Jump to content
stagnate

Avg XP best indicator of performance?

Recommended Posts

Most of the stats discussions revolve around how to evaluate performance, and are limited by what we can tell or know about players based on the API and other info.

 

Avg XP is thrown out because it is not possible to tell whether it is with or without premium, or a mix from a programming standpoint (including remote observers).

 

However, you personally can track xp and identify when it is increasing or decreasing by also accounting for when you have premium.  I realized that theh trigger for my to improve my game was when I thought I had good xp/game (650 or something, back in beta) and realized that my 'peers' had 100+ xp/game more than I did.  I would recommend that any player looking to increase their performance should pay more attention to xp/game than other stat measures.

 

That being said, does anyone have thoughts about xp as a performance measure?  I haven't really seen any threads constructively taking apart xp calculations - generally just some confusion from players and lots of whining.  Are there types of games that are poorly rewarded from an XP standpoint?  Possibly it weights kills and damage too heavily compared to spotting, and any assist tank (scouts) will have some trouble, but I haven't really heard complaints about xp in them (gameplay yes, xp no).

 

Also, one of the stats/replays sites should do a mass data analysis to figure out the xp parameters.  There should be enough data out there to do it (may violate T&C though).

Share this post


Link to post
Share on other sites

I think there are a couple snags. Sniper is too easy to farm for certain tanks, taking the pain of losing a game away from your XP. Next is XP multipliers; not all tanks earn the same XP. Take out a Church III or KV-5 and perform as good as your peers and you'll be earning 20%+ more XP.

 

At the same time, XP doesn't offer any sort of condolence for bad tanks. You're just going to get shit on.

 

If XP was normalized and adjusted to equally award all tanks it'd be one of the best. Its my favorite stat to keep track of using my own regulations, but one of the worst when looked at without deep knowledge of why.

 

With that said, since you should be able to answer that question about your own XP, it's great for measuring yourself (and only yourself.).

Share this post


Link to post
Share on other sites

Another problem with average exp, other than pain in the ass to track premium time if you have it on and off like a lot of people, is the platoon win rate "inflation" of it. In essence, people that win more will have higher average exp. "That's the point isn't it?" Well not entirely, if you're platooned vs solo or just had a good session with favorable MM and good teams (10-20 games are few enough where the swing of things can be big), you're going to have much higher average exp than your previous solo session just due to 1.5 exp multiplier on more games. While you'll do more damage solo, the exp from that won't come close to making up the lack of 1.5 multiplier. Now, if you normalize it, by taking out the winning multiplier for each session I can see it being viable. I've had the "win rate normalized avg exp" thought for a while, but I'm not sure if it makes sense to anyone else to take out the winning multiplier when using exp as a performance gauge, since winning is the goal.

Share this post


Link to post
Share on other sites

Pretty sure it's only the Churchill that gets the bonus exp Rexxie not the KV-5

 

But yeah it's one of my preferred methods of checking someone's stats , pity that premium affects it , would be cool if all exp shown was just base exp

Share this post


Link to post
Share on other sites

avg exp is the best indicator of skill if attention is paid to the following:

- the player plays only with (or without) premium account

- doesnt play churchill 3 (it has +30% extra exp)

- tds had a far less avg exp gain as tanks for a very long time

- arty still has a worse exp multiplier

 

also attention should be paid to the avg tier, tier 1-5 make (far) less exp on average, tier 6 and 10 are also a little below avg, while t7-8-9 give the best avg exp

 

however to compare tanks with eachother, avg exp is far, far better as wn8 or anything else, coz exp includes:

- dmg after spotting

- dmg after tracking,

- close range bonus (and close range is a FAR better sign of skill as shitting camping an ``supporting``)

 

a good player who plays only premium, not too much games with below t5 tanks / arty and not to much real turd tanks / total stock failures, should get 1k avg exp (overall) (and not just statpadding gold spammers, but normal tanks, be it a Tiger, KV4, JagdTiger or Centurion)

 

i myself have 982 avg exp for the past 10296 games (only premium account, thats why the interval is so strange :P) (checked with noobmeter) which is only below 1k due to td penalty and spg, without those 2 i would have had just above 1k avg exp, it DOES included wonder machines like SU-101, Churchill gun carrier and a whole bunch of stock tanks (i grinded brits, chinese tanks, 2e ru td line and french tds mostly aswell as some tanks i skipped before / didnt play yet)

 

http://www.noobmeter.com/recent/eu/GehakteMolen/500134362/500134362_00018376_20121225_171444/500134362_00028672_20131231_062926

Share this post


Link to post
Share on other sites

however to compare tanks with eachother, avg exp is far, far better as wn8 or anything else, coz exp includes:

- dmg after spotting

- dmg after tracking,

- close range bonus (and close range is a FAR better sign of skill as shitting camping an ``supporting``)

 

I've always felt suspect about using average exp as a skill indicator for all the reasons mentioned above, but I do think you make a very strong point when looking at someone's per-tank avg exp...

