Gentleman Jon — As the hive mind of European TF2 turns to the near future in anticipation of imminent new seasons one man hunched over a computer console in a darkened room, not watching sick frag videos or young ladies disporting themselves but instead gazing at reams of log entries, senses the winds of change and thinks “not yet”… Let me lead you by the hand (or drag you by the ear) back one last time to i49, back to that weekend, back to the Greatest TF2 Tournament of All Time…

Player stats are the main focus of most TF2 analysis but their performances all take place on the broader stage of the maps, which have their own characteristics to get to grips with. In this article I’m looking at who thrived and who choked on the maps played at i49, and through a meandering path of kills and damage I reveal how the teams fared head to head.

So without further ado a look at how the medics fared

Medics: Deaths, Ubers and Drops

4- medic by map

The uber drop count per map uses the bottom axis, deaths and ubers the top axis.

Viaduct is a little difficult to compare as the map times tend to be different, but it clearly has a very different feel to the other maps.

The medic is far more likely to drop uber than any other map. Popping or killing the medic and protecting uber isn’t any more important than on a cp map really (although the possibility of a back cap is gone) but the enclosed space makes pre-conceived and organised attacks and traps designed specifically to kill the medic possible, as well as the small size of the map meaning your own medic can stay safe but still be only a short walk away from combat. B4nny’s speciality medic killing sticky traps no doubt raises the drop rate on this map. Also Kritz use, which is more attractive on koth, increases drops.

Amongst the CP maps medic survivability is at its worst on Snakewater and Gullywash, yet avoiding uber drop and instead popping it is apparently easier on Gully.

Medics felt most comfortable on Process, rarely dropping and surviving most easily of any cp map. In contrast Granary is a nightmare, Medics frequently dropping and only managing to keep the number of uber deployments up because of the chokey spamming nature of the map with extra time to rebuild after the last visit to the spawn queue.

Map Rounds and Caps

4 - rounds caps map

The top axis is rounds per minute and the bottom is caps per map round.

Badlands it the only cp map here that consistently has an extra round or two than other maps over 30 minutes, the rounds per minute difference between the other maps being minimal. Viaduct’s average is lower as it has a minimum time limit, so no 60 second rounds.

Caps per round shows the potential for end to end games with teams swapping points rather than linearly ploughing from mid to last. Process and Gullywash are clear leaders in this regard promising more comebacks from holding last and more swapping of points in general. Granary and Badlands have the most linear play of the 5cps, but on Badlands it’s quicker. The difference may seem insignificant but over the course of a 30 minute map it adds up to 7 point captures on average difference.

The sample size is small and HRG and iM were willing to play a slow low scoring strategy, how that will have affected these stats isn’t clear but it’s possible there is more cap volatility with more aggressive play.

Map DPM

4 - map DPM

Clearly Viaduct promoted the most aggressive play style. The fast respawns, small map and pressure of the clock forced teams into almost constant combat. Koth strategy also encourages suicide runs and makes it possible to throw lives away in favour of damage on the right target.

Gullywash’s design with multiple routes, short distances between points and potential for back capping promotes the most fighting of any cp map even thought it has some of the most difficult chokes to negotiate. Granary and Snakewater give players more time to breathe.

It’s noticeable that Medic survivability and consistent protecting of ubers don’t have a relationship with aggression and damage dealing.

Player/Map DPM

In a desperate effort to shore up  War’s apparently crumbling ego have gone a whole article without being top, I’ve decided to revisit player DPM and see how it varies per map. The graph for the players is large, so only the top 11 is shown full size here (Gullywash’s highest damager wouldn’t have been in the list if it was left at ten) and the full list is a small clickable image.

Player DPM by map

So where is War laying down so much damage? The most obvious point to note is that his Viaduct damage reflects a huge game he had, stats indicate they were probably running Kritz for much of it (8 hits recorded in excess of 200 damage over several minutes). However he wasn’t a one trick pony leading damage output on Snakewater and Process. We’ll never know how he did on Gullywash, the log didn’t make it.

None of War’s other maps are shabby, they can’t be if you’re going to lead on overall DPM but the two traditional classic maps of Badlands and Granary are a relative weak point, Badlands was played against Epsilon and Team Immunity were one of the Granary games. Waiting for Sheep to get a headshot is bound to hurt anyone’s DPM figures. In the future when he inevitably takes on  Numlocked in a grand final he’ll need to do better.

Numlocked doesn’t lead on any map, but simply puts out massive damage on them all with no weaknesses.

It’s a similar story with  Brego apart from an abysmal performance on Snakewater. He has the highest output on Granary and Badlands and none of his other maps are anything other than highly competitive.

Mike’s figures on Viaduct and Badlands are up there with the top demomen, and it’s worth comparing Mike and Numlocked with the other demo/soldier pair in the list  b4nny and  Lansky.  The NA pair had better Process and Gullywash maps but on the others they were seriously behind, Mike’s ability to survive away from the Medic and get the damage out is the difference.

Yuki falls behind Mike on Snakewater, Gully and Viaduct but puts out top damage elsewhere.

 Ryb and  Zebbosai put out such neat amounts of damage it’s as if they were playing to a quota on every map, although they flex their muscles a bit more on Gully. Strangely they had quite low damage output on Viaduct, they must have adopted an unusually defensive strategy.

