You Are So Predictable

By Alec Meer on December 8th, 2007 at 4:37 pm.

The latest creepy output from Valve’s ever-watching stat-o-tron reveals how we, the human animal, play Team Fortress 2. And, well, everything you’ve ever moaned about during that one match where it was so unfair because they had too many of this and they did that and God, I might as well just stop playing if you’re going to be like that is, as it turns out, pretty much true.

It’s mildly distressing how absolutely it confirms everyone’s suspicions and stereotypes about the game. Yes, there are more rushy-bastard Scouts than any other class. Yes, campy-bastard Snipers earn the most points. Yes, hitpointy-bastard Heavies get the most kills. Yes, hidey-bastard Engineers live longest. And so on.

What’s slightly odder is the breakdown of which side wins most frequently on each map. BLU has the edge in every killing field except Dustbowl and Gravelpit. Why? How? I thought we were all the same! Damn you for not being neater, demographics.

The heat maps that reveal the most common death locations are a giggle, too. Here’s 2Forts:

Who’da thunk it?

__________________

« | »

, .

68 Comments »

  1. Andrew says:

    I quite like the Gravelpit one, which shows that quite a few people kill themselves in the spawn points. A few over-zealous demomen and soldiers attempting to farm the health cabinets, I guess.

  2. The_B says:

    I’m finding it fairly humourous that more damage is being dealt per hit by the Pyro’s shotgun than the flamethrower…

  3. Feet says:

    I cannot believe how few medics there are. Clearly it’s the most fun class. :

  4. Mr Pink says:

    This is really interesting stuff. I think it’s fairly obvious why Red wins on the Dustbowl/Gravelpit, since they are always defending, but those Blu advantages on the other maps are definitely statistically significant. I’d love to hear theories on why that is.

  5. Mr Pink says:

    B: I’d imagine that’s because a “hit” with a flamethrower is counted as a very short tick, since it’s effectively a constant fire weapon.

  6. Theory says:

    I’m finding it fairly humourous that more damage is being dealt per hit by the Pyro’s shotgun than the flamethrower…

    The flamethrower “hits” something like twenty times a second though. ;-)

    What’s really shocking is the dominance of the rocket launcher. The soldier’s popularity isn’t nearly enough to explain it.

  7. terry says:

    where are stickies on that list?

  8. Seniath says:

    I’m curious as to how the fire axe has a longer range than all the other melee weapons o_0

  9. roBurky says:

    That says there are no stalemates on hydro. Something’s gone wrong there, because it’s almost always stalemate when I play it.

  10. adambaum says:

    I’m a nerd, and I like stats… but I’m always amazed how many people fail to be annoyed by this level of data gathering from a gaming company. No, it’s not personal information, credit card numbers or things like that but still, it’s invasive.

    Everyone bitches and moans about things like rootkits, DRM and Starforce yet when Valve decides to follow all it’s players around in game like a stalker watching kids at the playground, we’re ok with that… hey, it’s Valve.

    Right?

  11. Matthew Gallant says:

    The red heavy in that screenshot is a spy! J’accuse!

  12. Theory says:

    Valve decides to follow all it’s players around in game like a stalker watching kids at the playground

    You’re weird.

  13. Monkfish says:

    I cannot believe how few medics there are. Clearly it’s the most fun class.

    It’ll be even more fun soon, as Valve will be adding loads of Medic-related achievments (assuming, of course, that attaining achievements float your boat). It’s great that Valve are acting on these stats in this way.

  14. Mr Pink says:

    adambaum: I don’t really think this is a problem. What possible use could they have for this data, other than improving the game?

  15. Steve says:

    Seniath: If you watch a Pyro attacking, he holds the very bottom of the handle and chops straight down.

    adambaum: You’re right, it’s not personal information. Plus, they are releasing it freely – so I for one know I’ll be playing differently because of this. This is helpful data collection, not creepy dangerous data collection.

  16. Monkfish says:

    No, it’s not personal information, credit card numbers or things like that but still, it’s invasive.