Share this post


Link to post
Share on other sites

Tank X vs Tank X (e.g. my T62 vs your T62)

Win rate, is the single best indicator of player skill because it is all-encompassing, all play conditions (platoons, ping, whatever) being equal

 

If play conditions are unequal, then DPG is the best metric

 

Tank X vs Tank Y (e.g. my T62 vs your Leo)

I like to use Minutes of Effective Fire here when the tanks have similar roles. MEF is basically DPG/DPM, and helps compensate for when tanks have different guns or when one is grossly better than the other

 

If the tanks are pretty different, I use XP

Share this post


Link to post
Share on other sites

Tank X vs Tank X (e.g. my T62 vs your T62)

Win rate, is the single best indicator of player skill because it is all-encompassing, all play conditions (platoons, ping, whatever) being equal

 

If play conditions are unequal, then DPG is the best metric

 

platoons totally ruin winratio comparisons, and so does CW, same for DPG, trow in platoons and CW and dpg also becomes very hard to compare...

 

tank for tank, exp is (by far) the best indicator, since the only ``disbalance`` is premium / no premium, even cw has not much influence on avg exp (a good clan will win most cw matches, which gives almost same exp as random battle)

 

it gets tricky when comparing tds from before / after avg exp buff, since avg exp went op sharply after that

 

my ISU had about 950 avg exp, all premium account, with 26xx avg / dmg and 66% win or so, this was all before the accuracy buff, arty nerf and exp buff, so this was the sort of ``max`` for a t8 td, after the buff, many t8 tds easy rack of 1k avg exp, showing how much more easy it is to get good exp with tds....

 

Also the avg exp went up a little because you now get exp for detracking an enemy, so comparing ``new`` stats with ``old`` stats (again) gets difficult...

 

all in all the game has changed so much, that old stats are becoming irrelevant for many comparisions, not due to learning curve, but due to changing mechanics / meta, while changes to HE and swapping tanks in tech tree had some influence, its influence is way smaller as the mega nerf to arty, which greatly increased hp pool + accuracy buff + td camping madness we now see...

Share this post


Link to post
Share on other sites

Generally for myself I use all of my stats together to see if I am actually improving. Main ones are WR, WN8, K/D, Average Damage, and average exp. If all of them are above my overall stats I know I have been doing something right, If one or more is below I try to figure out why. Sometimes it is just a night of potatoes, other nights I am the potato. I can feel like I played well one night and go back and my stats suck. Other nights I feel like I played terrible (usually because I get killed almost every match that night) but my stats are great.  

 

For me just avg exp would be a bad indicator because I tend to get on a kick and start playing more low tier tanks that I am grinding which would tend to keep the exp down. Most nights I do a mix of tier 4 to tier 10. Plus all the other reasons mentioned above tend to influence your overall exp for the night.

Share this post


Link to post
Share on other sites

Where is the info about the Church III 'officially' listed?  This is the first I heard about it, and already it's in conflict (20% or 30%, KV-5 included or not).  I generally keep good tabs on news and had never heard about any xp bonus.

 

Ech, the whole point of all the stats hoopla is that WR and DPG are not great measures.  Avg XP is going to be a far better comparison, because it includes the most facets of the game.  Trying to determine 'similar' tanks and use fire metrics doesn't work for non-dpm tanks (scouts, mostly) and any gameplay that isn't based around damage (tanking, scouting, etc).  Effective DPG or DPM is not a bad rough metric, but I can't see it being more accurate than avg xp.

 

 

Good point about lower xp for low tiers, and factoring grinds vs elite play also has some impact.  Since they have started using base xp calculations I continue to hope that is actually being tracked and will be available at some point.  Not holding my breath.

Share this post


Link to post
Share on other sites

Where is the info about the Church III 'officially' listed?  This is the first I heard about it, and already it's in conflict (20% or 30%, KV-5 included or not).  I generally keep good tabs on news and had never heard about any xp bonus.

