Should rosters be given more time?

19-9-2022 21:49

After another summer break filled with transfers, should teams be giving their lineups more time?

"I kinda dislike seeing so many changes every break, because it feels like teams restart their progression. They don't try to find the full potential of the lineup." That's what Russel "⁠Twistzz⁠" Van Dulken had to say during BLAST Fall's group stage when he was asked for his thoughts on this summer's edition of the silly season. So, does he have a point?

There was considerable excitement for new-look projects at G2 and Vitality last off-season, but neither got close to the results their name-value demanded. Instead, it was the teams that stuck with their original lineup (Natus Vincere, Cloud9, Movistar Riders, Spirit) or made one single change but kept their overall system and structure (FaZe, ENCE) that ended the season with the best results. Over the pond, a huge shuffle in North America left Liquid (3 moves), Evil Geniuses (3), and Complexity (2) with fresh lineups, but all three sides floundered.

Aleksib was only given six months in G2 before HooXi was handed the reins

Which begs the question: Are teams pulling the plug too early? If teams kept faith in their lineups for longer, would results eventually turn a corner? How long does a team need before we can safely write them off as having reached their peak? Let's try and answer these questions.

First things first, let's take an overall look at our sample. 71.3% of our roster moves included just one move, 23.5% two, and 5.3% three. Of rosters that are not currently active, 58.7% of them survived for longer than 100 days, meaning nearly half of teams made another change within three months. The median roster timespan is a measly 174.5 days (128 if we include active rosters).

Lifespans of every inactive team from our sample that peaked in the top 20

Of our active teams, 31.8% of teams have been together for longer than 100 days, though this number might be higher if we weren't collecting data just after a player break. Still, the bar chart below illustrates the point: Teams, despite contracts being harder and harder to buy out, are quick to twist rather than stick.

Cloud9 are the ultimate outliers, picking up Abay "⁠HObbit⁠" Khassenov all the way back in July 2020 when they were still Gambit Youngsters. The only other team to have not made a change in 2022 is FORZE, who added Aleksandr "⁠shalfey⁠" Marenov and Evgeny "⁠Norwi⁠" Ermolin in October 2021.

So why do teams do this? Surely now, with data scientists and more and more general managers or coaches controlling the roster, five-man lineups should be given more time?

Well, one answer is that we're not quite there yet; as much as teams love to publicly preach about their professionalism, their decision-making reveals that they are not as well-run as they act. This is a viewpoint vocalised by Aleksandar "⁠kassad⁠" Trifunović and Marco "⁠Snappi⁠" Pfeiffer in recent times, the latter reflecting that "I think a lot of GMs don’t have their finger on the pulse when it comes to talent and what to look for. You have to find out through other channels how their attitude is, work ethic, mentality. Are their numbers inflated by good roles, are they selfish?"

It’s a very cynical business in the sense that if something can improve you, then you need to do it

Marco "⁠Snappi⁠" Pfeiffer on removing Olek "⁠hades⁠" Miskiewicz

But Snappi also employs a ruthless, but effective, roster management in his own team, such as cutting Joonas "⁠doto⁠" Forss for Pavle "⁠Maden⁠" Bošković or most recently hades for Alvaro "⁠SunPayus⁠" Garcia; the skill is knowing when a lineup can be improved in the market. Roster moves, when done right, are an opportunity for a team to evolve. But how do we measure this?

One way to measure the effectiveness of a roster move is to look at a team's HLTV ranking on the final day of their old five-man lineup, and compare it to the peak of their new lineup. So, let's take a look, using only teams that were in the top twenty before they made their change to remove huge, disproportionate, jumps made by teams like ENCE and Movistar Riders.

This chart only includes 103 rosters that peaked in the top 20 and didn't make a change within their first 100 days together

As we can see, nearly every team peaked at least slightly higher than their rank before they made their roster moves. The outliers here are FaZe, who could only peak at No. 15 (-7 positions) after losing Nikola "⁠NiKo⁠" Kovač to G2 and re-instating Olof "⁠olofmeister⁠" Kajbjer, and EXTREMUM, who never went higher than 22nd after replacing Justin "⁠jks⁠" Savage with Hansel "⁠BnTeT⁠" Ferdinand. But, a good 80.5% of rosters improved on their prior ranking by at least one position at least once in their lifetime.

Naturally, comparing a team's ranking at their peak is doing some heavy lifting here. So what about comparing the new roster's peak to the old one's?

Now, we get a much more even distribution, with only about 42% of new teams one-upping their previous iteration. This implies that a lot of roster moves are essentially coin flips, gambles that have no guarantee of reaching the heights of the previous lineup.

Yet, this isn't as stark as it may seem for the 'roster moves are good' camp. Most roster moves are teams making changes to get out of a rut, so a 42% chance of improving, however small a jump, is actually pretty high. This sample also includes teams who made involuntary roster moves, whether because of a player being poached or taking a leave of absence.

Twistzz: "I dislike seeing so many changes every break, it feels like teams restart their progression"

Returning to Twistzz's interview, there is a level of irony to his opposition to roster moves in that FaZe's dominance of 2022 so far was jumpstarted by a roster move. The Europeans jumped eight places in a 98-day spell all the way to the number one spot when they replaced olofmeister with Robin "⁠ropz⁠" Kool, winning IEM Katowice and ESL Pro League Season 15 within their first few months together. Without ropz providing that final piece of the jigsaw, it's unlikely FaZe would be where they are now.

