PDA

View Full Version : Basic stats on Centurion&shinobi popularity based on 2 weeks of gameplay



SikanderAzam
07-05-2017, 11:13 PM
I'm sure this has already been done more comprehensively elsewhere, but I just played for two weeks and recorded the number of centurions and shinobis in every game to see if they are truly picked far more than all others or if we're just whiners imagining things. After recording it and then dividing the total number of centurions I saw over the total number of players I saw in game, I got the following:

Given that there are 14 classes, if they were all picked equally then each would be about 7.1% of all players in all games at any given time.
About 18.5% of all players in all games are centurion at any given time, or about 260% what you'd expect if you assumed classes were balanced.
About 12.8% of all players in all games are shinobis at any given time, or about 180% of what you'd expect if you assumed classes were balanced.

So yes, they're picked disproportionately by a fairly large number, especially given that they're by far the most expensive to recruit and therefore there are many people that still don't even have them unlocked; but it's not game-breaking numbers. It actually means we're only seeing an average of about 1.5 centurions per game, which is far more than other classes... But it isn't the "3 cents in every game" that we see people complaining about.

Centurion play-style was fairly normal, with a lot of gank squads.

Shinobi playstyle was almost exclusively running away and avoiding combat then ranged GB'ing people from behind to try to kick them into holes. Shinobi players based on this observation over 2 weeks are, honestly, almost exclusively just trolls that want to mess around with Shinobi's gimmicky ranged moves. Encountering one such shinobi was also the most glorious experience I had in the 2 weeks... A shinobi standing over a hole tried to range GB me to pull me in, but I countered it which pulled on his chain which tugged him in... I almost died laughing, it was pure poetic justice.

Gray360UK
07-05-2017, 11:26 PM
Did you record the % for other classes though? Assuming an equal pick is a slightly invalid starting point, because I don't think that has ever been a thing. Would be interested to see % pick figures for some of the more popular OG characters in order to have a true comparison.

UbiNoty
07-05-2017, 11:33 PM
Those are some pretty interesting numbers (even though I think you do also need to consider your small-ish sample size relative to the playerbase as well as other factors such as region, and your matchmaking rank). I wonder if I can encourage the devs to divulge some of the hero pick rates and similar stats on a stream sometime...

bmason1000
07-06-2017, 01:17 AM
I would be very interested to see these numbers. Not just for cent/shin but for all. Feels like I'm seeing wayyyy less peacekeepers, shugoki and conqs...and valk, wait and...okay, almost all of the season 1 characters haha. But yeah, would love to see the numbers. Think that would be really cool.

Gray360UK
07-06-2017, 01:27 AM
I made a note of my last 2 Dominions (huge sample size I know) and there were 2 Shinobis, 0 Centurions (I know, impossible right) and 3 Raiders ;)

https://i.ytimg.com/vi/iNPpWSfo7S8/maxresdefault.jpg

bmason1000
07-06-2017, 01:46 AM
It seems like the centurion population has really plummeted in the last couple weeks

SikanderAzam
07-06-2017, 03:10 AM
Did you record the % for other classes though? Assuming an equal pick is a slightly invalid starting point, because I don't think that has ever been a thing. Would be interested to see % pick figures for some of the more popular OG characters in order to have a true comparison.

I'm sure that it's never been a thing, but it's a good starting point to see who is relatively popular. It's an idealized baseline that you'd expect if the classes were perfectly balanced and picked equally, which is certainly what we all want and what the devs intended--nobody makes a game with 14 classes with the deliberate intention of them having radically different degrees of usefulness. Hence a basic null hypothesis. I'd also be interested in seeing % for everyone, but I don't play dozens of games per day so my sample size is only a couple dozen games over a couple weeks. I just don't have a large sample size so I didn't bother trying to make it into something more official. I just know that centurion and shinobi do take up a significant slice of the pie... but not the amount that some people seem to think.


Those are some pretty interesting numbers (even though I think you do also need to consider your small-ish sample size relative to the playerbase as well as other factors such as region, and your matchmaking rank). I wonder if I can encourage the devs to divulge some of the hero pick rates and similar stats on a stream sometime...

It is a small sample size and isn't representative of the entire playerbase, but I highly doubt that there are radically different numbers in different regions. The same meta probably applies everywhere. I'd love if you could get the devs to publish the pick rates, we all know that they have them.

