As the sun sets on the first fixture of ETF2L season 13, a select few are left feeling satisfied that they have done all they could to win it or at least try and drag their team away from the oncoming train; and an even more select few will feel like they were the ones steering the wreck.
DDR RATINGS
The DDR rating is an arbitrary value that has been calculated by asking a load of prem players and myself to give a mark out of 10 for how well they perceived each player as doing. Players taking part in the game were only allowed to give values for the opposing team and all of the values have been averaged and standard deviation calculated.
Player-Rating/10 (σ)
Epsilon eSports
KnOxXx – 9.6/10 (0.50)
numlocked – 10.0/10 (0.00)
Mike – 10.0/10 (0.00)
GeaR – 9.6/10 (0.73)
schocky – 7.5/10 (1.3)
wltrs – 8.4/10 (1.3)
BFF
Mirelin – 7.0/10 (0.93)
kaidus –6.8/10 (2.32)
Darn! – 4.7/10 (3.53)
CaptainHax (merc) – 6.2/10 (0.69)
cookye – 9.0/10 (0.55)
bybben – 10.0/10 (0.00)
For those of you that haven’t studied statistics, that little squiggle in the brackets in the Standard Deviation – this helpful little value enables us to reflect upon how split the opinion was over this player; the higher the value, the more varied the panel’s opinion was. Anything over a 2.0 is what I’d class as controversial among our 20-something panelists and in this instance, the rabble rousers were kaidus & Darn!, who split the panel a considerable amount. The most valuable and least controversial players in this game are numlocked, Mike and bybben all netting themselves straight 10s across the board.
INSIGHT
After the game I caught up with kaidus and Mike for their opinions and a little insight about what went so right or so wrong for their respective teams:
kaidus – “Was a fun game. A bit disappointing that we (read: I) slipped in the last 10 minutes of gully, but we’re happy with the overall performance. The result never really mattered to us, it was more just an opportunity to see how we’re developing. Conveniently disregarding the last few rounds of gully, we’re satisfied. If zeb[bosai] had been available we’d [have] taken badlands in my opinion, but [we] need to work on gully mids against epsilon as well as last defences (it’d be nice if in PCWs some teams would give us practice at defending last??)”
Mike – “GG and wp to BFF, we were surprised by how they played on badlands and were on our back foot for most of it, however we got used to their pure-dm style of play and managed to get a lucky cap right at the end due to a successful push. We got our heads together and resumed what I feel is our normal playstyle on gullywash and we were very pleased with the score on gully and of course the final score, just the start we wanted :). Although I felt everyone in epsilon played well, Bybben really stood out for bff and did so much work for them, his dm was insane tonight and he saved his team’s behinds countless times with regular 3/4 mans, wp borje(?)”
A few of my panelists (who shall forever remain anonymous) gave abnormally low or high scores, and so I asked them for their opinions on the matter; this is what they had to say:
Panelist #1 -“Schocky & Wltrs just didn’t make any impact because Bybben was on fire; epsilon couldn’t handle him!”
Panelist #2 – “Well, Darn! did nothing, just look at his stats! He jumped into nothing, did no damage and basically swanned around for the game, poor show!”
Panelist #3- “I think Darn! has been poorly criticised for his role in Gullywash; a lot of his bombs and plays would have worked excellently if he had the backing of kaidus who was being utterly shut down and crushed by Epsilon”
Panelist #4 – “Mirelin was caught out way too many times and generally outclassed by KnOxXx, even on the close Badlands game, KnOxXx had 5 more ubers than Mirelin. Maybe it would have been different with Zebbo[sai].”
STATS
The DDR ratings & Insights are great tools; however, it’s purely subjective. With the presence of TF2logs on the server we can draw some graphs and more conclusions and see what we come by:
So far so similar; although we do see a skew in the balance between cookye and schocky, and Darn! and GeaR, there is no real overall damage difference to worry about. However we do notice something very interesting we take Damage/Death into account. The damage per death of the Epsilon soldiers is approximately double of their BFF counterparts and kaidus has 5 more deaths and 165 damage less per death than numlocked (almost a fully buffed scout less). This doesn’t paint a particularly accurate picture of how the game went, but these statistics are interesting at the very least and slightly revealing at best.

The Gullywash stats really don’t warrant reading into too much; as you can see by the information in the logs, things were actually relatively close damage-wise right until the last third of the game where shit really hit the fan for BFF and Epsilon started to dominate. You can draw any conclusions you like from the logs here & here.
