World of Tanks: Rating Systems

Everyone’s Stats Obsessed

Statistics are becoming an ever more popular feature in online games, especially those with a competitive element. Whether it’s because players like to brag, improve their skills, or need good stats to get entry into a clan, they’ve now become an indispensable element to any online game worth its salt.

World of Tanks is no exception, and it produces a wide range of statistics. However, unlike a number of other high profile online multiplayer games, its official rating system is always overlooked in favour of third party systems. These rating systems have not only matured with time, but recognise each other. Players haven’t just adopted one; they’ve chosen to use a combination.

Yet these are not without fault for two reasons. The first is obvious: Player-made systems tend to favour what the player who made them thinks is important. The second is a core problem with Wargaming – they either do not record statistics on a number of key things, or they include them within their dossier cache or API. Areas such as damage done upon spotting, damage done to vehicles you’ve tracked, or damage taken are seemingly done away with for no known reason. In addition to this, no system apart from Wargaming’s own rating system takes into account normalised Experience, because that data is not published by Wargaming.

WHAT DO THESE SYSTEMS MEAN TO ME? SHOULD I CARE?

The answer a humble man should give is no, unless you’re trying to do some self improvement. However this is an online game which is no place for the humble or modest individual. With the large scale adoption of XVM, anyone who has it installed can see your scores next to your name and not only prejudge your performance on them but possibly call you out. Why did that dude call me a tomato? Because it’s likely you have a low score. Why did the enemy team all gun for me at the outset? Because you may have a good or great score and you are the #1 threat.

This is the default view of the latest XVM – showing total number of battles, WN7 score, overall win ratio, number of battles in tank currently being used, and your win ratio in it. Had World of Tanks been around in Darwin’s time, he may well have given his theory of evolution a second thought…it takes effort to be that bad after 13,000 battles.

Another reason to care is clans. Much like how some EVE corporations want players to have X number of solo kills or Battlefield 3 clans require a certain kill-death ratio, World of Tanks clans will often set a minimal set of scores such as average damage across your Tier X tanks and a minimum acceptable WN7 score before you can gain entry. Whether you agree or not, it is a standard recruiting filter for middle and top flight World of Tanks clans.

So why talk about this now? Because Wargaming introduced their own revised rating system into World of Tanks in 8.8. This leaves an obvious question: What are the various rating systems and how do they work? How do they stack up against Wargaming’s new system, and which one really represents player skill in the fairest manner? In addition to this, we’ll look at some other questions like: How do I find out my statistics? Why do people laugh at me for having a score of ‘500’? Why do other players say “Oh god no” then drown themselves when I drive my E100?

Rating systems

First, we’ll look at the common rating systems available – those which have been adopted by the community and often lead to insults being thrown across the battlefield and forums alike.

Efficiency

Often referred to as Eff or incorrectly as XVM – XVM being a mod that just displays stats among other things – Efficiency was one of the first player made rating systems that appeared in World of Tanks, made by WoT News For a considerable length of time it was seen as a guide to a players ability despite being easier to manipulate than an alcoholic when you’ve got vodka to tempt them with.

Efficiency has gone through a few changes, but at its core it remains the same:

DAMAGE * (10 / (TIER + 2)) * (0.21 + 3*TIER / 100)
FRAGS * 250 +
SPOT * 150 +
log(CAP + 1,1.732) +
DEF * 150

The system takes damage done, your average Tier, number of kills, number of tanks spotted, and cap and defense points into consideration when working out the score. The big controversy is the use of cap and defend points, as they can effectively be farmed with not much effort, unlike doing damage or killing opponents.

There are arguments for and against the use of cap and defend points in any rating system, and both have validity. It’s fair to say that if you’ve fought your way to the cap and then proceed to try and win by capping – the point of the game – then cap points should be included. The same can be said of defence points; defending the cap by shooting enemies accumulating cap points aids your team.

However there are some fair, if slightly cynical, arguments against the use of these. Chiefly that people either racing to the cap to get the points, or those that hang round the spawn area to kill attackers once they start capping, are of no overall use to the attacking portion of their own team. It should be noted that defence points – the amount you’ve removed from a capping enemy – are recorded even if you lose. It’s possible to wait until an enemy has built up 90+ cap points, which you then remove, saving the match and banking a huge Efficiency increase.