    Nah – this kind of stat gathering I embrace. There’s nothing nefarious about it, and they’re genuinely using it to improve their games. Valve have been pretty up-front about their stat-gathering since HL2:Episode 1.

    And it’s hardly invasive. I mean, it’s not like I can feel Gabe Newell’s beady eyes watching my every move. It’s not like CCTV, or anything…

  17. Joel Esler says:

    It’s interesting to see the edge BLU has in what are functionally symmetrical maps. Could the different colour schemes affect the psychology of players and their willingness to storm through the enemy fort?

    Consider, Red construct their fort out of wood and hay. Clearly, they are idiots and shall be crushed like bugs.

  18. JP says:

    Everyone bitches and moans about things like rootkits, DRM and Starforce yet when Valve decides to follow all it’s players around in game like a stalker watching kids at the playground, we’re ok with that… hey, it’s Valve.

    I’m not really grasping your analogy with rootkits and DRM.

    Also I thought it was pretty clear that anything you do in an online game is equivalent to a Public Act, like walking to a shop down a street with CCTV cameras.

    When you play a multiplayer game, you’re installing client software (probably on Windows no less), connecting to Valve’s master server, then connecting to some random person’s server. There’s not really any step of that process that’s secure or private.

    If your stance is just hardline against any and all public data collection, that makes sense… but if it is you are probably going to want to abstain from all online gaming.

  19. The Unsung Hero says:

    Has anyone else noticed the security cameras on walls in game that follow you around? I always wondered what they were there for….

  20. Rockeye says:

    Maybe it’s an issue with with the colours themselves. Red is traditionally a danger/enemy colour in nature, everyday life and a lot of games. Possibly this means the hostile reaction to the red team members is more immediate, people identifying them as a threat more rapidly. Blue is often used as a colour to denote things that are safe, so maybe the response to shoot them is delayed?

    From what I remember, the eye is pretty much equally receptive to the blue and red parts of the visible spectrum in bright conditions (it’s more receptive to the yellow/green part) so that shouldn’t make a difference. Maybe red is more of a contrast to the background in these maps, so the red team are more obvious?

    Or maybe Valve have balanced things against those damn commie reds.

  21. Lacero says:

    Maybe more people click blu than red when choosing a team? Or the auto balance puts odd people into blue first.

    Theres no stats for “average number of people on a team over time” so we can’t tell, but I think that’s all it is.

  22. roBurky says:

    I don’t think the blue victory is going to be down to any differences between the mechanics of the two teams. I know a lot of people always pick a particular colour team. It’s just going to be that those who always go blue happen to be doing better than the people who always choose red.

  23. terry says:

    Rockeye: I was wondering that myself earlier. I tend to spycheck compulsively when red… anyway

    The stats are very interesting to me as a player and I daresay more than helpful to Valve in tuning the game. I think that also these stats serve as a way for software companies to further finetune/outright change frustrating areas, and guinea pig new ideas.

  24. King Awesome says:

    People playing red would get less practice on a standard map cycle as on the defensive maps their respawn times are longer.

    Clearly this is the total and sole reason for the discrepancy. I don’t know how you can’t all see it.

  25. Hump says:

    I wouldn’t know, I play Fortress Forever.

  26. Stick says:

    Heh. The four classes with the shortest average lifespan are my top four played. In exactly that order.

    Not sure what this tells me.

    Anyone else confused by syringe gun having lowest average distance?

  27. malkav11 says:

    Not really. I expect that a lot of syringe gun usage is a medic turning around, going “oh shit, I’m under attack” and spraying syringes wildly. (Not to say that it can’t be effective under other circumstances, but I for one am generally healgunning as a medic. For evident reasons.)

  28. roBurky says:

    I wonder if maybe the syringe gun and fire axe are somehow swapped around on the range chart.

    I can’t think of any time I’ve used the syringe gun in melee range.