It's necessary to point out that Twistzz, when he criticises teams for "restarting their progression", is not talking about roster moves like FaZe adding ropz. That move was planned for months, with Finn "⁠karrigan⁠" Andersen and Robert "⁠RobbaN⁠" Dahlström building the rest of the side, including Twistzz himself, with ropz in mind. This was the kind of roster move that a general manager makes their money on.

Rather, Twistzz is arguing that teams like G2 and Vitality are reaching for the red button too soon, before their potential was reached. This is a reasonable point to make. Vitality's lineup with Kévin "⁠misutaaa⁠" Rabier could be far better than they are now in six months as the players get used to communicating in English. G2 had barely learned Aleksi "⁠Aleksib⁠" Virolainen's system before deciding it clashed with their preferred style.

Take this scatterplot: the vast majority of teams make changes in the first six months, which skews our data quite a lot. Every team that changes in the first six months therefore peaks in the first six months. We can also see in the exceptions that teams can be patient: Gambit again being the perfect example.

However, we can also see why these teams are in such a rush. No team in our sample took longer than 399 days to reach their peak, even including the online era re-jigging the order of things. Speaking specifically about Vitality, Peter "⁠dupreeh⁠" Rasmussen and Dan "⁠apEX⁠" Madesclaire are both 29 years old; if this Vitality project is going to win Majors, it needs to soon — with a player like Lotan "⁠Spinx⁠" Giladi available, why wouldn't they make that move?

As for G2, Aleksib's system not suiting their players is a good enough reason to remove him, despite not answering why they didn't foresee the ideological clash when they picked him up. 25-year-old NiKo, too, is running out of time to cement his legacy with a Major while he is in his peak. It is very much a chicken and the egg situation; do teams make roster moves early because they feel they can peak no higher, or do teams peak quickly because they change quickly?

When we only look at teams that reached the No. 1 spot — which Vitality and G2 are aiming to be — nearly every single one reaches that zenith within the first season or so of playing together. This bar chart even excludes those who were number one before the roster move (Natus Vincere adding Valeriy "⁠b1t⁠" Vakhovskiy for Egor "⁠flamie⁠" Vasilyev, and FaZe using Richard "⁠Xizt⁠" Landström as a stand-in for olofmeister in 2017).

Every single team to reach No. 1 on LAN since 2017 did so less than six months after making a roster move

Gambit once again stand out as outliers, but we can see that most teams reach No. 1 within half a year of making their roster move, practically still within their honeymoon period. Moreover, we can see the same trend among the teams that never make it to No. 1 — of the teams that improved in the top 20, the median length of time for them to reach their peak was just 103.5 days.

Teams can repeat and maintain their peak, of course, but this data seems to suggest that — in the majority of cases — we know how good a team is going to be after 3-6 months of them playing together. Number one teams are the exception here, with teams like Astralis going through peaks and troughs all while maintaining their No. 1 spot. But, for everyone else, the first six months of a roster are the most crucial and, usually, their peak.

Of course, there are limitations to this hypothesis. We do not know if teams would go beyond their peak if they waited longer before making changes. The pandemic has given teams unreasonable rankings, and community expectations to match; Heroic's move for Jakob "⁠jabbi⁠" Nygaard is a direct result of them trying to return to the heights of the online era, heights that they may have never reached had LAN and crowd play never went away.

Peak team ranking is also rarely fully representative of a team's quality. Vitality were ranked No. 2 in the world when they removed Jayson "⁠Kyojin⁠" Nguyen Van and Richard "⁠shox⁠" Papillon for Emil "⁠Magisk⁠" Reif and dupreeh, but they only reached their rank once the decision to go international had been made. For most of 2021, the same five-man lineup hovered around the bottom of the top ten, providing a solid basis for Vitality's management to make a change in pursuit of silverware.

Liquid didn't wait for LAN to return before changing their Intel Grand Slam winning roster

This happens the other way round, too. Liquid were eighth in the world when Nick "⁠nitr0⁠" Cannella departed for VALORANT, but that squad was the best in the world only a few months before. Throw in the pandemic, and it's easy to argue that Liquid may well have returned to their peak of No. 1 had LAN play returned and nitr0 stayed.

It is an inarguable fact that there would be fewer roster moves (and better teams) if every transfer was conducted with care, working out role, personality, and style clashes before any player gets near a dotted line. In the same way, teamplay like prime Astralis was only possible because their core three players had learnt each other's automatisms by playing together since 2013. There are clear benefits to giving rosters more time.

Six months is a long time in esports. We do not have 'multi-year projects', at least so far. When teams waste six months shuffling between stand-ins, they are potentially wasting two runs at the number one spot. Careers are short, and the odds of the last five years tell us that most number one teams reach that point with the help of a roster move one or two off-seasons prior.

Astralis may not have become the greatest team of all time had Kjaerbye not chosen to leave the team

In fact, every team that reached the No. 1 spot outside of the online era did so less than six months after making a roster move. Like Twistzz says, "you don't necessarily have to [add] a star player to reach" the top, his example of Ilya "⁠Perfecto⁠" Zalutskiy joining Natus Vincere being a good one of how a tweak to role balance can make every player improve.

This is a route of several number one teams, whether by luck like when Markus "⁠Kjaerbye⁠" Kjærbye leaving Astralis gave dupreeh his preferred roles back, or deliberately when Perfecto was added to Natus Vincere or Ricardo "⁠boltz⁠" Prass to SK.

Adding firepower is just as common in our number on teams of the past: FaZe adding ropz, Natus Vincere adding b1t, or FaZe adding Ladislav "⁠GuardiaN⁠" Kovács and olofmeister.

FaZe took no time at all to start winning trophies after adding ropz

A change of IGL — with the exception of Evil Geniuses adding Peter "⁠stanislaw⁠" