pauli reacted to BoilerBandsman in WN9 aesthetics poll
I don't know if you can still find it, but Garbad did a challenge/experiment where he only allowed himself 5 shots a game, to see if he could quantify the value of spotting/tanking/positioning. He pulled something like a 54% winrate. Whatever you think/thought of Garbad, for a guy who routinely pulled high-60s winrates on solo challenges, it's pretty clear that using your gun dwarfs the impact of anything else in almost all cases.
Obviously there are anecdotal counterexamples. Everyone above green has vision-carried a match on Prok or Mali at least once, but WNx is designed to rate over many battles, and over many battles your gun activity is by far the most important.
pauli reacted to Fabunil in Object 268 Version 5
The 268-tumor is actually about the same size as the one on the E4 and it is weaker than on the E4 (the E4s tumor can still produce some bounces with the upper part of the cupola which has ~260mm EFA plus it has a 30mm overmatchable roof.
As for the turretcheeks, this isn´t really true either, several areas on the turretfront actually have less than 250 EFA and against gold rounds your turret front turns into hot cheese:
330mm HEAT vs E4 and 268-front:
So the E4 gets better turretarmor, much better hullarmor, better mobility/acceleration, a better gun (slightly more dpm, better(still shitty) softstats, much better goldround) versus the 268s 8mm of extra penetration and fully traversable turret.
pauli reacted to GehakteMolen in Has anyone noticed the client?
Imo all those captured tanks (aswell as ``historical versions``, so King Tiger with porsche turret and 88 L71) would make great reward tanks for a historical game mode.
Said historical game mode should also be mostly PVE, a team of humans vs bots, this way:
People can play historical stuff People can l2p People can drive their tiger here, and not in randoms Adds something usefull to the game Good way to add unique vehicles who dont fit in the tech tree yet are not worth beiing premium But that would make a lot of sense, so im sure WG wont ever do that....
pauli reacted to leggasiini in Swedish tanks
Nothing. The fact some nation was notable in WW2 or something doesnt mean that it should be instant priority to get tree, unless we talk about obvious stuff like Germany, USSR etc. Let me explain:
Czechoslovakia: big playerbase AND enough somewhat unique tanks for 1 line
Poland: big playerbase but lots of clones
Sweden: small playerbase but LOTS of unique designs for branches of all types (expect heavies are China-esque minibranch, from tier 8, but still)
Italy: both small playerbase AND lots of clones
Czechoslovakia is first of these "minor" nations (although China is sorta like one too, it was made mostly so WoT would be more succesful in China). It was chosen first because it has somewhat large playerbase, and enough candidates for atleast 1 tank line that doesnt have too many clones. Lowtier are sort of clones, but they are actually originally Czechs so German ones in the game are actually clones if anything. Poland has even bigger playerbase than Czech, but the problem is that they lack their unique tanks, especially for high tiers. Still, Poland's playerbase is so big that WG wants to implement Polish tree, a way or another. Sweden is mirror of Poland: it has small playerbase, however, it also has lots of very interesting tank designs that would be shame to be left out. Then comes Italy. Italy doesnt have particurarly big playerbase (bigger than Sweden, but still not as big as CZ/SVK or Poland), but it also has same problem as Poland: they lack unique designs. Or lets be more specific, they do have actually quite lots of unique designs, the problem is just that they are tier 6 material at maximum. At tier 7, they would have Panthers, Pershings and Patton, not even really modificated ones, just full clones of Panthers, Pershings and Pattons. And at tier 10 they would have just a Leopard 1 with autoloader, which tbh doesnt sound like it would have something special that TVP, 50B or Batchat doesnt have. The Czech and China is better because clones at low tiers + not full clones at high tiers > unique low tiers and clone high tiers, since +95% cares more about unique tier 9-10 than unique tier 1 - 5.
Thats why Italy doenst have that high priority, but its actually considered to added, maybe even next year.
pauli reacted to RichardNixon in T-22 to be nerfed
Armour efficiency has little dependence on skill. Near-identical tanks with inferior players tend to have near-identical armour efficiency. Armour efficiency does have a strong correlation with tank power, at least when measured by winrate. The T-22 did at last count have an extremely high skill-adjusted winrate. I need to re-run this with data from the last three months.
pauli reacted to lesterquaestor in Team death ratios- are roflstomps too common?
I did this, and posted on the official forums:
The results were that for perfectly balanced teams the chance of a blowout is about 20% (or 1 in 5 battles). A blowout was defined as a result worse than 15-5. For badly imbalanced teams the probability rose to about 33%.
In the thread someone else posted his personal statistics he has kept and showed that they were similar to what I calculated.
pauli reacted to Private_Miros in Missing Expected Values of New Tanks
In a probably useless effort to mediate the ever returning threads of often mediocre players demanding the addition of expected values for their new toys (in the vain hope their stats will get an insignificant boost to a value no one cares about), I will compile in this stickied thread the reasons for their loss.
Most of this information is originally from our resident statistical lump of gold, @RichardNixon
1. Why are new tanks not giving me stats:
- WG does not always immediately release the stats of newly added tanks in the API, meaning that WoTLabs cannot read the data to be included into its calculations.
- Added to this: WN8 data doesn't typically include any tank details except the ID. Sites used to do automated encyclopedia lookups, but WG murdered half their API team a few months ago, and no new tanks are currently being added.
For example: The Mutz was in the newer tankopedia API and the KR was in the deprecated one. The M41 isn't in either/
Not all is lost though: Sites (like WoTLabs) have to enter the name, tier, class etc manually. The data's there in the API though. This, however, takes time and manual labour, and known expected values (or replacement values).
It's not always easy either, as for example, there are two M41 90s in the API. The GF version (ID 50961) has something like 30-40k players on EU, while the plain one (ID 64017) has one player on EU and ~100 on RU, which is typical for supertest tanks.
2. Why are other sites giving me stats for this tank and WoTLabs not?
- WN8 didn't define a method for handling missing expected values, so sites vary. Some methods will universally inflate WN8, while others are in the right ballpark for most tanks or most accounts.
- A lot of sites pull updates automatically from wnefficiency.net. If that's not updated, then nothing happens though, unless replacement values are being used. Sites and session mods sometimes pick their own values for missing tanks when the official expected values are taking too long to update.
- WoTLabs "official" WN8 works with "official" expected values for each tank, using replacement values from other similar tanks, still need internal discussion, much like a full update of the WN8 expected values. This takes time.
Tanks will get added sooner or later. WN is a measure only relevant over larger number of battles. Replacement values also take time as enough players (and not only the good ones that free xp or grind faster) should have played the tank to get a relevant sample.
The people who take care of all this, do this voluntarily and in their own free time. They have no obligations to anyone, least of all to a random dude who's irky because his overal or recent meaningless value is temporarily slightly lower than it might otherwise be.
If you were referred to this thread because you made a thread about this subject without first checking this stickied thread, you are bad and should feel bad.
pauli reacted to Cyral in wottactic Alternative
I'm a bit late to this thread but I should mention that I am developing a web-based version of StratSketch that should be available in the next few weeks, I know a lot of people are holding off using it because the current desktop version has to be downloaded.
pauli reacted to Kalith in wottactic Alternative
Hi, I'm the creator of wottactic.tk.
First of all, the problems with wottactic.tk should be over now. Today 21/04 everything was smooth. And I'm confident I can keep it that way for the rest of the campaign.
I am truly sorry about the problems the last 2 days. I was not prepared for the stress the campaign would put on it, I've been on the other end of a tactic tool failing, it sucks and it always happens at the worst possible time.
pauli reacted to RichardNixon in WN9 candidate prototype
This is a WN9 candidate based on an idea of bjshnog's from a few months back, shortly before I quit forums to finish another project. WG are still ignoring requests to add assisted damage to the API, so it's not ideal, but I think it's still a sufficient improvement to be worth posting.