 

Ech, the whole point of all the stats hoopla is that WR and DPG are not great measures.  Avg XP is going to be a far better comparison, because it includes the most facets of the game.  Trying to determine 'similar' tanks and use fire metrics doesn't work for non-dpm tanks (scouts, mostly) and any gameplay that isn't based around damage (tanking, scouting, etc).  Effective DPG or DPM is not a bad rough metric, but I can't see it being more accurate than avg xp.

 

 

Good point about lower xp for low tiers, and factoring grinds vs elite play also has some impact.  Since they have started using base xp calculations I continue to hope that is actually being tracked and will be available at some point.  Not holding my breath.

 

 

dunno, churchill III is one of the older/oldest vehicles of the game, and it has that bonus since forever i think (and its 30% and only for the churchill III afaik)

 

to give an idea how old the churchill III is, it was in game before there where US tanks, or arty stronger as an SU-14 aka schoolbus

 

ps: fun fact, Panzer IV was the first tank of the game i think, imo suprising its a german tank, and not a russian one :P

Share this post


Link to post
Share on other sites

As someone who played almost half of his games without premium, I could never understand why WG didn't just track average BASE xp in their stats. 

 

The objection based on the ease of sniper medals is somewhat valid, but I think it has a negligible effect. I get a LOT of sniper badges, but if I had to guess, I'd say that fewer than 20-30 have granted me winner XP while losing. It's not a large enough factor to warrant throwing away average XP as a ruler. 

Share this post


Link to post
Share on other sites

Given all the noise factors that can't be corrected for I don't think you can use Average XP to compare two players stats.

 

As the OP was actually talking about, you can use it to benchmark your own performance, especially for short periods of time (single game/1 day) where WN scores become noisy.

Share this post


Link to post
Share on other sites

http://www.reddit.com/r/IAmA/comments/1u75hh/i_am_the_guy_with_two_penises_ama/

 

Also, they removed all the exp buffs and nerfs in the arty patch.  Before that there were many premiums that got an exp/game buff.

 

i simply wont believe that...

 

i have never seen lowe/kv5/premium tanks get more exp as they should, with the SOLE exception of the churchill III, which gets ridiculous often ridiculous good exp (almost any1 with a churchilll III has a max exp game of over 2500, and even games of 2700-3000 are not ``that rare``)

 

also ace tanker of churchilll 3 is (exp wise) ridiculous high, 2400 exp was no ace tanker on EU server... (atleast, before i got ace tanker myself)

 

but thats 1 tank, i never seen any proof about prem tanks get more exp

 

ps: talking about normal exp now, not crew exp

 

edit, just example of what churchill 3 can (expwise) do:

http://wotreplays.com/site/312463#prokhorovka-dyaebl-churchill

 

23380 exp (with a x5), thats 4676 normal premium exp....

 

edit 2:

this numbers are heavy flawed (does it include premium etc?)

 

but churchill crushes all competition BY FAR exp wise (on tier 5)

http://www.vbaddict.net/statistics.php?tier=5&tanktype=0&nation=0&premium=0&team=0&battles=1000&groupby=0&fieldname=experience

Share this post


Link to post
Share on other sites

We all agree XP would be the ultimate skill rating.... with a few normalizations.

 

Aside from obvious prem/non-prem, there is also single tank XP bonus, like KV-5, church III, T14, 13-90 and others.

 

Also, there is a big difference between XP gained at different tiers. When was the last time you saw someone average 1000XP at tier 2? Tier 10s also get normally less XP than tier 8-9s due to tier difference bonus for damage.

 

 

 

Lastly, it would also need to be normalized by win%. A good player who platoons all the time for 85% wins, will get more avg XP than a really good player who solos 65% wins.

Share this post


Link to post
Share on other sites

We all agree XP would be the ultimate skill rating.... with a few normalizations.

 

Aside from obvious prem/non-prem, there is also single tank XP bonus, like KV-5, church III, T14, 13-90 and others.

 

Also, there is a big difference between XP gained at different tiers. When was the last time you saw someone average 1000XP at tier 2? Tier 10s also get normally less XP than tier 8-9s due to tier difference bonus for damage.

 

 

 

Lastly, it would also need to be normalized by win%. A good player who platoons all the time for 85% wins, will get more avg XP than a really good player who solos 65% wins.

