Tobold's Blog
Monday, October 31, 2011
 
World of Tanks balance

Above is my current overall result from the statistics screen of World of Tanks. Of 3,596 battles fought, I have won 1,748 and I have lost 1,748. For the math wizs among you who notice that there are 100 battles and 2% missing, those are the draws. My win loss ratio is perfectly balanced, and that after thousands of battles.

Recently a reader wrote me a mail complaining that he didn't see it like that. While his victories to defeat ratio was 50:50 like everybody else's (you can only derive in a major fashion from that by doing tank companies and clan battles), he was looking at the last line on that screen. He only counted battles as victories when he was still alive at the end of the battle. Dying early felt like a loss to him, even if his team won. He didn't care much for his team, they were random strangers to him anyway.

As you can see on my screen, my battle survival ratio is extremely low. That is because how likely you are to survive a battle depends strongly on the kind of tank you are driving. Artillery has a high survival rate, if your side wins. But personally I play mostly light tanks, nearly exclusively these days, and those tend to die in the first minutes of the battle. A scout simply isn't worth anything when staying behind and hiding. At most you might want to wait a minute or two before going scouting. If I manage a good scout run I might light up a tank or two for our artillery, or even kill an enemy artillery, before dying. But I next to never survive the battle. Of the 620 battles survived shown above only 4 were with scouts.

And I don't mind dying early with my scout in World of Tanks. I still believe that if I scouted well, I contributed to the eventual victory. Even if the people in my team are random strangers, many of which don't even share a common language with me (European servers), I still consider World of Tanks as a team sport. And I don't need to be always the hero, the most valuable player of my team for that to be fun. The obvious advantage of that is that I consider any win a win. And even if I played well but my team lost, I'm not unhappy with the result. I can only advise anybody who gets extremely unhappy when he loses or when he isn't an outstanding player in a win to stay away from balanced PvP games like World of Tanks. Losing gracefully is a lost art.

Comments:
Two thoughts occur.

1. If over time the outcome of a game is predictably even, if you can expect to win the same number of matches as you lose, is that a game? And if it is, is it a good game?

2. Is it morally acceptable for a team game to be designed to expect the sacrifice of a member or members of the team? If it is morally acceptable, is that morality affected when the game uses analogs that represent sacrifices made by real people in a real situation?
 
I love the new matchmake services that are gaining popularity these days.

Take SC2. After a few matches you will get a roughly 50% chance to win a game. Quite the difference with the olden days where you could either steamroll your opponent or get crushed.
 
How is progression in a light tank? You probably don't rack up many kills but you get to fight a lot more battles because you are one of the first tanks knocked out.

I played Tank Destroyers mostly. Their stealthy nature means they are often the last vehicles alive along with the artillery. I really enjoyed some epic end game battles with slowly moving TDs and even slower artillery hunting each other but of course you aren't earning any extra experience by extending the battle so long.
 
Hmmm another interpretation of the win/loss balance would be that, on average, your role in the battle is irrelevant. If you're better than the average player, that should help your team, resulting in a slightly biased distribution.

I'm at 49%/49% as well, 25% survival (I play arty more than lights).

BTW I'm having enough fun to keep playing, so to the hell with stats :)
 
If everyone is special...

In a well matched PVP game you're going to lose as many as you win, and if you're really awesome you might nudge that up to winning 52-55% of the time. Maybe Call of Duty type games kind of mask this fact by primarily presenting you with a k/d ratio that sort of disconnects you from your team, but that's pretty much how the numbers end up working. Also respawns allow you take successfully play a much high risk strategy that allows you to do crazy stuff. In WOT the insane kamikaze charge works sometimes... but if it doesn't you're out of the game. I know when I play COD my K/d ratio is the stat I look towards to see if I am playing well, and my team winning is a secondary factor.

In WoW you could dominate simply by having way better gear than the average BG player.
 
I have to agree with your reader here: there's a fundamental difference between losing in World of Tanks and in, say, a World of Warcraft battleground.

Even if you're losing in WoW, you still feel like you're contributing to your team because of the resurrection aspect. You can score some more kills while holding off your defeat for as long as possible. But in WoT, you can do nothing but stare at a Game Over screen or watch a movie of your teammates. A WoW analogy to WoT is random Arena. Little wonder that Arena is much less popular than Battlegrounds.

To me, the difference between victory and defeat isn't a number on the screen: it's how I feel after the game is over. A hard-fought victory is of course the best, but even a hard-fought defeat is much better than the feeling of dying to a dumb mistake in the beginning of the game, no matter what result your teammates achieve in the end. I don't have to be an "outstanding player", but I don't feel part of a victory when I don't perform a single action for the vast majority of the game.