The general idea is that instead of using a full set of expected values for each tank, you have a single set per matchmaking type, or in a simplified version, per tier. Per-tank adjustment is only added at the last stage. The principle (for what it's worth) is that the value of a point of damage, frag or spot towards winning the game is the same regardless of which tank you're driving.
The practical differences are small for most tanks, but there are significant effects on the two problem classes: Scouts and arty. For scouts, spots are naturally better rewarded, while damage-padding is less rewarded. For arty, the big random chunk of WN8 provided by spots vanishes.
Improvements from other WN9 candidates were also included: True-recent expected values, per-tank skill-scaling and a linearised formula. These account for the vast majority of the advantages over WN8.
Part 1: Expected values
I based the initial expected values set on EU tier averages, although there's nothing special about those. With this method, you can effectively implement different formulas depending on tier and class by varying the expected values. For example, if frags are relatively important in the mid tiers, you can drop the expected frags value for mid tiers. I was lazy and just shifted scouts by a tier.
Part 2: Formula
I did a fair amount of formula testing with my custom multiple linear regression solver. To avoid repeating WN8's primary mistake, platoon-padded players were filtered out using a model based on topgun, BIA and CC medals. Each component was automatically pre-linearised to avoid tracking the slight non-linearity in contribution vs winrate.
Rather than chucking all the data into one test, I ran separate tests on various tiers and classes, with the following highlights:
Damage totally dominates frags at high tiers, but frags and damage have roughly equal importance at tier 6. This is probably because high tier tanks see a far smaller hp variance than lower tiers. Cap points have a strong influence at high tiers and for scouts, although not for mid tiers. I'm not going to advocate including them, but it does suggest that the pathological fear of the cap circle you see in high tier games is counter-productive. Scouts really like frags*spots, and to a lesser degree dmg*spots. The correlation is actually quite good these days (R2 ~=0.83), not far below mediums. Mediums also like frags*spots more heavies do, but the difference isn't huge. Artillery just wants damage & frags, nothing else. Correlation is pretty bad, probably because artillery had a big performance dive since the missions started. Defense tends to come and go, and it's never large. Frags*defence is generally preferred to plain defence. Parameter caps are generally counterproductive as long as the components are roughly linearised. Now, formulas aren't magical. You can mix and match various components and the correlation doesn't change much as long as they're roughly linearised. There are also reasons to prefer a formula with an inferior overall correlation:
Damage is far more accurate than other parameters over short runs. Important if you want a good per-tank stat. Trivially-paddable parameters should be avoided, especially if they may harm the game. Obvious examples are cap points and survival rate (which is generally negative anyway), and it's another reason to lessen the use of frags. You may want the formula to be more accurate for players who care about the result. Hence the formula may prioritise accuracy for high tiers and medium tanks at the expense of mid tiers and artillery. I compromised on 0.7*rDmg + 0.25*sqrt(rFrag*rSpotC) + 0.05*sqrt(rFrag*rDef), although there are likely benefits from class-splitting the expected values for scouts and arty. rDef should probably be capped and scaled to prevent it blowing up for tanks with a handful of games.
Single battle formulas (as used by mods) shouldn't have multiplications in them because they make the result too noisy. Something like 0.65*rDmg + 0.12*rFrag + 0.15*rSpotC + 0.03*rDef would be fine.
Part 3: Per-tank adjustment
Only two values are used per tank, an offset and a scale. They're not easily comparable cross-tier, because they also encode tier difficulty variations. Within a tier, tanks with a higher offset work better for bad players, while tanks with a higher scale work better for good players. If both values are low, then that tank is simply bad.
The values are generated using a regression method from a 6-week sample of 300k high-activity EU players. Earned XP values are used to exclude stock tanks and to perform some mild crew-skill & player learning adjustment. Non-random battles are filtered to an extent, with a reduction to around 1% of all battles. Most tanks from tier 5 upwards had sufficient data, although a handful had values substituted manually: One version each of the KV-220 and Pz V/IV, plus the IS-5 and ISU-122C.