  29. Bulk says:

    I’m guessing a bunch of people have managed to hop the train, and died all the way out there

  30. c0wb0ys7y13 says:

    I can tell you why blue wins more than red… the maps arnt symetrical. Look at grainery. The last cap point on grainery is totally different for each side. blue has an extra pipe running through their area. and on 2 fort you can see that the objects outside the base are different. I use the reds barrels and boxes as cover to hide despensers.

  31. Zeh says:

    Great stats.

    Sometimes on symmetrical maps Blu might win more then red simply because it’s the “easier” to join. It sits on the middle and if you’re joining a map and doesn’t care about the team, it’s likely you’ll end up on Blu. It just shows that the “random” option has to be more prominent.

    Similar stuff goes for the difference in deaths on symmetrical maps, specially on spawning points. That’s specially visible on 2forts above. My theory is that some people join Blu by mistake, simply because it’s the easier option, then quickly return to the team selection screen and picks red (causing a Blu death on the spawn point).

    It doesn’t help that everytime the team selection screen is shown the mouse is reset to the middle of the screen (which is where the Blu selection door is). I’ve more then once selected the Blu team by mistake simply because the mouse position was reset.

  32. Radiant says:

    Or there is a ten percent variation either way in which case it is fairly even.

  33. Radiant says:

    Oh and the syringe distance is because you only use it close up.
    You either have the healgun out all the time with a ‘backup’ unit killing people or you’re in an ‘oh shit’ moment where you’re face to face with someone trying to splat you.
    You and the syringe gun.

  34. Radiant says:

    And at that distance you’re gonna use your bonesaw anyway.

    [sorry for the 3 posts]

  35. Jeremy says:

    “quite like the Gravelpit one, which shows that quite a few people kill themselves in the spawn points. A few over-zealous demomen and soldiers attempting to farm the health cabinets, I guess.”

    Dunno about you guys, but after winning the round, I often have to chase the losers back into their spawn point to finish them off. That’s probably why.

    As for the big brother thing, I think we agree to it in the EULA, so not much can be done. If you don’t like it, don’t play, I guess is the only option. I for one think it’s fun.

  36. Jay says:

    2Fort heat maps just tell me that there are too many idiots who insist on trying to go through the courtyard where there’s ALWAYS a sentry on the walkways. I usually rocket jump onto the platform, or failing that, through the grate

  37. kadayi says:

    Hump says: “I wouldn’t know, I play Fortress Forever.”

    And you are whom exactly? I’d like to know so I can remember not to care in future.

    As for the comment about the stats being some kind of invasion of privacy, technically yes..but until such time as Valve start posting up hilarious and humiliating footage of individual players being owned relentlessly going ‘look at this tard’ youtube style I doubt anyone will be complaining.

  38. yns88 says:

    I’m actually rather surprised at how unbalanced the invasion/defense maps are and how Hydro seems to be the most balanced map there is.

    From what I know, many people have complained about control points in Hydro being off balance, though the mere fact that Blu and Red are constantly changing places probably puts that off (oh, and the reason why they put no stalemates there is because a round of Hydro is never “over” until one team completely dominates the entire map. It’s the same with Gravel Pit and Dustbowl).

    It’s still kind of odd though, since many people consider gravel pit and dustbowl to be their favorite maps, and apart from fast respawn servers (which obviously off the balance) I have not seen any greater difficulty attacking than defending. Perhaps it’s due to people who don’t know the game too well and get owned by defensive engineers or something.

    It may be true that the “symmetrical” maps have slight differences, but I only really just noticed how different they are when I looked at these maps just now. In any case, the difference certainly isn’t very noticeable. What I’m thinking is that when you join an evenly matched server, your mouse pointer defaults to the Blu door, so if you want to join really fast you’ll probably go Blu.

  39. Chris says:

    “2Fort heat maps just tell me that there are too many idiots who insist on trying to go through the courtyard where there’s ALWAYS a sentry on the walkways. I usually rocket jump onto the platform, or failing that, through the grate”

    Sure, but only demos, soldiers, and scouts can avoid the courtyard. And if you want to bring a medic with you, you have to go through that room.