Forcing his way onto the end of the list in 11th spot is  Exon who had the highest damage output on Gullywash. He also put out very high damage on Badlands but struggled to match the top players on the other maps.

Click here to view the full list

Player DPM full

Kills Assists/Deaths

All maps have the same number of kills and deaths on them, and no map seemed to have a strong bias to encouraging more assists than any other. However individual player KAD varied significantly by map. To avoid a massive graph the top 15 players are presented here, the full list can be opened below.

Player KAD top

The outstanding single map performance is  Stefan on Viaduct, he had the pleasure of savaging the invite section minnows DIS and then an open section team so it’s no wonder he padded his stats. Matching this is  Mirelin on Process where TCM battered EV but more creditably also had to face Immunity.

Yuki could have put out Numlocked type numbers if he hadn’t had a comparatively terrible Gullywash. Looking at the record they only played one game on Gullywash, and it was against Epsilon which goes to show what playing Epsilon can do to someone’s stats.

A case in point are Mirelin and Ryb’s Snakewater performances. They played the map twice, both times against Epsilon. In Epsilon’s case Badlands seems to have been their worst map (worst being a highly relative term in Epsilon’s case) and although they had a couple of steamrolling games on it, it’s also the map they lost on golden cap in the final.

HRG’s Lansky and Shade struggled horribly on Gullywash where they had to face both Epsilon and TCM. However it’s notable that b4nny is the HRG player who put up good performances regardless of who he faced.

Mirelin and Ryb also had poor Viaduct games in this metric, but the scouts (see below) both had good games. As discussed before it’s expected for TCM’s stats to reflect an aggro style and playing Epsilon by far the most, and therefore generally be a bit worse off in KAD.

Flippy’s stats reflect playing Epsilon, Infused and Immunity on Granary and HRG and Immunity on Badlands.

Player KAD full

Team vs Team DPM

The more we look at individual player performance on a map, the clearer it is that the small sample size means that the opposition is the bigger factor. These are the head to head damage per minute figures for the top 8 teams. They’re presented as differentials, a negative figure showing that the other team had more damage.

4 - h2h Team DPM

Head to Head Team DPM

The strongest teams tended towards a positive DPM but it’s notable that Epsilon overall actually had a DPM disadvantage in their head to head games against TCM despite clearly coming out on top in them, confirming that’s it’s far from the whole story.

Apart from handing out beatings to the usual suspects TCM actually struggled a bit against stronger teams who more than matched them for damage output.

Considering War’s tournament leading DPM output it’s a little surprising EV are so comprehensively out damaged by almost everyone, but they can rub Infused’s face in their comparative equality in damage, if not the match result.

However the beating of the tournament goes to Epsilon’s performance against the Frenchies whose aggression was met with superior deathmatch and coordination at every fresh encounter. Kudos to the French that they obviously didn’t stop throwing their bodies in the way of Mike’s rockets, no white flags in the face of Epsilon’s pan-european blitzkrieg here.

Team vs Team KAD

On the flip side of the raw aggression of the damage output are the kill/assist/death stats. These are average differences in kills and assists over deaths in games between the teams.

Head to Head Team KAD

Head to Head Team KAD Table

Epsilon look even stronger by this measure, they didn’t give up a negative KAD to anyone, although it was a close thing against HRG. The games between them weren’t particularly close, but HRG did fare slightly worse overall than TCM against Epsilon so there is a good case to say these sets of stats are a reasonable representation of the hierarchy we saw in the tournament.

Fellow finalists TCM on the other hand had a negative KAD against Epsilon, HRG and Infused. They also had an inferior DPM against HRG, and throughout the tournament they drew their total maps against them, coming up short twice on Viaduct but winning Gullywash and Badlands. If it’s a decent representation how can this happen? We’ve seen Viaduct has relatively high DPM so given that HRG won two maps against TCM there that probably skews that overall stat against them.

EV and DIS tie for marginally the biggest beating at the hands of Immunity and HRG respectively, the two teams that pursued a strategy that produces good KAD numbers most clearly.

Conclusions

The map pool played at i49 and that the competitive scene has gravitated towards over time provides a lot of variety in a number of key gameplay stats, no two maps seem to have the same types of characteristics despite following similar broad design philosophies. There are no simple links between the obvious statistical categories.

Player performance varies from map to map but the opponent makes the biggest difference, although judging from the logs you can say some maps favour certain classes, Viaduct seems to be strong for scouts, Process good for medics, etc, which then feeds into the strengths or weaknesses of the team, or just strengths for Epsi.

Numlocked’s consistency and head to head capability is his strength, not even Mike managed to be ultra-strong on every map (his Process threatens to be merely human). Numlocked gives Epsilon a reliable core of top damage output to play around on every single map in the game regardless of who he is up against. How things will work with the new Medic we’ll just have to wait and see, but getting a few extra rounds of Kritz in to cement overall top DPM spot has to be on the to-do list.

Despite people apparently becoming bored of giving him 3 awards per season Stefan still seems to be the strongest scout in the EU. The logs still love him, his very high DPM and his class leading KAD. Flippy was no slouch at lan as well and his performances deserve to be one of the stories of i49. Outside the EU Sheep is the other scout with the best showing, but you have to accept a massive amount of sniper in his play which wouldn’t suit all teams.