There are plenty of balanced team PvP games out there without this aspect of permanent death. In addition to WoW, some games such as Starcraft II allow you to take control of your allies' units if all of your forces are destroyed.

I greatly respect that you're able to feel victorious even if you only play for the first few minutes. But I don't think that it's uncommon to not being able to do so while playing with a group of random players.
 
I'm going to have to agree with your comment of "losing gracefully is a lost art", especially in MMOs. About every 5th game you'll see somebody quit the game right when the timer hits zero. Usually because the "matchmaker sucks" or "there is too much arty". Nobody wants to adapt anymore. They want to just "pwn everything" without working for it.

I myself pull about 55% Wins 43% losses, playing almost exclusivly german armor over a period of 6000+ games. When I hear remarks about "russian armor is OP", "Russian Bias" or even complaints about Artillery, I just roll my eyes. I will be the first person to admit I'm not the greatest WoT player in the world. I only pull good stats because I'm willing to adapt to what I'm driving and what I'm facing. I don't have to "wtfpwn" to be effective.

People need to stop looking at their preconsieved notions of what a Light, Medium, or Heavy should be doing and look deeper at their tank. Yes, the Tiger is a heavy, but that doesn't mean you sit up front and brawl as it can't handle that well. It's a heavy in that it can carry a bigger gun with superior penetration and wicked accuracy to snipe at extreme range. My Leopard is wicked fast, but I probably won't last long slashing through the lines - so I'll sit quietly with my spotting mods and camo net on in this bush and spot the baddies for my sniping teammates.

It's not all about getting up front and slugging it out, the game has some subtle ways where the player can contribute if they are willing to look past their kill count and see it.
 
Wonderful post, and right on the money.
I have only been in about half as many battles as you (a relative noob, you see?), and the average won/lost is still tilted a bit for me, but I am hoping it will balance out in the longer run.
Some days, it seems like I cannot buy that "first win," though. Yesterday, my teams lost 12 matches in a row. Since I was playing a light or medium in most of them, I do not think my skill level could have affected the outcome so dramatically.
No, I agree that long-term, the game is probably well balanced. But sometimes, the coin lands "on edge" ... and there are some days when I seem to be in nothing but 12-2 or 12-1 matches (either for or against "my team") and those can hardly be considered "balanced."
I think what would be more balanced is if the matchmaker went for a more homogenous mix, when that is possible given the available players. Mixing and matching sometimes ends up with very peculiar teams (like six artillery on a side - boring!) or jsut completely un-winnable matches (like one side completely unable to penetrate the top tank on the opposing side).
I enjoyed your thoughts, and concur that overall, the game finds balance. In the short run, however, battles can be frustratingly lopsided.
Of course, so is life, but no one logs on to play that, do they?
 
I won't pretend I'm the best scout in the world (okay, maybe I'll pretend it) but my survival rate with the light scounts is very, very, very low while my win-loss ratio is very good and my XP per battle is high.

That leads me to agree that a good scout can make a difference even if you die early.
 
If you are playing a Tier 5 light, you should try shifting your playstyle. The german and russian models are survivable enough to make several passes if they don't intentionally suicide, and the Chafee frankly has enough gun to be better as highly mobile harassment than a suicide scout.

It's not unusual for me to have 3 or more kills on my Chafee.
 
I also feel that WoT is very well balanced and most people's stats reflect a fairly balanced win/loss ratio to supports this. I consider myself a team player as well because no matter what tank you drive you will need the support of your team to have any success. Given that, what you do to help your team is your primary concern...and if you do your job well you might just see your win/loss ratio tip to reflect it.
 
I expect people to point out that the role of scout is a bit different and staying alive is an important part as arties are not machine guns, I think the point you are trying to make still stands - most of the battles I've been in end with a "kill all" victory instead of taking the base and the victorious team rarely has more than half of the people alive. That means the average survival rate is well under 1/2 and people who primarily care about their own survival will rarely be happy with their results.
 
Hmmm another interpretation of the win/loss balance would be that, on average, your role in the battle is irrelevant. If you're better than the average player, that should help your team, resulting in a slightly biased distribution.

Somewhat. It's very rare to see anyone above 51%, though, and those are usually people who platoon (three man group) in nearly every match.

The thing is, the influence isn't symmetrical. Bad payers hurt the team much more than good players help. I see few people with really high win scores -- but I do see a bigger handful with very low win scores. I interpret this to mean that bad players can cause a loss a lot more than good players can promote a win.

For that matter, I think that is the real platoon influence. You are taking two of your teammates out of the random distribution for terribads. Because the team with more bad players is the one that loses.
 
