world. Here are fifteen games we're looking forward to in the first half of 2020.

You Might Also Like

Leave a Reply

Your email address will not be published. Required fields are marked *