In addition to this, unlike some other systems, there is no maximum amount of cap/defense points the system will allow. WN7 – which we’ll look at next – will not count defense points above a 2.2. average across all battles. Efficiency scores divide down into these brackets:

Bad – 0 to 649
Below Average – 650 to 859
Average – 860 to 1139
Good – 1140 to 1459
Great – 1460 to 1734
Unicum – 1735+

The idea here is that an average player would fall into the 48-49% win ratio category – which is the current win rate average across World of Tanks.

Efficiency does have some positives, especially over Wargaming’s new system. For instance, it takes into account the player’s average Tier when working out the average damage, does not factor in number of battles played or the number of battles won or lost. Despite the problems with cap points, it’s subjective whether or not Efficiency still holds its value as a valid rating system. It is, however, falling out of favour.

WN7

WN7 is the latest evolution of the Weighed and Normalized stats system. Since WN6, it has seen widespread adoption by elements of the World of Tanks community to the point it is now overtaking Efficiency as the number 1 method of judging players skill. This has likely pleased the creators, as they purposefully designed the WN system to address the shortcomings in Efficiency. It proved so successful that most sites which offer players the ability to view their own rating systems also show them their WN7 scores.

The formula looks like this:

(1240-1040/(MIN(TIER,6))^0.164)*FRAGS
+DAMAGE*530/(184*e^(0.24*TIER)+130)
+SPOT*125*MIN(TIER, 3)/3
+MIN(DEF,2.2)*100
+((185/(0.17+e^((WINRATE-35)*-0.134)))-500)*0.45
-[(5 – MIN(TIER,5))*125] / [1 + e^( ( TIER – (GAMESPLAYED/220)^(3/TIER) )*1.5 )]

Its implementation has a number of features to prevent common forms of stat padding, such as collecting cap and defence points, playing Tier 1 & 2 with questionably powered tanks and multi-skilled crews, or boosting win rates by playing Company Battles with experienced Clan Wars hardened players. In addition to this, a number of stats which can be effectively farmed have reduced importance in the overall calculation. The creators also decided that capping points are not an indication of player skill, so they completely dispensed with them.

So how does it do this? For Frags, it uses a series of arbitrary numbers while taking average minimum Tier (or Tier 6 if it’s higher) into account. The Damage stat is similar – taking into account average Tier – it allows someone with a Tier average of Tier X and Tier VI to be compared accurately. Both Frags and Damage have equal weight within the final score. The spotting score is more simple – it needs no Tier modifier since the number of tanks in a battle will always be the same. Defence points, as mentioned above, are hard capped at 2.2 and require no Tier modifier. Lastly, win rate roughly works like this: for every % above 48% – the accepted server average – you get additional points; however, this is done on an S curve, so it gives diminishing returns the higher you go above 48%.

The last calculation is certainly the most interesting – penalisation of ‘sealclubbers’ – and rather than describe it, here is a copy of the original formula from one of the creators, Praetor77:

What this shows is a sliding scale of penalisation. If your average Tier is under V you’ll be penalised with maximum deduction at Tier 1. It also scales with number of battles played, so a new player who’s working up through the Tiers sees a reduced penalty

Sounds great, but surely it must have faults. Like Efficiency, it suffers from lack of data from Wargaming, and it has received some criticism in respect to the balance of kills vs damage done. In WN7, kills have more importance than damage done. Yet this is still highly subjective, and players with higher than average damage stats (in relation to average Tier) tend to also have higher than average kills per battle stats.

Like Efficiency, WN7 breaks down a player’s performance into categories:

Very Bad – Under 500
Bad – 500 to 699
Below Average – 700 to 899
Average – 900 to 1099
Good – 1100 to 1349
Very Good – 1350 to 1499
Great – 1500 to 1699
Unicum – 1700 to 1999
Super Unicum – 2000+

Unlike Efficiency, Unicums and Super Unicums are particularly rare; the bulk of players sit between 500 and 1099. WN8 is soon to be released as well, which more accurately deals with Artillery damage. However, the largest change is how higher scores are handled. The scale for the system for scores above 1500 will change significantly, stretching out to 2800.

So how do you find out your WN7 score? As mentioned previously, the WN7 score is used on a number of player rating systems and websites. The most common one used is NoobMeter but at the bottom of this article we’ll provide a complete list of ways to check out your stats and scores on a number of sites.

Performance Rating

Performance Rating and the site that hosts it, NoobMeter, is a relative upstart in the world of stats. Like the other systems, it uses various statistics published by Wargaming to build up a score based on player performance. Its key difference is the in algorithm used to get that result. Unfortunately, the people who run NoobMeter will not publish this algorithm, leading to accusations of various sorts and preventing wholesale adoption by the community. All is not lost, though, as they have published information on the elements that go into making up the overall score.

