Greedy Goblin

Tuesday, November 22, 2011

Helping M&S kills you and you don't even help them

Tobold posted his World of Tanks results:

If we consider the draw matches to half win, he won exactly 50% of his matches. Based on this data he consider this game perfectly balanced. He is wrong.

While I don't agree Tobold much, I consider him an intelligent man, so him playing WoT made me accept that WoT is not another dumb shooter that you outgrow after the (mental) age of 20. (Note: I did not call every shooter dumb, but most are). So I tried out the game, read the manuals, did the proper meta-gaming and here I am:

If we consider the draw games half-wins, I won 62% of my matches. Also, my survival chance is more than 2x higher than his. Since I survived 28 out of 64, I died 36 times. I destroyed 58 enemies, so my kill:death ratio is 1.6:1. Also I did not even pay a single euro for this microtransaction game, unlike him.

Wow I'm awesome! I mean I must have "ub3r l33t skillz" to make such difference all alone, without money-trained crew, without money-paid upgrades, without the shamelessly overpowered money-paid ammunition. On the top of that, I played a tier 3 tank that was on average the 10.8th position out of 15 on the power list (the start screen lists tanks due to tier). All hail me!

Or not. I'm clearly having no better dexterity than Tobold, and considering the low amount of battles, it's obvious that I am just learning the game. I think I can get better, and even 2/3 winrate is reachable. The difference between us is not in playing, it's about thinking about the real world. He thinks: "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." He sacrifices himself for his team. Very noble. Totally stupid though. Not because of his low survival rate. I agree that win is a win, and I would prefer a win where I die first over a loss where I'm the last man standing with 3+ kills. His sacrifice is stupid because it's clearly not getting wins. Despite being only 1/15th of the team, I can increase the win chance from 1:1 to nearly 3:2. So could he.

His problem is that he is sacrificing himself for morons and slackers. Scouting enemies worth dying for in a clan match. You lose your light tank and the enemy loses several heavier tanks due to artillery. But in a random map you helped mindless kids who don't know the basic game mechanics so can't use the help you gave them. The "meleeing" artillery and the zerging tank destroyers won't take out the heavy tank you spotted at the cost of your life.

My view of the World is that everyone that haven't proven himself a smart person is considered a useless idiot. I planned my playing from this, and it clearly worked. After learning the game basics on the tier 1 and 2 tanks, I played exclusively a tank destroyer. Tank destroyers are less mobile than other tanks (except SPGs) and also has less armor on the sides and rear. But they have a good gun and their purpose is "taking out enemy armor from long distances completely unseen". That's what I do. I hide in the back of the map, near our base. I don't have to wait long before a "Rambo" shows up. A light tank, rushing trough the lines to take out the artillery or just "have fun". He either stops to shoot at the arty or runs a predictable circle around it. I take that out. Unfortunately for my kill statistics they have the nasty tendency to have 10-20% HP left after my first shot so the AoE of the arties defending themselves give them the killing blow. I swear, with money-paid ammo I would have 3:1 kill:death. Anyway, the point is that I actively participate in killing them. Then I keep hiding. The randoms will attack unevenly so on one lane we win, in the other lane they do. Our winners approach their base "defended" by arties only. Their winners walk directly into my gun. Not only I can take out one or two, but also I slow them down and illuminate them for the arties. Alternatively I let them capture the base up to 80% when I shoot, breaking the capture bar and may kill one before going down, giving time to our front guys to capture their base.

The above, well-working strategy is based on the fact that our randoms are not worth helping. They are fodders, and should be considered as the mobs on League of Legends. Their only purpose is to march forward mindlessly, taking out enemy mobs while providing me cover. If anyone would say "the key to win in League of Legends is defending mobs at all cost" (same as "the key to win Alterac Valley is protecting Balinda at all cost"), everyone would consider this person a total newbie or an idiot. In World of Tanks, the very same thinking is considered "teamwork" because they are people. Well, I don't care if they have feelings, if they love and are loved, if they cherish watching the sunset holding hands with one another. If it moves like a mob, shoots like a mob, ignores messages like a mob, it is a mob.

The point is not the standard goblinish "use other people as tools for your own end". The point is "using dumb people as tools is good for the whole group". At the end of the match, the dead random has the win bonus credit, just like everyone else. He also had some use on the map, even just as being a decoy for me. Helping them doesn't give them win because by helping them you sacrificed the most valuable resource the team had: yourself. Exploiting them gave them win, as they were utilized for maximum team effect.

The same in real life: giving the M&S aid doesn't help because it decreases your resources. Exploiting them in our factories and mines help them as it provides the whole society GDP.


Happy Forum said...

I agree. In League of Legends, you should only sacrifice yourself to save someone if you know they can capitalize on it now or later in the match.

Rohan said...

There is another possibility.

Presumably, World of Tanks uses a rating system to match people. These rating systems take some time to settle at the correct rating.