To anyone who complains about down time, the solution is simple: play 2+ different tanks. If you die early and don't want to sit and watch the match unfold, leave the battle and play another of your tanks! You get to keep playing WoT and not be bored. As for Bhagpuss - it's a GAME. Is it morally acceptable to sacrifice a team member? Cut me a break. Anyone who answers "no" to that question should never pay chess, as sacrificing pieces is often integral strategy. It's an online GAME. No one really dies. And, guess what? In war, such sacrifice is built into scenario builds. If the Army plans a Mass TAC jump, their planners include x amount of causalties occuring during the actual jump. In the real world, such sacrifices ARE part of the deal. WoT is like most PvP pickup games (assuming you aren't in clan wars). If you get together with 1-2 good players and co-operate in the match, you can often siginficantly impact the outcome. As free games go, WoT rocks!
 
I just had a look at my stats and I also got 50% win overall, but my KV-13 that I played 20 matches so far has 30%.
So I assume the matchmaker variation must be quite big, which also supports streaks of very extreme matchmaking. (not to forget that a stock KV-13 is actually quite a liability in tier 10 matches for my team).

Another thing I realised in random matches is that coordination is extremely difficult, with team members that are afraid to get scratches in their newly painted tank and don't care a rats ass about a "random noob". Especially when you play spotter or the single medium tank in higher tiers, it shows.
 
If you are a better player. You are offset by a better player on the other team. The teams are fairly large, so odds are someone on the other team is fairly comparable to your skill.

The only time I think being a "better player" would have a recordable impact is if you always took over your team and gave them direction and they followed it. However, I don't know many people that are willing to do that every game.

In WoT equipement is even. Skill can very, but it is expected the distrubution of player skill would be a normal bell curve. And the teams are fairly large.

The 1 life per match also is a good balancer. High risk moves can't be repeated time after time. For example a solo rogue in WoW trying to capture the same tower 5 times. If the rogue fails 4 times but suceeds once that may still win the match. In WoT the first failure it your last.

1 Life also puts more focus on "luck", be it tactics or the RNG. AKA: Who lands the first shot, or breaks their gun, ammo blows, or when you cruise into the enemy camp are their 0,1, or 4 defenders?

Pretty much it is similar to how a real battle would be expected to finish. If you made two marine groups (from randomly selected marines) and put them into a war game in unfamiliar territory with identical gear and only communication via texting, assuming the skill level was equal across both teams, there is no way you can call the match ahead of time.
 
Just so you know it is possible to be better than 50:50 without tank companies or clan battles. I have an overall win percentage of 54% with 3582 battles, 1925 wind, and 1579 loss (technically a 53.74 win percentage). I have only been in tank company or clan battles in one tank, my t-54. If you remove the 438 battles, 236 wins, and 183 losses in my t-54 from my stats my win percentage drops all the way down to 53.72 or by .02%. I also have only platooned 3 times ever. I know a lot of people who are better than I and have much better win percentages without tank companies or clan war battles. Conversely, I do know a lot of people with less than 45% win ratio so people should not expect 49% wins and 49% losses without being mildly skilled.

I would recommend figuring out strategies and play to increase you survival rate in order to increase your win percentage. My most played tank is an A-32 which I got with the pre-order light package deal. I have 494 games played on it. 283 wins (57%) and 202 loses (41%). I have a 34% survival rate. The A-32 is very close to an A-20. In the unofficial Russian server stats the A-20 has one of the lowest win percentages of all light tanks. With my A-20 I had only 47 battles played (so still fairly susceptible to luck and swings in stats but) I had 28 wins (60%) and 17 loses (36%) and 32% survival rate.
 
How many battles do you feel it takes before your win% becomes statistically significant?
 
Similar stats on League of Legends myself. Very close to 50:50 ratio.

Although I do notice a difference when I play a tank in classic or Karthus/Rammus in Dominion.

These are strong choices that impact the game a bit. Yet, when you consider a 5-person team, you're gonna end up with 50:50.

With WoT, it is even more natural to be 50:50 considering the team size.
 
Here are my stats:

# Battles: 3268
Victories: 1647 (50%)
Defeats: 1554 (48%)
Survived: 951 (29%)

I don't play arty, just mediums, a few heavies and destroyers. My lowest tank is the Marder II (tier 3) which I've kept for sentimental reasons and because it's shitloads of fun, while the highest is the tier 7 Tiger (which I also love). I don't use premium so I have to work hard for my credits.

What people fail to realize is that in WoT it's not just about killing the most tanks and surviving the most battles. It's not even about winning a large number of battles.

While I'm not a big expert myself, I took the time to do a lot of reading on the forums and the wiki so I have an idea of how the mechanics work and how the game rewards players with experience and credits.