Performance Rating is no different to Efficiency or WN7 in having scores divided into categories:

Bad Player – 0 to 1150
Below Average Player – 1150 to 1250
Average Player – 1250 to 1450
Good Player – 1450 to 1750
Great Player – 1750 to 1950
Excellent Player1950 to 2000
Unicum – 2000+

So what are the pros and cons of PR? It’s hard to say, because they do not publish the exact algorithm. This reason alone is a negative, leading to accusations from some members of the community that the score is engineered to make some players look better than others, i.e. the creators and their friends. It is, however, worth noting that a players PR score often tallies with their WN7 score – if one goes up or down, so does the other. In addition to this, the creators of Performance Rating do explain – often at length – some of the elements that go into making up PR, including what tanks are considered overpowered and Tier pentalies.

So while there are arguments about PR’s validity, it is a consistent method of rating players. If the owners considered publication of the full algorithm, WN would get a serious contender. The other downside of PR is it’s only available on noobmeter.com, whereas WN and Efficiency are available on a number of sites and in-game mods.

Wargaming’s Player Rating

Since World of Tanks release, Wargaming has displayed player statistics on their website. However, this was largely, if not completely, ignored by all and sundry due to it being a terrible system. This led to the growth in player-created systems using the limited tools Wargaming provides. At the same time, players have asked, complained, whined, and begged Wargaming to either create their own or add in an third party player rating system. Wargaming finally gave in with 8.8, and in true Wargaming fashion added something which they claim is perfect, yet to the discerning eye is hopelessly flawed. Not that SerB cares about it being flawed or what you think about it.

So how does Wargaming’s Player Rating (we’ll refer to it from now on as WPR) system work? It’s not entirely clear because Wargaming will not release the algorithm for it, but they hint at a few details. Win rate is one of the core statistics it uses, as is average damage and kills. However it does two things the other systems actively avoid. WPR system takes into account the number of battles played and does not factor in average Tier.

What is the big deal with that taking into account the number of battles played? Unfortunately, the more battles you play, the higher your score will be. Clearly this has nothing to do with player skill. Is refusing to take account average Tier good? This is subjective to a degree, but you do more damage as you go up Tiers. To some, playing Tier X is the pinnacle of the game and something all players should do. On the flip side, many veterans with multiple Tier X tanks are seen playing lower Tiered matches. The game play at mid Tiers is, after all, exactly the same as those below or above.

Apart from the obvious fact that a Tier X vs a Tier X is not much different from a Tier III vs a Tier III – they are balanced to fight each other – it can be seen as yet another carrot and stick method to get players to face the grind and advance up the Tiers, even though the core game doesn’t change. Players need to do more damage to get a better rating; to do more damage you need higher Tier tanks, to get higher Tier tanks you need to grind more or spend cash to reduce that grind and make playing Tier VIII sustainable. Remember that your average player on a free account does not break even after a Tier VIII battle.

We have to play Devil’s Advocate and put Wargaming’s reasoning across as to why they’ve done this. They claim they wish to prevent people using secondary accounts, hence the number of battles factors in. They also claim that by not having an average Tier when factoring average damage, the are preventing “sealclubbing,” or as SerB puts it, “peadobears”.

Does it actually work and give an accurate report on a players ability? It’s a little to early to tell, but as discussed, the system is deliberately setup to favour those with more battles in higher Tier tanks, so make your own judgement. What’s also clear is that a player with a high WN7 or PR rating doesn’t always have a high WPR.

Understanding stat padding

The age old holy grail for developers- at least on terms of collecting player statistics – is a rating system immune to abuse. It’s fairly easy to pad statistics in Efficiency, but it’s also easy to see where it has been done. For instance, you can now look up players on numerous statistics websites and see what their average statistics are – you can tell if a player’s 2000 Efficiency score is from playing Tier 2 games and acquiring cap points.

Despite the protection built into WN7 and the sliding minus scores for tanks played under Tier V, it’s easy to exclusively play Tier 1 or 2 and end up with a WN7 score exceeding at least 1800, if not higher. You can see below how WN7, despite its built-in protection, can still be abused:

Although the average damage will be low, survival rate, win rate and number of kills per death will all be extremely high. Once again, a quick look at a players stats on a site like noobmeter.com will tell you this information. Performance rating has similar protection to WN7 in respect to Tiers played:

First penalty threshold:
clearedFromPenalties1 = 1500
expectedMinBattles1 = 500
expectedMinAvgTier1 = 6

Second penalty threshold:
clearedFromPenalties2 = 1900
expectedMinBattles2 = 2000
expectedMinAvgTier2 = 7

In some respects it’s unfair on those that enjoy mid and low Tier battles, but in its defence, it’s an effective way – compared with WN7 – of reducing stat padding through ‘seal clubbing’.