In chess, the rating for a new player is referred to as a "provisional" rating, until a certain threshold has be reached, and the system is certain that the assigned rating is correct.

Tobold has some 3,600 games played, and as such his rating should be perfectly identified. You only have 65 games, and the game is still determining your rating. Because you are still in the provisional phase, the rating you have is too low for your actual skill, and thus your win percentage is higher than 50%.

I would keep track of your win percentage as you play more games. I would expect it to converge to 50%, as your game rating converges on your true skill.

latusthegoat said...

As someone above me said, you need more games played before you can draw any statistical evidence from your results.

Also, Tobold showed ALL games, you are looking only at your favorite/best. I can pull out a tank with around 75% win during 80 games to show how leet I am... But overall all my tanks come out to around 50-50 in 1500 games played.

Something to note as well is that I have never seen the total m&s you talk about in the higher tiers. Maybe its just part of the low tiers with easily accessible tanks that go really fast and attract herpderp behavior...

madscorpion said...

I'm sorry but you have absolutely no idea on WoT game mechanics.

1) The matchmaker puts you in your first 50 games in matches that you're likely to win. This is to allow you initial fast xp/credits so you can research the upgrades for your tanks fast. You've played just over 60 games. It's very customary for me to score 60%+ in my first 60-70 games even though my average winrate is 53% (which is considered pretty decent btw)

2)What type of tank you play affects your stats a lot yes BUT playing scout is far from a stupid idea (if you play scout properly that is - the blind suicide rush is NOT scouting) A good scout WILL win the game for the team.

3) The type of tank you play affects game outcome a lot. A good scout will change the game if the artillary isnt TOTALLY stupid. A TD will very rarely affect the outcome otoh (my jagdtiger has a 45% winrate despite awesome individual performance, while my heavy tanks all boast over 53% winrates with much worse individual stats- it's because i have a greater effect on the outcome in a mobile, turreted tank than i have with a TD.)

After playing over 8000 games, i can say that you dont know how to play a tank until you've played well over 150 games with it (and i've played a lot of different tanks)

There are merits to the points you bring up in your post. WoT also suffers from a high number of M&S. It's customary for me to play awesome on my side of the map while the other 14 people get rolled over by much crappier tanks. It happens. But despite a horrible matchmaker the WoT game is a lot more balanced that people give it credit for. Individual "every man for himself" attitude is prevalent in random matches and your success depends more on how many idiots there are in your team. But it also rewards team play and a single 3-man platoon out of a team of 15 can totally run over a team that doesnt do teamplay (me & my friend have a screenshot where we had 12 kills)

TL:DR: While i follow & agree with most of your posts, please refrain from posting about WoT - you need a LOT more time with it to make conclusions.

Anonymous said...

Rohan may be right. Knowing nothing about the game but what you've said, I strongly suspect that you really can't be the only player who's figured out this strategy.

Dalrian said...

Now, I have never played this game myself, but since this blog is often about WoW, I'll just assume that the things you said also apply to WoW.

You say that dying for your team is a bad tactic. Often, I'd agree, but in WoW that's definitely not always the case.

It can be very beneficial for me (as a Rogue) to stealth up to a base and attempt to ninjacap it (for example AB or BoG). This is basically suicide because in half the occasions they catch and kill me, but in the other half it's a successful cap. Therefore, it's worth the risk.

Other examples are simply slowing the enemy down with AoE spam whenever it's needed (WSG/TP, flag carrier is running away, you jump in to slow the 10 Horde chasing him on mounts).

Sure, you shouldn't always sacrifice yourself, but there are definitely a lot of occasions when "suicide", is the best option for your team and sometimes, it can win you the battle.

Other than that I do completely agree that Random BGs, or any other way of playing with randoms, are mainly won by 1 player deciding the entire battle.

One player can have a very big influence on the outcome of a game, which creates a difference in winrates.

If only the forums understood that, then I'd have a lot less work to do.m,

Anonymous said...

rather bleak outlook in the last paragraph, tough essentialy corect.

but is this not like most BG strats ?
find a task important that everyone ignores (Gys on Sota fo instance) and defend capture ?

Bworf said...

Basically what madscorpion said, except that a tank destroyer such as the MarderII Gevlon posted stats from can most definitely definitely affect the outcome of a game. MarderII and Hetzer especially since they both have guns that can take out higher tier enemies while still in stealth mode. Hetzer also has awesome sloped and thick front armor while the marder has better accuracy.

But the comparison of one tank's stats with Tobolds total stats is wrong for one more reason: Some tanks you barely even bother to upgrade because they are tanks you have to drive for some games in order to get to a desired higher tier tank. You will likely have pretty crappy win stats for these tanks even if you play well, which will of course affect your total score.

My worst tank has something like 44% wins over 100 games. My best (Hetzer) has 63% wins over 60 games and my total wins are 54% to 44% lost.