POST GAME JUDGEMENT
If you read my pre-game write up, you will have seen that I made a number of predictions about who was going to do what in the game; this is a summary and a little confirmation that I was right, like I am about everything:
Lynchpin – Wltrs
I think wltrs played passably in this game; I don’t think he contributed quite as much to the game as bybben did, but I don’t think he should be disappointed by his own performance.
Man to Watch – Numlocked
numlocked played really well, netting himself straight 10s all the way across the board, man of the match deserved.
Clash to Catch – Cookye vs Schocky
schocky really wasn’t motivated before the game today, and it showed. cookye ftw.
Most likely to Choke – GeaR
It’s always hard to pick a ‘most likely to choke’ award when all of the players in a game have been playing at such a high level for such a long time, however, GeaR managed to not only stave off a failure, but also gave a sterling performance, making me feel like my pre-game award was unjust! wp!
IN CONCLUSION
I feel that I could have gone much deeper into the stats and vods from this game, but this is already much longer than your average VanillaTF2 article; if you think this has been of any worth then I’ll make it longer next time, if you think I have
Darkdwarfed all over the place in a realm that utterly doesn’t belong to me, then please comment below and tell me; this was simply meant as me testing the water to see if this kind of content is something that is wanted in the scene.
I’ll leave you with a great summary by the Epsilon lads in teamchat after the game:
Epsilon GeaR (team) : Deserved?
Epsilon numlocked (team) : no :D
Epsilon Mike (team) : I warped all over the place and got frags
This has been the Chaplain perspective and I’ll see you next trainwreck.







55 comments
TweetVery nice article. We should have more of these.
Btw, the logs don’t count overhealing, so healing stats are kind of useless.
Great initiative. I don’t really see why you should keep the panellists anonymous though, it makes their insight kinda empty.
Kaidus just butthurt at variance.
fat bitch gibbz
- # - nice! +4I think I’m at least two of those panellists… but which ones? :o
- # - nice! +3I’m one as well, although Chappers did use some artistic license on it :D
- # - nice! +0Its hard to get something meaningful when your insights are given through a haze of digestives, breakfast blend tea and dog walks
- # - nice! +3:DDDDDD
- # - nice! +0I love these statistics, definitely moving to the right direction!
- # - nice! +1yay graphical stats
- # - nice! +3I like it. More stuff like this is good for TF2. I felt like each section could of had more direction though and maybe deduce more from the stats. Atm we’re just being given a lot of stats.
Keep going though.
- # - nice! +2That’s not how you make piecharts
- # - nice! +5my condolences
- # - nice! +1holy shit chaplain you’re really good at this writing shit
I would’ve liked to have seen your comment if you’d got a 4 out of 10.. :D
- # - nice! +5holy shit chaplain you need brain surgery
Chaplain, I was sceptical when you first said about doing this and the whole off site posting etc, but this is totally where it belongs, and I can see this becoming a regular feature – just don’t pull an Admirable!
Mad props yo, great post, really interesting read :)
- # - nice! +4“this is a summary and a little confirmation that I was right, like I am about everything”
You were completely wrong about our game though?
- # - nice! +1It was a joke, I’m sorry, won’t happen again
droso for the love of christ stop getting so butthurt every time someone doesn’t say nitrose is the best team in europe :D
aaaaarfhfhhhghhh to much thinking from information to my heads :(
- # - nice! +1I like the ratings section but I think you should hone your graph skills – they need some polish.
- # - nice! +2Chaplain scores a 4/5
(point deducted for missing bodywarmer reference)
Staying anonymous? pussies…. (and no, I dont mean in the british sense of the word)
- # - nice! +3You mean the entire panel is….made up of kittens?
- # - nice! +3YES!!!! no, not really, just trying to make a poor atempt at a Arrested Development reference :(
- # - nice! +4Thanks for the feedback guys, keep it coming
- # - nice! +1Chaplain would be nothing without my editing skills, where is my fucking shoutout you bodywarmer wearing bender?
DDR rating = Darkdwarf’s Retard Rating rofl
- # - nice! +5I thought it meant how good each player was individually at Dance Dance Revolution :’(
You’ve shattered my dreams.
I’d probably get an even worse rating in that, so no thanks.
- # - nice! +3www.youtube.com/watch?v=wLTs_K...
neva 4get
- # - nice! +1Why has it been viewed so many times :(
- # - nice! +1FUCKING SIDEWAYS BLONDE CUNT
- # - nice! +0Damn… I’d do pretty well!