Various issues with the per-tank adjustment:
Tiers 1-3 (especially 1) are much more difficult since the newbie protection was introduced. Older sealclubbing performances will be overrated. There is an option of using older data for the low tiers instead. Arty have been doing relatively badly since the missions, so arty played previously will come out relatively high. As this is based purely on recent data, performances in pre-nerf tanks will be overrated. The T-55A and Object 260 are probably overrated, due to players switching back from completing missions to padding damage.
To test how good various metrics are as overall metrics, I used the following method:
Throw away accounts with significant platoon-padding. Calculate the result of each metric for each account. Place the results in bins split by tank-adjusted winrate. Calculate the standard deviation of the results in each bin. Scale the results for each metrics so that they're roughly comparable. Lower standard deviations mean that the players with similar "skill" have less numerical variation in that metric. Tank-adjusted winrate is used for "skill", because it's probably the best account metric for solo players. When scaling the metrics, I left the zero points alone, so to get an idea of relative error you need two graphs:
WN8 and especially WG-PR should gain an advantage here from using winrate in their formulas (and proxies such as survival rate and base XP), at the expense of accuracy in platoon-padding cases. The vast majority of high-skill players are somewhat platoon-padded. I used true battle-weighted WN8 and WN9 rather than the usual fudge, because all the other metrics involved are battle-weighted. This may give both metrics an unfair advantage over WG-PR: I didn't check which weighting method resulted in higher error. Using rWin (WN8 component) instead of tank-adjusted winrate made little difference to the relative position of the metrics, although it helped WN8 slightly at the middle of the skill range. All errors increased with rWin, because it's an inferior metric. Zero points are optional. I chose zero-means-zero for WN9 to retain linearity for player comparisons, even though players who do nothing don't really exist. WN8 uses something like zero-means-average-bot. It wasn't possible to get accurate top-end values for WG-PR because you can't calculate it for a subset of an account. Observations:
WN8 is an increasingly poor metric with increasing skill, as expected. On absolute error it cheats at the low end, because the non-linear terms and zero-point caps drag the bad players closer together. TEFF is terrible in the middle, but better than WN8 at the top end, as expected. WG-PR runs pretty close to WN8 on relative error, except at the low end where it's the single best metric. This is probably because the bottom of the graph is heavily populated with deep campers, and so access to spotting damage and direct use of winrate are big advantages. The WN9 candidate is a much bigger improvement than I expected. Ideally I'd also do a per-tank error comparison, but because I changed my data collection methods a few times I don't have sufficient data for it. Eyeballing recent tanks on various accounts suggests that the per-tank error is very low compared to WN8.
The account WN9c2 values only use account/info and account/tanks, so they have essentially the same overhead as WN8. The tier and tank values use tanks/stats instead. Ideally the tier & tank values would only use random data, but the API's so broken that you end up with a lot of negative junk, so currently they include clan, team & stronghold battles as usual.
Tweak the expected values to improve some tank classes. Make a decision on what to do with tiers 1-3. Cap & scale defense to prevent it blowing up on low game counts.
pauli reacted to Tringapore/CAball in The T110E3, because you don't give any fucks.
we all fell for the armour meme
tip: wiggle harder than hillary did out of benghazi, it'll keep you alive alot longer
wiggle like the clinton
Shoot like the donald
Dont back down like bernie
Bounce all ghey accusations like cruz
Keep your engine oiled (i mean 100 octane) like rubio
pauli reacted to Never in "Where the hell is Never?"
So, hi guys.
I've been very inactive the past few months, and I probably should have made this post a while ago, but well, here it is.