Lastly, the killed per game total depends a lot on what type of tank you play and its matchmaking. A Leopard light tank will very frequently be placed in matches where it can not even damage 80% of the opposing teams tanks. Imagine what that does to your kills/game ratio if one of the tanks you drive is a Leopard.

However, I still agree with your conclusion.

Killan said...

That's why I think that playing healer in random BGs is plain stupid and should be avoided.

Playing healer only results in random idiots getting more HKs in the mid, and not winning a BG, since random idiots usually don't play for objectives. Most of the time I manage to heal twice as much and the best alliance healer - and we still lose. So what's the point?

Another reason why playing healer, i.e. helping idiots, should be avoided is lack of versatility of healers in PvP. As a druid, when you play feral, you can get enemy melees from your healers, you can grief enemy healers with heavy CC and interrupts, you can carry flags and defend bases or go bear and tank 3-4 dudes in a heavy zerg. In the end, you can simply go and kill some enemy terribad players and have fun kicking their butt. When you play healer, you can only heal. If your DPS dudes are terribad, no matter how much you heal, you'll still lose. If you play DPS, there are options.

Clockw0rk said...

The problem here is that this would be no different from someone playing battlegrounds at level 60 in WoW then declaring they fully understand the PvP in the game...some things are going to change. Keep in mind that most of Tobold's commentary is from his much higher tier tanks where the game does change. For example at low tiers the enemy SPG's can't one-shot as many things nor can they hit across the map. This higher tier games a scout is invaluable to the team's success, even if he is likely to die early.

Gev's Panzerjager has never seen anything bigger than a Hetzer if at all (as he is presumably playing on the European servers where the tier spread per game is supposedly much smaller as it is where I am on the USA servers).

Playing a TD then deriding your teammates as idiots is like playing a sniper in battlefield then laughing at the assaults because they're dying on the point. You might have a better K/D than they do but they're the ones winning the game. Each tank has a role...except the M3 Lee, that thing's a piece of crap.

Lars Norberg said...

Whether or not Gevlon understands the mechanics in World of Tanks is to me a moot point. Sure, the statistic material is too narrow to be conclusive, I won't argue that. But it doesn't change the basic philosophy behind the post.

Sacrificing yourself based on the assumption that it will benefit the group is only a smart move if the group actually follows up on whatever reason you sacrificed yourself for. Unless you know for a fact that the group consists of smart players, this is in other words a gamble. And gambling is not a good way to win a fight.

I support the strategy of "wait and see" when it comes to this. There is no point sacrificing oneself for any sort of objective until you know for a fact that it will benefit the group in total. You are of no use to anybody dead.

When playing any sort of game with a randomly selected group, you should always analyze the group before making any sort of strategic decisions. Anything else is a waste of time, and an invitation for major headaches.

Phelps said...

Madscorpion is right. First, your sample size is WAY too small. Get over 1000 games so you actually have a sample that is predictive.

Second, you assume that the skill of your teams players are the driving factor behind wins. It isn't. The driving factor is how bad your opponents are. WoT punishes mistakes much more than it rewards skill. Winning consists almost entirely of having fewer m&s on your team. That is why platoons significantly effect win/loss -- it isn't the skill of the platoon as much as making is 15 potential terribads vs 12 potential terribads. And that is why even platoons only push win/loss up into the 55-60% range.

If you want to know how skillful a player is, you don't look at win/loss, you look at xp/battle. Otherwise it is the same mistake as rating a WoW dpser on gearscore instead of DPS.

Phelps said...

Sacrificing does make sense in WoT for a couple of reasons. First, a win garners you 2x experience. Second, you don't get xp only for damage done. You also get xp for first spotting a tank, and you get half the xp for damage done by someone who is spotting their target through you (long range shots and arty).

So a scout actually gets decent xp is they are doing thier job, and taking a shot for someone else after you have done a good amount of damage makes tons of selfish sense, if they get you the 2x bonus.

maxim said...


All you said is true, as long as you accept that M&S is not doomed to remain M&S forever and can actually become as useful as yourself, should he choose to.

Naturally, higher levels of play in any game seem to have more of these reformed ex-M&S. Heck, i'm pretty sure you yourself is one of those.

Anonymous said...

I agree with you on philosophy and the above commenters on the mechanics.

If you don't know anything about them, just assuming your fellow players are M&S is quite reasonable and your behavior should be planned accordingly.

But if there is a rating system, unless you are really world class, then your win rate will be around 50%. Its difference will be random numbers and the effectiveness/speed of the rating. But people with a 66% rating past the initial will not win 2/3 of the time against 66% people.

Bworf said...

Average xp per battle is not really a good measurement of skill either unless you factor in tier as higher tier tanks gain more XP for their efforts simply because they can hurt/spot higher tier targets.

Phelps said...

That's not my experience past 2000 games or so, Kaj. I make as much experience on my elited tier 5-6 tanks as my T34 or Patton. (and more than I make on my Tiger).