Upcoming BR changes.

[Stubbz]

https://forum.warthunder.com/index.php?/topic/475670-planned-economy-and-battle-rating-changes-28012020/&ct=1580242594

Thoughts?

[jaxjace]

引用自 Sobolewski：

引用自 Mirage：

Regardless of who uses it, it clearly shows hundreds of thousands of battles being logged, we're supposed to dismiss the results of that, and take your word for something instead? That's a massive sample size..

Okay...

Today's most popular tank for tank RB battles (according to TS) was the M18 GMC. As of the time of writing, it has 48 550 (forty eight thousand, five hundred and fifty) battles logged "since last month". As of the time of writing,501 RB battles were logged today on it.

As of the time of writing, there are a little over 850 (eight hundred and fifty) battles occurring simulteaneously. The number tends to fluctuate between 500 and 1000 in my experience. Let's average it at 800.

Also, in my experience, the usual match from start to end lasts around 20 (twenty minutes). Some take 10 or 15, some take 30, a few take over 40. So lets put a safety margin in case I'm wrong and establish them at 30 minutes average.

If we make the assumption that they all happen strictly in simultaneous cycles of 30 minutes (in order to be as generous as possible to your theory that TS has a massive sample size), so we have 48 cycles a day, each with 800 matches. 38.400 matches a day. Wow, it's about 1.3 day worth of play in a 30-day period, right? Not exactly.

The 501 battles "recorded" today don't make up even a single cycle. Now let's also take into account that we have, on average, WAY more than a single m18 hellcat in every 4.7-6.7 match. Up to four in the 4.7-5.7 range and all the way up to sixteen in the 5.7-6.7 range. And Thunderskill collects data of individual players, not directly from every match that happens. And thus one match often ends up becoming logged as multiple matches in the TS database.

If we establish an arbitrary (and actually probably on the lower end) average of 4 M18s per match, suddenly we only have around 125,2 battles recorded today. For the most popular tank in tank RB. Out ot 38.400 possible battles (in a very ideally stable environment that doesn't correspond to the chaotic ending and beginning of matches in a minute-by-minute basis that means we actually have a lot more than 38.400 battles going on every day).

That's a 0,3% sample size day-by-day. In this assumption of what War Thunder matchmaking is that is incredibly favourable towards your claims about Thunderskill. Then we must remember also that countless other tanks share a BR range with the M18 and thus countless other "battles" logged in Thunderskill might just be actually a single battle where the M18 and these other vehicles either fought togheter or faced off.

Considering the extremely complex nature of the interaction of tanks in matchmaking (each tank can see dozens of different tanks simultaneously in the matchmaking, with thousands upon thousands of possibilities of how a match can come out in terms of tanks, maps and map modes), this is a field of statistics akin to that of an economy. In Economics, strong inferrences about the economic situation of a country and the measures (economic interventions or incetives) to be taken in order to ensure economic progress often require sample sizes in the order of 25% to 35%, taken by national agencies with 1% or less error margins. Even if we cut it to half (12.5% to 17.5% sample size and 2% or less error margins), it still doesn't come even close.

So, I'm afraid you have a lot of disproving to do now if you still want to protect that claim of yours.

All I got from this wall of text was that the m18 is the most played vehicle in the game, gee colour me suprised. its americas favorite clubbing tank. Its about as Meta as it gets and is a perfect example of "handholding"

[The Wiggly Armed Man]

引用自 jaxjace：

All I got from this wall of text was that the m18 is the most played vehicle in the game, gee colour me suprised. its americas favorite clubbing tank. Its about as Meta as it gets and is a perfect example of "handholding"

You didn't just miss the point, you scooped out your eyes and stuffed them in your ears.

He used the M18 as an example as to how Thunderskill represents a very small minority of the actual number of battles that take place, try reading it next time.

[jaxjace]

引用自 Shovel Knight：

引用自 jaxjace：

All I got from this wall of text was that the m18 is the most played vehicle in the game, gee colour me suprised. its americas favorite clubbing tank. Its about as Meta as it gets and is a perfect example of "handholding"

You didn't just miss the point, you scooped out your eyes and stuffed them in your ears.

He used the M18 as an example as to how Thunderskill represents a very small minority of the actual number of battles that take place, try reading it next time.

I did, Im just doing what you guys are doing and using your data to fit my narrative.