My.Insanity
07-06-2017, 07:06 AM
wow... this Theard is so sensless... the Shinobi and Centurio were released 9 weeks ago!!!

Ofc the nummber of poeple playing those Heros was decreasing but also the playerbase was decreasing because too many players were sick facing those broken tryhard Heros.

and just like UbiNoty mentioned it there are a lot of faktors you need to beware of:

- Gamemode
- Region
- Time
- Orders (yes if you got play X Games as a Hybrid you can be sure its 3x Cent in the Gamemode)
- Playerbase

so sorry but i dont see any use in this post.

SikanderAzam
07-06-2017, 12:13 PM
wow... this Theard is so sensless... the Shinobi and Centurio were released 9 weeks ago!!!

Ofc the nummber of poeple playing those Heros was decreasing but also the playerbase was decreasing because too many players were sick facing those broken tryhard Heros.

and just like UbiNoty mentioned it there are a lot of faktors you need to beware of:

- Gamemode
- Region
- Time
- Orders (yes if you got play X Games as a Hybrid you can be sure its 3x Cent in the Gamemode)
- Playerbase

so sorry but i dont see any use in this post.

Cool.

Gray360UK
07-06-2017, 12:26 PM
I'd love if you could get the devs to publish the pick rates, we all know that they have them.

Would love to see this too. I'm tempted to do what you have done myself. I do play a lot and could provide a fairly large sample size quite quickly. Let's face it, we spend enough time on the VS screen that quickly putting another tick in the Centurion / whatever column isn't going to be a problem. I'm just hoping the Devs publish the rates to save me the time, but overall it's something I'm really interesting in seeing.

Horus-31
07-06-2017, 01:05 PM
I'd love if you could get the devs to publish the pick rates, we all know that they have them.

I think we all would. While we are on the topic of statistics, publishing winrates, K/D and K/D/A per hero and per game mode would also be nice... And a way to close any "that hero is OP/UP" debate.

But I get a feeling that they won't if those stats point out that a certain hero fares better than others in 4v4 (better K/D), is a better help to a team in those modes (higher winrate) and/or tends to gank more (higher difference between K/D and K/D/A)...

bmason1000
07-06-2017, 08:42 PM
I think we all would. While we are on the topic of statistics, publishing winrates, K/D and K/D/A per hero and per game mode would also be nice... And a way to close any "that hero is OP/UP" debate.

But I get a feeling that they won't if those stats point out that a certain hero fares better than others in 4v4 (better K/D), is a better help to a team in those modes (higher winrate) and/or tends to gank more (higher difference between K/D and K/D/A)...
You think the community would react negatively to that sort of thing? I think its common knowledge that certain heroes are more of an asset in teams, this hero is better at 1v1 etc. I can't imagine there being much of an outcry over the actual statistics of this but then again...it is the internet haha. What sort of reaction do you predict from the playerbase?

SikanderAzam
07-06-2017, 11:19 PM
You think the community would react negatively to that sort of thing? I think its common knowledge that certain heroes are more of an asset in teams, this hero is better at 1v1 etc. I can't imagine there being much of an outcry over the actual statistics of this but then again...it is the internet haha. What sort of reaction do you predict from the playerbase?

People complaining would have tangible evidence to throw in Ubisoft's face if there were disproportionate winrates, and as you said it's fairly common knowledge that some heroes are more of an asset than others. Imagine what would happen if they released statistics that gave proof that centurions have a crazy high winrate in 4v4? Evidence to back all the complaints. Ubi wouldn't screw themselves over by giving complainers such good ammunition.

Horus-31
07-06-2017, 11:26 PM
You think the community would react negatively to that sort of thing? I think its common knowledge that certain heroes are more of an asset in teams, this hero is better at 1v1 etc. I can't imagine there being much of an outcry over the actual statistics of this but then again...it is the internet haha. What sort of reaction do you predict from the playerbase?

When I said "a certain hero", I actually had one particular in mind... No, it don't think Ubisoft would release stats that shows that the Centurion is indeed OP. Because yes, we would react even more negatively to stats proving he's OP than we are reacting now to just getting that feeling when we play against him.

Edit: cross-post with SikanderAzam... who actually translated my thoughts into words better than I did myself. :)

bmason1000
07-07-2017, 01:18 AM
Well you both make a good point. Didn't they publish sonething similar towards the end of season 1? It wasn't mode specific though