I've already asked this question in the Newcomers' Forum but it didn't lead to any results. So I'll try again here and hope that some person actually knows about the values that WN8 is based on. XVM is used by many thousands of people, so there has to be someone knowing, right?

This is going to be very detailed below but the actual question is just one line, that's all.

- The values are updated daily here on the modxvm.com website.

- I assume the values are global as there is no statement made to which region they would belong. I guess they are obtained with the help of the XVM modification. So their would be no reason to differentiate between mod users from different countries.

- WN8 is based on the expected tank values. They are needed to calculate WN8 together with players' individual statistics.

My question: How do the expected values come about? How are these values collected? How are the expected values calculated from other values (as you can't find them in their pure form in the game)?

- I know what an expected values is in mathematics (stochastics) but it seems to me that this doesn't fit here.

Details:

Although stochastics and statistics are closely related in stochastics you need a random variable to calculate an expected value and you need probability values for the different values the variable can take. If your random variable would in the simplest example be "winning a game" that can take two values: "win" or "no win (= draw or lose)" and you would use empirical probability values, this would just lead to and average value. In this case the expected value calculation would be the same as the average value calculation if I'm not mistaken. At least I understand it this way. But way of calculating the expected values can't be the right way because in reality they don't match the actual average values.

- In regards of World of Tanks it might be the obvious solution that the values are simply average values. This is what I first thought about and what I was given as an answer on my previous topic. But this is not true as I'm going to proof below (premised the information on my source websites is correct).

(My research in detail: )

For my research I've used this website as another source of information: https://wotlabs.net

- Comparing the values on this website to the values modxvm.com offers they are basically the same. The tiny differences are probably down to the difference in time. I assume wotlabs.net simply uses the table provided by modxvm.com as its source for the expected values. F.e. comparing the DMG values of a tier X tank (Leopard 1): 1875.128 - 1874.131 = 0.905 DMG difference (modxvm.com first, wotlabs.net second). That's a deviation of 0.905 / 1875.128 * 100 ≈ 0.05%

- Comparing the values on this website to the expected values from the modxvm.com website as both tables are updated daily: The values are different. F.e. comparing the DMG values of a tier X tank (Leopard 1): 1875.128 - 1862 ≈ 13 DMG difference (modxvm.com first, wotlabs.net/eu second). Now this example values seem very close to each other but the deviation is still 10 times higher than it was when comparing the expected values from modxvm.com and wotlabs.net. Another example: Comparing the DMG values of a tier V unfair artillery (leFH18B2): 662.886 - 586 ≈ 77 DMG difference. That's a deviation of 77 / 662.886 *100 ≈ 12% which in this case is not a 10 times higher but a ≈230 times higher deviation than it was comparing the expected values from modxvm.com and wotlabs.net. So you can see the difference varies a lot and I didn't look up more than these two examples so there are probably even more extreme examples.

No these examples were simplified. I'm trying to compare the expected values to average values. To be exact the global expected values and the global average values. But in my examples I didn't use the global average values, I only used the EU average values as I said "(modxvm.com first, wotlabs.net/eu second)". This is because I can't obtain the global average DMG values over all battles played. Because to do so I would need to know over how many battles the EU DMG value was averaged and over how many battles the NA DMG value was averaged as they are unlikely to be the same. There are far more players playing far more battles in EU as in NA. So we can't know the exact weight of each of that regional average values.

But there is one thing we know: The global average has to be in between all the regional average values. And comparing the highest average DMG value for leFH... from EU, NA, RU and Asia to the global expected value it is still way below the expected value. So this means expected values are indeed different from average values.

This confirms that expected values are not simply average values. Still premised that my source of information isn't wrong and titles the values they give correctly. But I think this result is likely to be true because

1. why would there be two different tables for the same thing if expected values were just the same as average values?

2. why would there be two different names for the same thing if expected values were just the same as average values?

So I hope someone can help me and explain to me what expected values truly are?

Greetings

schw4rz3r_Engel

Edited by schw4rz3r_Engel, 24 March 2020 - 11:21 PM.