- # - nice! +1rofl :D
- # - nice! +1Amazing write Chaplain. I really love it :)
- # - nice! +1I like the idea of an after match “review” if you will. However after all the nice stats, players that went big and an opinion from each side, whats missing is getting insight into what actually went down. This + Byte analysis for a couple of top/close matches would be sick.
- # - nice! +3If you exclude the bad punctuation, the overuse of semicolon, and some other minor shit, this is a great article. It has room to improve and get better next time, but it is definitely something we’d like to see more often.
- # - nice! +1Only 12, thats an all time low
- # - nice! +2Awesome article, really nice to see some stats presented and not just the log files pasted in the chat. I really hope we’ll see more of these. I’m not sure how long time it took to compile it, but asking 20 people a bunch of questions might be quite time consuming, if you can keep it up, awesome; but be careful that you don’t end up “burnt out” by the third article.
Though I agree with the sentiment, it would be nice to further analyse the data, for example: you conclude that the BFF soldiers did less dmg/death than their Epsilon counter-parts – if you could explain how this affected the game for example it would be really nice, do they die more because they suck? Or is it more a case of play style? If you could take, the statistics, knowledge about how the team likes to play, e.g. two suicidal soldiers, look at how they actually played (did Darn’s suicides do enough?!) – all the information is more or less already in the article, it just needs to be tied together a tiny bit more, some more conclusions beyond the obvious things the data tells us would be nice.
As a random addendum, it would be cool if we could see a players stats over the course of a game, look at when the demoman did loads of damage, was it after or before they played shaky – nothing to do with this article, but could be a cool statistic to be able to read from logs.
- # - nice! +6awesome article. remove the healing stuff (as f2 pointed out they’re bogus stats – could look at ubers and deaths instead) and you’ll be on to a serious winner
- # - nice! +3I don’t really like the whole… damage per death thing. It’s no surprise that a team’s players get double the damage per death, if they’ve wiped the enemy team on middle, and then get a second battle in before dying. Might as well just look at who died the most, and you will probably get the same results anyway.
Overall damage in the game is a far better statistic.
I’d actually be interested in seeing capture point patterns / routes. Don’t really know of a better way to put it, but I’ll try to explain…
Let’s say, the course of a round went: Team A caps mid, Team A caps 4th, Team B retakes 4th, Team A caps 4th, Team A caps 5th.
I’d kind of want to see that in a diagram form, so you could come to conclusions such as, “TCM fail their last CP push this much, or badlands spire tends to be much harder for teams to retake, compared to badlands mid, or Once Epsilon have captured the mid point, they only ever get pushed back beyond that, 10% of the time.
Could find team’s strong and weak points, as well as get to know the imbalances of each map (not just for a team, but as in… the map’s design).
Even better, would be to kind of, combine these stats, and get an individual player’s damage output, per capture point push. So between capping spire and mid, the demoman for each team deals X damage, soldiers dealing X etc…
those aren’t stats that tf2logs.com produces though. it would be an incredible amount of work to put that type of stats together, a la the earlier-days JimmyBreeze articles.
- # - nice! +1good old report cards
- # - nice! +1A+ comment
- # - nice! +3It can be done cause logs tell you what point was captured, but it would be quite a lot of work to do from scratch. Maybe if TF2 logs guy would open source that shit… -_-
- # - nice! +0github.com/barncow/tf2logparse...
Have fun.
- # - nice! +1Back when JB articles were good!!! :D
- # - nice! +2Or an hour or two of messy PHP scripting to crunch the numbers from the logs. ;)
Damage output broken down by capture point push can be found here: pastebin.com/gt6My5QV
That’s sick, nice work CVM!
- # - nice! +1Taking it a step further, it would be quite simple from here to summarise who got the frags and who died for each capture point push. If there’s interest, I could add this in.
I also completely forgot Gullywash. Will process the log when I get home tonight.
- # - nice! +2Really nice!
- # - nice! +2Here’s the same damage output number crunching applied to Gullywash: pastebin.com/CFwWPneH
- # - nice! +3Aaaaaaand stats for both maps reworked to also include who killed who:
pastebin.com/vz9PK4w2 (cp_badlands)
- # - nice! +4pastebin.com/w9Zgpygs (cp_gullywash)
great post, keep it up
- # - nice! +4