  40. Arathain says:

    I think what a lot of players fail to realise about the syringe gun is that it’s actually got one of the best effective ranges in the game. Once you get the hang of the arc you’ll find the spread is actually quite small (compare it to the functionally similar Sniper SMG), and it benefits nicely from that large clip size. I think the stats reflect play styles rather than the capabilities of the gun. I also think this will change when the medic achievements are released and we see a lot more people learning some offensive medic-ing.

    That said, it’s a lovely circle strafe weapon which can really catch players by surprise.

  41. Spacegirl says:

    I am nearly positive that the Fireaxe and Syringe Gun stats are switched there, it really doesn’t make any sense otherwise.

    Also, I think the Blue edge in the symmetrical maps is nearly statistically irrelevant. Although, It is interesting…

    And Defense is easier on Dustbowl, especially against a bunch of noobs. I really only play on 2 servers, both with very high caliber players, and on them it is VERY hard to win gravelpit on Defense, and dustbowl is not that slanted in favor of D. I’d hazard a guess that it’s 60/40 Defense / Offense win ratio. Tough to say though…

    Most of the other stats are pretty obvious and completely in-line with their class roles. Scouts capture points the best. Soldiers are the main offensive class and thus blow the most people up. A heavy and a Medic are going to completely mow down any garbage team and by the same token a good Sniper is going to completely dominate any noob team. Especially since the top 2 played maps are 2fort and Gpit, the most sniper friendly imo. The Pyro is garbage…

    The One thing of interest is the relative lack of demomen high up in the stats. I suppose they aren’t very noob friendly, but if you have ever watched high-level demo play, they can be the most dominant class.

    The most interesting things to me were how much damage the fist and wrench do lol. nerf wrench!!!

    Space, Out!

  42. Spacegirl says:

    O and to the poster above me, I agree about the syringe gun! It’s pretty good!! It has a big clip and is pretty easy to keep directly on some1 while you run around. It’s better than the pistol and smg 2 b sure. Medic is probably my best class, and I have a lot of success with the syringe gun.

  43. Morte says:

    “The Pyro is garbage…”

    Them is fighting words Spacegirl (if that really is your name). Used effectively the Pyro can be devastating, especially when defending, there is no better sight in TF2 than seeing your enemy scatter as you jump into the middle of them spreading your cleansing flame.

  44. Morte says:

    Oh and laughing maniacally is mandatory at that point.

  45. Stick says:

    Pyro: it’s not about the kills, it’s about the burning and the screaming. And the awesomeness of the firebug isn’t really reflected in the stats. (E.g, on Dustbowl, ambushing the offense interrupts their momentum and wastes their time – even if you don’t manage to get any points.)

  46. MacBeth says:

    My main habit is ‘Chaos Pyro’ – i.e. disrupting attacks before or just as they gather and get their ubers ready. Forcing enemy medics to trigger ubers before they are really in position is a very effective technique for your team, as is physically pushing attackers back, exactly as Stick says. I’m resigned to the fact I don’t get many points for it. The times when the top right screen fills up with MacBeth (flames) other team are worth it.

  47. Lu-Tze says:

    Frankly, the Bungie stats for Halo 3 are much more irritatingly stalkerish because they don’t watch everyone as a collective… they watch JUST YOU.

    People can call up your profile and find out what maps you suck at, and your own personal heatmap so they can see where you like to hang out… it’s more than a little big brotherly.

  48. demm says:

    What surprises me the most is how many bad players there are.
    Look at the achievement page and you’ll see that only 60% have the ‘Hard to Kill’ achievement. That means 40% of all players have never killed 5 people in a row, which should be quite an easy task! (even if it’s pure luck, you just need to do it once to get the achievement)

    Only 17% of all players have the ‘Head of the Class’ achievement. I think that everybody who wants to be a good TF2 player should try to get this achievement, because you have to know the strengths and weaknesses of all classes in order to be one.

    One other thing that surprises me, is that so many players like to play engineers, even though it’s the least rewarding class (point-wise).