 

no need to normalize, just remove the win bonus (and that shouldnt be so hard, if you win 50%, the winbonus is 1.25 so [avg exp / 1.25], if you win 66%, the winbonus is [0.5*0.66 = 1.33] (50% winbonus * % wins +100%), after that just divide avg exp by this ``bonus factor`` and you now the ``real`` avg exp

 

ps: this is, assuming that the avg exp in a defeat is equal to the avg exp in a win, this is ofc not true due to various reasons, yet winning/losing is the goal of the game, so thats imo not a problem (and good players will get much better defeats as bad players)

Share this post


Link to post
Share on other sites

no need to normalize, just remove the win bonus (and that shouldnt be so hard, if you win 50%, the winbonus is 1.25 so [avg exp / 1.25], if you win 66%, the winbonus is [0.5*0.66 = 1.33] (50% winbonus * % wins +100%), after that just divide avg exp by this ``bonus factor`` and you now the ``real`` avg exp

 

ps: this is, assuming that the avg exp in a defeat is equal to the avg exp in a win, this is ofc not true due to various reasons, yet winning/losing is the goal of the game, so thats imo not a problem (and good players will get much better defeats as bad players)

 

Isn´t what you just described normalizing by the % of wins? :D

Share this post


Link to post
Share on other sites

Isn´t what you just described normalizing by the % of wins? :D

ah, i read what you wrote more as divide exp by win%, or create a new ficitional number or whatever :P

 

ps: but yeah, we then agree, without win bonus avg exp is best indicator (except for that stupid premium bonus...)

Share this post


Link to post
Share on other sites

no need to normalize, just remove the win bonus (and that shouldnt be so hard, if you win 50%, the winbonus is 1.25 so [avg exp / 1.25], if you win 66%, the winbonus is [0.5*0.66 = 1.33] (50% winbonus * % wins +100%), after that just divide avg exp by this ``bonus factor`` and you now the ``real`` avg exp

 

ps: this is, assuming that the avg exp in a defeat is equal to the avg exp in a win, this is ofc not true due to various reasons, yet winning/losing is the goal of the game, so thats imo not a problem (and good players will get much better defeats as bad players)

Just normalize the final average won't be accurate, if you're looking to track minor changes. It makes a difference which battles were won if you're going to normalize the values.

You have two games of 100 and 200 base exp. The true average base exp of these two games is 150. But if you won the 100 exp game, your normalized avg exp through dividing by 1.25 would be 140, where as it'll be 160 if you won the second battle.

Share this post


Link to post
Share on other sites

 

 

Went through the entire thread to find a Churchill-III reference. Dude, that would have been more appropriate for discussing http://ftr.wot-news.com/2013/09/20/ensigns-qa-14/

 

Where do you find that stuff? The knitting post made my wife choke XD

Share this post


Link to post
Share on other sites

now i know where i saw that exp multiplier, wotdb, they (used) to have an (incomplete) list of (former) armor homoganisation values aswell, and so do/did they have a list with all exp / credit modifiers

 

http://forum.worldoftanks.eu/index.php?/topic/213558-xp-modifier-for-premium-tanks/

 

Outdated info
1.38 Experience factor
0.05 Free experience factor
1.15 Credit factor

 

in this case, Serb is clearly wrong or lies (as he says churchill 3 doesnt have a modifier) since WG also always stated that TDs get same exp as tanks (and later they put in patchnotes (and Q&A) that they removed this penalty....)

 

wotdb used to be the best source for all the hidden value stuff (like ground resistance and camo) and unlike all those other ``sources`` wotdb datamined the wot folders, so NO crappy ``hmm, looks strange`` numbers, like current camo stuff, but REAL hard numbers (straight from the game files), this was all in beta, and it was a huge uproar, also coz it showed armor values, firechance, and some other stuff :P )

 

WG after that (short before or after release i think) removed ALL data from the game files, and what was not removed got split up / decoded / cut apart, making it for example almost impossible to see the armor / hitzone models,, this made it impossible for almost a year or so to show gamemodels

 

here the numbers are discussed in detail, all the pictures (with the actual information) is however gone

http://forum.worldoftanks.com/index.php?/topic/25650-wot-secrets-english-version/#topmost

Share this post


Link to post
Share on other sites

Just normalize the final average won't be accurate, if you're looking to track minor changes. It makes a difference which battles were won if you're going to normalize the values.

You have two games of 100 and 200 base exp. The true average base exp of these two games is 150. But if you won the 100 exp game, your normalized avg exp through dividing by 1.25 would be 140, where as it'll be 160 if you won the second battle.

 

It should average out in the long run.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...