The reason I've been so far from pretty much everything is that, to put it bluntly, my health is going to shit. I've been to the doctor more times than I can count in the last 3 months. I've been having chest pains and sometimes I can barely breath. If I don't change, I'm going to die. And that pretty much put me in a pretty bad depression. Then when I remembered WoTLabs and everything I promised everyone I'd get done, I'd feel worse and worse.
Well, this week has been a small ray of sunshine for me. I've since started a heavy diet with close medical supervision. For the first week since I can't even remember when, I've felt no chest pains at all, I've not been out of breath for no apparent reason and I generally feel a bit better.
So first things first, sorry for not being present here, barely answering PMs and generally not getting anything done on WoTLabs 3.0 or the replay hosting. Right now, I gotta be honest with myself and with you guys. I can't get these things done. I've barely been to my computer at all and that's not going to change for a little while. I'll do things here and there, but I'm mostly gonna be focusing on, well, not dying.
Well, that's it. Love you all <3
pauli reacted to Orrie in WN8 Expected Values Update
New test, new tanks, new ID's:
Missing Tanks: Name ID Tier Type Nation Status Grille 15 L/63 19217 10 td germany WT-E100 replacement leKpz M41 90mm 64017 7 light germany prem Turán III prototípus 63249 5 medium germany prem Т-44-100 (Р) 62977 8 medium ussr prem (RU only) Sentinel AC I 52817 4 medium uk prem Chieftain/T95 15441 8 medium uk prem (CW gift) Supertest 0.9.15: T-45 50945 2 light ussr new - R125_T_45 Object 268 Version 5 50689 10 medium ussr new - R126_Object_730_5 KV-4 Kreslavsky 63233 8 heavy ussr new - R128_KV4_Kreslavskiy T71E2 / T71 CMCD 57633 7 light usa new - A112_T71E2 Supertest: Name ID Tier Type Nation Status Т-44-100 61953 8 medium ussr Object 244 60161 6 heavy ussr Kirovets-1 62721 8 heavy ussr Object 777 Version II 61185 10 heavy ussr KV-4 KTTS 62209 8 td ussr Panzer 58 63761 8 medium germany VK 45.02 (P) Ausf. B7 62225 7 heavy germany VK 45.03 62993 7 heavy germany VK 100.01 (P) 62737 8 heavy germany Krupp-Steyr Waffenträger 61713 7 td germany Rheinmetall Skorpion 62481 8 td germany T25 Pilot Number 1 57377 8 medium usa M48A2/T54E2/T123E6 56865 10 medium usa Sentinel AC IV 56145 6 medium uk Chieftain Mk. 6 15185 10 heavy uk AMX M4 mle. 49 62785 8 heavy france Bat.-Châtillon 25 t AP 62529 9 medium france 121B 63537 10 medium chinese
I've found the 121B on RU supertester accounts, same with Object 777 V2.
Could you guys post the ID's of accounts with supertester tanks?
pauli reacted to stagnate in Changes to CW
Ugh. When are they going to understand that 7v7 is really poor on these maps. It leads to one-dimensional gameplay and does absolutely no good in helping a caller practice leading larger groups (ie leading up to 10v10 or 15v15).
I can partially understand with strongholds that they want to create a low barrier so that it's vibrant and active since it is on-demand, but for clan wars that will happen for a few battles each night they really should be sticking with 15v15.
pauli reacted to Sidus_Preclarum in IDF Tanks in WoT?! Not quite... (M51 Super Sherman inbound)
Tbh, reaching 1st mark with a 100% survival rate is not bad.
pauli reacted to sela in [0.9.17.1] sela's mods - v126.96.36.199 released 26/03/2017
Pack updated to v9.7.1. If you are having CTDs, please install the latest update (old engine_config.xml possibly causing issues). There may still be issues as the sight mods are not officially compatible with 9.7 yet.
"..\Games\World_of_Tanks\res_mods\configs\xvm\hotkeys.xc". Edit "keyCode" value on line 3 "minimapZoom"; I have this set to "29" (Left Ctrl).
Here is a list of common key codes: