Sign in to follow this  
Shaggy996

WN8 and lower tier arty

34 posts in this topic

So is the WN8 expectancy of arty damage especially lower tier skewed hard from pre-nerf   numbers?

 

I have been working on the M41 and a good game on that thing is about 1200 damage with most games being more like 600 or 700 damage. Now this is with BiA and using food. Im also not bad on arty. Im good at first light hits and leading moving targets. But in tier 5 most of the games go so fast you dont have time to do a lot on arty and half the targets are fast light tanks.

 

So at 1200 damage with 1 kill it was a 1400 WN8 game. I just dont see how the AVERAGE damage is that high for that arty. Not with arty in its current state at least.

 

So if you want to level arty are you just screwed by the stats of what it USED to be,or am i just bad and don't realize it? lol

Share this post


Link to post
Share on other sites

expected Stats for arti with the latest update (v15):

 

 

2 + 3 = 7

 

 

That happens when you blindly follow some numbers and stop thinking about whether they make sense or not. v14 of the expected values were pretty good for arti, because some people took the time to fix nonsense manually. Overall arti and light tanks need a two point system (WN9) to be measured correct or much manual tweaking.

Share this post


Link to post
Share on other sites

Numbers are skewed from prenerf arty, and when you manually start tweaking numbers instead of relying on data, people will then question all of the numbers. If it bothers you, don't play arty; if you want the 53/55 then continue grinding.

Share this post


Link to post
Share on other sites

Numbers are skewed from prenerf arty, and when you manually start tweaking numbers instead of relying on data, people will then question all of the numbers. If it bothers you, don't play arty; if you want the 53/55 then continue grinding.

 

v14 was a good compromise between pre 0.8.6 numbers and current ones.

Share this post


Link to post
Share on other sites

v14 was a good compromise between pre 0.8.6 numbers and current ones.

When we update the numbers next, make sure to point it out (I agree; arty expected values are incorrect for current play).

Share this post


Link to post
Share on other sites

When we update the numbers next, make sure to point it out (I agree; arty expected values are incorrect for current play).

 

I already did it with v15 for artis - sure I only mentioned mid and hightiers cause and I dont play low tiers, but no one was really interested.

 

The thing is that for v14 we had at least 3 people, myself included, who went over the arti numbers seperatly and checked them for plausibility and errors. Thing is for that you need "arti players", not arti owners.

Share this post


Link to post
Share on other sites

I still think the expected damage for arty in general is too high. TD42 got knocked off the top wn8 players and everyone knows he's the arty god. Idk. I remember looking through them and seeing the expected damage to be obnoxiously high.

Share this post


Link to post
Share on other sites

Ok, I did an expected values run on post-8.6 accounts, comparing to v15:

 

- Tier 8-10 arty expected values are pretty decent.

- Tier 6-7 arty are generally overrated by v15.

- Tier 4-5 arty are about right except for the M41 and SU-5, which are massively overrated.

- I don't have enough data for tiers 2-3.

 

Tier 6+ will probably fix themselves over time. The M41 and SU-5 need manual correction.

Share this post


Link to post
Share on other sites

Ok, I did an expected values run on post-8.6 accounts, comparing to v15:

 

- Tier 8-10 arty expected values are pretty decent.

- Tier 6-7 arty are generally overrated by v15.

- Tier 4-5 arty are about right except for the M41 and SU-5, which are massively overrated.

- I don't have enough data for tiers 2-3.

 

Tier 6+ will probably fix themselves over time. The M41 and SU-5 need manual correction.

 

 

With a few exceptions expected values for t8+ artis are much to high or in some cases totally unrealistic. I challange u to show me the players that are able to achieve 3K damage solo random in arti for ~2500 WN8 - u wont find them. Even Baines Obj 261 DpB is/was going down - now probably around 2850 and still sinking.

And even before 0.8.6 3K DpB was unrealistic in randoms without massive platooning and favorable platoon setup. The only times you could play 3K DpB were probably up to 1 - 1.5 years after release, after that autoloaders, arti parties and TD alpha were starting "stealing your damage" with arti.

 

 

edit: Just to make it clear - I m talking about EU server here.

Share this post


Link to post
Share on other sites

With a few exceptions expected values for t8+ artis are much to high or in some cases totally unrealistic. I challange u to show me the players that are able to achieve 3K damage solo random in arti for ~2500 WN8 - u wont find them.

 

I would guess that this is because arty skill scaling is terrible. Players with 1565 recent WN8 really do deal 1.9-2k damage in their tier 10 arty since 8.6.

Share this post


Link to post
Share on other sites

Well it seems Im glad i brought this up since it started a good conversation about this. 

 

As for the high tier arty. It might be a little high but the big thing I notice on my GWE is there is very little difference on WN8 if i do 3k damage or 5k damage compared to normal tanks. The spread seems much higher. 

Share this post


Link to post
Share on other sites

I would guess that this is because arty skill scaling is terrible. Players with 1565 recent WN8 really do deal 1.9-2k damage in their tier 10 arty since 8.6.

 

Its even more "complicated" - if I would play only for damage in my arti, my DpB would be higher but my WR lower. Its impossible to gauge empty damage farming in arti - for that we would need pure solo stats and put WR and DpB in relation like in PR from Mr Noobmeter.

 

 

 

Well it seems Im glad i brought this up since it started a good conversation about this. 

 

As for the high tier arty. It might be a little high but the big thing I notice on my GWE is there is very little difference on WN8 if i do 3k damage or 5k damage compared to normal tanks. The spread seems much higher. 

 

u need to make kills also, without them ... .

Share this post


Link to post
Share on other sites

Its even more "complicated" - if I would play only for damage in my arti, my DpB would be higher but my WR lower. Its impossible to gauge empty damage farming in arti - for that we would need pure solo stats and put WR and DpB in relation like in PR from Mr Noobmeter.

 

 

 

 

u need to make kills also, without them ... .

 

Oh if im doing high damage ill have a few kills. I would rather take a gun out of the game then farm damage.

Share this post


Link to post
Share on other sites

Sometimes - its very situational with arti, because of your long reload.

 

well yeah of course. if they guy is dying in 10 sec anyways and I have a high health target, im not going to snipe the kill.

Share this post


Link to post
Share on other sites

Its even more "complicated" - if I would play only for damage in my arti, my DpB would be higher but my WR lower. Its impossible to gauge empty damage farming in arti - for that we would need pure solo stats and put WR and DpB in relation like in PR from Mr Noobmeter.

 

 

Oh, that's interesting. Being an arty noob, I assumed that damage & frags would correlate pretty well with winrate for arty. In fact it's barely better than the correlation for scouts.

 

 

As for the high tier arty. It might be a little high but the big thing I notice on my GWE is there is very little difference on WN8 if i do 3k damage or 5k damage compared to normal tanks. The spread seems much higher. 

 

Other tank classes get a lot more freebie WN8 from the spots*frags term. For arty, the scaling would be relatively close to linear.

Share this post


Link to post
Share on other sites

Oh, that's interesting. Being an arty noob, I assumed that damage & frags would correlate pretty well with winrate for arty. In fact it's barely better than the correlation for scouts.

 

 

 

Its pretty simple to explain, lets take Sandriver as an example. You spawn NW base with your arti.

 

You have usually 2-3 possible target areas.

 

1) J/K line

2) Some middle spotting often revealing base camping Maus, E100s and and other pro tankers around F/G 9

3) faster tanks whirling around the NE corner of the map

 

Usually you should support either 1. or 3. with your arti, and let the pro tankers around F/G9 lose the game for the enemy, if you want to win the match. Ofc there are exceptions to that like high priority targets, but usually the base camping Maus is as dangerous to your team as a Loltraktor when it comes to winning a match. On the other hand, if I d play for damage, I would go for the pro campers in the enemy base. 

Share this post


Link to post
Share on other sites

Once you start adjusting tanks that are obviously rated in an inaccurate fashion, all the other tanks that are not so obviously out of whack would need to be adjusted also 

in order to be fair. You don't want the only tanks being re rated to be the tanks that are driven by the people that contribute on wotlabs.

 

Then when you re rate all the other tanks you might as well not even have the wn8 framework and just assign rating by vote.