[The Wiggly Armed Man]

引用自 jaxjace：

引用自 Shovel Knight：

You didn't just miss the point, you scooped out your eyes and stuffed them in your ears.

He used the M18 as an example as to how Thunderskill represents a very small minority of the actual number of battles that take place, try reading it next time.

I did, Im just doing what you guys are doing and using your data to fit my narrative.

What?

[Ichigo]

and people here wonder why everyone thinks steam forums is a joke

[Dakota]

The irony of these guys comparing us to flat earthers while they turn a blind eye to every bit of information we have against them is the best.

[Kotory]

引用自 Dakota：

The irony of these guys comparing us to flat earthers while they turn a blind eye to every bit of information we have against them is the best.

except there is NO information from you
"beliefs" and "assumptions" from fanboys are NOT information.
Numbers - are.

[Ouroboros]

引用自 Kotory：

引用自 Dakota：

The irony of these guys comparing us to flat earthers while they turn a blind eye to every bit of information we have against them is the best.

except there is NO information from you
"beliefs" and "assumptions" from fanboys are NOT information.
Numbers - are.

Sobolewski did give you numbers, though.

[Franss the Coffee Addict]

引用自 Shovel Knight：

引用自 Franss the Coffee Addict：

Ya know what would also be pretty? Putting the research trees in the order of date they were produced.
Unless its already a thing. I dunno I havent touched this game since last year's january

Are you suggesting historical matchmaking?

Eh not really. It would make the game hella unbalanced. Just keep the battle ratings, but put the order of research according to history

[Scheneighnay]

引用自 Franss the Coffee Addict：

引用自 Shovel Knight：

Are you suggesting historical matchmaking?

Eh not really. It would make the game hella unbalanced. Just keep the battle ratings, but put the order of research according to history

That would be a mess too, especially coming to American vehicles, where the engineers played musical components

[Haegler]

引用自 Franss the Coffee Addict：

引用自 Shovel Knight：

Are you suggesting historical matchmaking?

Eh not really. It would make the game hella unbalanced. Just keep the battle ratings, but put the order of research according to history

What exactly do you mean by that?

It sounds like historical BRs/research trees, which are, as you aptly put it, unbalanced.

Or do you mean that the stuff like an ASU-57, a 4.0 tank, should be in Tier V for research but still 4.0 battlerating, so you can, once unlocking T-54s, also unlock an ASU-57 which you can finally put into your 4.0 lineup?

And that would help in what way again?

[Scheneighnay]

引用自 Haegler：

引用自 Franss the Coffee Addict：

Eh not really. It would make the game hella unbalanced. Just keep the battle ratings, but put the order of research according to history

What exactly do you mean by that?

It sounds like historical BRs/research trees, which are, as you aptly put it, unbalanced.

Or do you mean that the stuff like an ASU-57, a 4.0 tank, should be in Tier V for research but still 4.0 battlerating, so you can, once unlocking T-54s, also unlock an ASU-57 which you can finally put into your 4.0 lineup?

And that would help in what way again?

that's already annoying with China too. You have to grind through the 4.3 B-25 in order to research the 3.7 Mosquito.

[Sobolewski]

引用自 Mirage：

引用自 Le Troisième Empire Français：

Sobolewski did give you numbers, though.

He gave his personal interpretation of the numbers and most of it is guess work or outright wrong to promote his view on things.. you disagree with stats that are direct copies from ingame performance but some numbers and statistics someone throws together is solid evidence.

Yeah, sure. If a simulation that is MADE to try and induce your proposed result of "Thunderskill being statistically reliable because it has a massive sample size" fails while using data collected from in-game and also from Thunderskill as well, then a neutral reality is bound to be exactly as you propose, yes?

Kinda like marxists saying "you don't understand dialectics, and thus you fail to understand anything else" as a counterargument when people from every reputable economic school prove them the impossbility of economic calculus on a centrally planned socialist economy.

If you think my simulation is too "biased", then provide me what you believe to be the average duration of a match and the average simultaneous matches in-game. We can run the simulations again and try to induce a result that proves your statement again.

But I have the impression you'll just keep doing the same old three-line dismissal by accusing me of bias.

[DanieL]

M4A1(76) and M4A2(76) moving to 5.3... nah, it's total trash now... another ♥♥♥♥♥♥ up by Gaijin