Really, the best way to pad WN7 and PR – and Efficiency to a degree – while keeping a higher average Tier is to simply be a better player in whatever tank you chose to play. Get more damage in, do more spotting, and most importantly win while surviving the round. If you look at most players with a good or great WN7 or PR, they’ll have a 35%+ survival rate, 55%+ win rate, and an average damage at least equal to the hitpoints of a heavy tank at the same Tier as their average played Tier. However it’s worth noting that Clan Wars and Company Battles may increase win rate, but on a whole they reduce most other statistics.

Statistics sites and tools

noobmeter

NoobMeter started out as the platform for the Performance Rating statistic, but has grown into something bigger: a one stop shop for World of Tanks statistics, including the other core 3rd party rating systems. Featuring tables containing a plethora of player data, the ability to view data for individual play sessions, interactive graphs along with Clan statistic, and a place to store replays, this site does pretty much everything. The graphic interpretations of the data alone are perfect for monitoring a players improvement, checking out those new Clan applicants, or crying in shame as you realise you’re terrible at World of Tanks and join the crowd of people who say statistics mean nothing.

Wotlabs

WoTLabs is similar to NoobMeter; it displays a number of statistics broken down into time periods, including Efficiency and WN7. It notably doesn’t include Performance Rating. While it’s a little light on features when compared with NoobMeter, it does offer signatures. If you’ve ever wondered where players on the World of Tanks forums get their signatures with the total and 60 breakdowns are from, it’s here.

Wotlabs also contains a community forum, blog, and a store with World of Tanks-related shirts and paraphernalia. They’re not all that bad either.

vBaddict – Performance Analyzer

NoobMeter has some pretty interesting peformance metrics, but if you really want to dig deep into your performance then Vbaddict is the place you’re looking for. It uses the Dossier Cache file, which is a file that sits on your local machine recording all World of Tanks data points during play. Simply upload this file to the site and it will spit out colourful charts and tables indicating trends and performance metrics.

If that isn’t enough, Vbaddict collates this data, along with similar data from other sources, to give a complete picture of World of Tanks player habits across all tanks and maps. The number of data points collected and displayed are staggering. In addition to this, it also contains leaderboards, map data, and a heatmap feature.

Exactly how did someone get that far into the lake?!

Wot-Dossier

Wot-Dossier is similar to Vbaddict: It requires you to upload that Dossier Cache file, and it will parse the information and present it to you in a simple format. Nevertheless, it’s a interesting and concise tool to use. You can see a number of statistics for each individual tank, from your average accuracy to the amount of time you’ve played in that vehicle. You too can feel somewhat depressed after discovering you’ve spent 4d 14h 38m 53s in a Lowe.

Wotlogger

Be a stat obsessive in real time with WOTLOGGER, a session based tool which is run while you play World of Tanks and updates at the culmination of each battle. There really isn’t a great deal to talk about here other than it shows you your individual battle statistics as you play. A fairly handy thing if you’re completing the new Missions, which neglect to tell you how many kills you’ve made.

In game statistics tools

XVM

Love it or loathe it, the eXtended Visualisation Mod has become a standard tool across the World of Tanks community. XMV is an all in one UI mod, which can be customised to show a number of different statistics, including those gathered in real time from the game client. It comes preconfigured to display WN7 scores and win percentage next to each players name, but can be customised to show any number of statistics through editing a small config file. In addition to this, XVM comes with some other neat features which we won’t go into.

XMV comes in many flavours, with dozens of different customisations. My personal choice is the OM Mod Pack installer, as it’s fully customisable on install and is packed with some excellent third party mods on top of XVM.

While you personally might not give two hoots about statistics, a vast collection of World of Tanks players do. So while you may not care what your PR or WN7 is, it’s something to bear in mind if you’re wanting to participate in Clan Wars, perform a bit of self improvement, and maybe find out something before issuing orders/raging at other team members while in the online battlefield. Nothing gets quite the same reaction as a dead player with 500 WN7 telling their team how to play.

This article originally appeared on TheMittani.com, written by knobber.