IMO, the highest goal (besides bragging rights) in WoT is to maximize your XP + credits gain in each battle. This is not always done by surviving a battle or killing more enemy tanks. For example, actual damage done plays a greater role than the number of kills. That's why I laugh at those poor sods who sometimes whine "U stole mah kill!!1!" when you kill an enemy with 1% life. Actually dealing the killing blow to a tank does give you a little extra XP but it's very little compared to the damage that was already done to that tank.

Being on the winning side also gives a bonus to all the tanks that won the battle and those who survive receive yet another bonus.

So how do you best maximize your earnings? Deal lots of damage (or scout enemies and survive long enough for those enemies to be damaged by your team), win the battle, survive. Kind of in this order. If you deal TONS of damage and then die and your team loses the battle, you are still rewarded nicely, but even more so if they win and/or you survive.

WoT is a team game. You can't succeed on your own and a team effort is required for the best "spoils of war". A 50% win/lose ratio proves that, all things considered, the game is actually pretty well balanced. Of course, it takes a large number of games to reach this "perfect" ratio.

Now, if you look at individual tank performances you might be surprised. I'll give you an example: My Marder II (tier 3) has won 62% of battles, lost only 35% and survived 45% of them. Why? Well, not to brag or anything but I'm pretty good in the Marder and the tank is also pretty much maxed out in terms of capabilities with the crew operating at top efficiency. A very good player in a good tank can change the tide of battle. My highest number of kills in one battle is 9. You can imagine that if a single guy kills 9 adversaries, his team is pretty much set to win.

In the end, and despite all the naysayers, I'm pretty happy with WoT's balance. The matchmaker is sometimes frustrating but for every shitty battle there's also a satisfying one and occasionally an amazing one.
 
To Bhagpuss - World of Tanks simulates war. In war, people die. The aim of a commander is not to keep his people safe, but to accomplish the mission with the minimum of casualties. That's a very different proposition from 'zero casualties at any cost'. Probably the highest praise for a commander is that he doesn't expend men's lives unnecessarily, but to achieve the mission sometimes he will have to put men in danger, or even into situations he knows they cannot survive.

as for the morality - if it helps with that, everyone in World of Tanks is a volunteer who has freely signed up to be shot at. Unlike the real battles of WW2, not even the Soviets here are conscripts with NKVD battalions waiting behind the lines to execute anyone who doesn't advance fast enough :)
 
Tobold, a scouts job is NOT to rush ahead of everybody else just to get mauled by the incoming zerg in the first 2 minutes of a game. A scout that dies early is a useless scout. Rather wait a few minutes for the battlefield to spread out and THEN go scout because that gives more time for the enemy arties to be unprotected and less chance for you to bump in to 10 tanks... also, don't use the FOTM paths... :)
 
That's why I said to wait a minute or two before scouting. If you wait too long (5+ minutes), you risk arriving after the medium tanks who broke through elsewhere, or arrive after the battle is already decided.
 
The whole idea of "scouting" has been perverted by the WoT community. Every tank is a scout. What we think of as scouting (spotting) is really the job of the mediums anyways.

When we say scouting, we really mean deep strike. The problem is, too many scouts forget the "strike" part of it and just concentrate on deep. They put themselves in situations where the best they will be able to do is trade for an artillery -- if the friendly arty can get on target in time, which relies on the OPFOR nubbing it up. Against a competent opponent (granted, rare enough in pubs) a deep run like that is useless. You disrupt their movement for a little while, but it is rarely worth the tank.

The tier 5 lights have amazing guns for the tier. The are comparable to tier 6 medium guns. The top-tier scouts should concetrate much more on the strike part of a deep strike.

The Chaffee in particular is suited to this. I almost never go running off into their base until we have the midline secure. Until then, I am better off on defense harassing the other team. Once we have the midline, the heavier tanks will be in place to support me, and then I can deep strike into their base, and put some shells up the behind of multiple pieces of arty. (Being situationally aware helps, too. With multiple arty, I almost always try to put the one I am attacking between me and the other arty. I've forced a TK many times that way.)

Like you said, I don't mind dying early. I'm fragile, and if they pushed through and tagged me while I'm on D, they got me. I don't like dying early for no other reason but in the hopes that maybe our arty tags something before I go.
 
Personally I think that MM has always been based on keeping players playing. If your doing well today tomorrow you need to be punished Bounce more shots be on a team/Teams that are all underachievers and lose 7 to 1 ratio. What doesn't come into the picture is Kids. I have been on teams where there jumping Tanks for Distance while the team is being beat Second the whole thing Should be based on your XP alone nothing else you earn your way out of the crap games and into the good games tough games close and strategic battle .Not like it is now constant 15 to 2 losses orWins
 
Post a Comment

Links to this post:

Create a Link



<< Home
Newer›  ‹Older

  Powered by Blogger   Free Page Rank Tool