Share this post


Link to post
Share on other sites

Once you start adjusting tanks that are obviously rated in an inaccurate fashion, all the other tanks that are not so obviously out of whack would need to be adjusted also 

in order to be fair. You don't want the only tanks being re rated to be the tanks that are driven by the people that contribute on wotlabs.

 

Then when you re rate all the other tanks you might as well not even have the wn8 framework and just assign rating by vote.

 

Strictly speaking, player/unicum experience should be used to find flaws in the formulas, rather than adjusting things by hand.  The difficulty is dealing with the inertia of 'yet another system'; I think that better versioning on WN would have helped, although that wouldn't be obvious until several iterations in.

 

If we had: WN8 = v1.5, WN7 = v1.0, WN6 = v0.9, WN5 = v0.8 that might make it easier to release updates by making it a more clear progression rather than a 'new rating'.  I'm not sure it would make that much of a difference though.

Share this post


Link to post
Share on other sites

Once you start adjusting tanks that are obviously rated in an inaccurate fashion, all the other tanks that are not so obviously out of whack would need to be adjusted also 

in order to be fair. You don't want the only tanks being re rated to be the tanks that are driven by the people that contribute on wotlabs.

 

Then when you re rate all the other tanks you might as well not even have the wn8 framework and just assign rating by vote.

 

 

it still stands v14 > v15 when it comes to arti by a great margin. As long as there is no way to come up with a automated system to measure arti or lights correctly, carefull adjustments by hand beats the nonsense that we currently have there with the expected values.

 

Look up the v14 for artis and look up some ingamestats, and you will see, that we didnt hand out free WN-points. There were very few regular arti players, who came close to 3K or more WN8.

Share this post


Link to post
Share on other sites

I agree, the current iteration of wn8 is more flawed than its predecessor.  It also has less bias.

As many have stated, the entire approach that wn8 takes is questionable and it seems to me a better system would

be a wn7b or a system based entirely on the inherent qualities of each tank. (which is fully known )

 

The latest balancing that WG is in the process of makes the entire rating system of wn8 feel extremely dated , what do you guys think?

 

There is a long list of tanks(including low level arti) that anyone wanting a high wn8 should avoid. That list grows every patch.

If we continue to use wn8 I'll have to start pre-padding ala the garbad thread. (play tanks that will be nerfed, then cashing in on the

free wn8 after the update of wn8 after the nerf bat hits. )


Strictly speaking, player/unicum experience should be used to find flaws in the formulas, rather than adjusting things by hand.  The difficulty is dealing with the inertia of 'yet another system'; I think that better versioning on WN would have helped, although that wouldn't be obvious until several iterations in.

 

If we had: WN8 = v1.5, WN7 = v1.0, WN6 = v0.9, WN5 = v0.8 that might make it easier to release updates by making it a more clear progression rather than a 'new rating'.  I'm not sure it would make that much of a difference though.

Right on for the most part.

 

I would say that wn8 should not have been in the wnx sequence at all though as its an entirely new approach.

Share this post


Link to post
Share on other sites

Once you start adjusting tanks that are obviously rated in an inaccurate fashion, all the other tanks that are not so obviously out of whack would need to be adjusted also

 

Not really. The automated method used for v15 causes most tanks to converge to a reasonable value over time. This only breaks down if a tank was vastly more popular before the nerf and the difference between the old and current capability is very high. This doesn't currently happen with any non-arty tanks.

 

Here's a summary of the expected damage values from v14, v15 and a run I performed on post-8.6 accounts. Clear convergence problems are highlighted in orange. Possible convergence problems that may fix themselves eventually are marked in yellow:

 

https://docs.google.com/spreadsheets/d/1cJSNCqzgZWFoXI3bPWubpJ4cC9d3cESvU42oCEsaSOA/edit?usp=sharing

Share this post


Link to post
Share on other sites

I already did it with v15 for artis - sure I only mentioned mid and hightiers cause and I dont play low tiers, but no one was really interested.

 

The thing is that for v14 we had at least 3 people, myself included, who went over the arti numbers seperatly and checked them for plausibility and errors. Thing is for that you need "arti players", not arti owners.

the skill trees for arty are fairly flat, whats the issue with using the real data from the dossier searches? 

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.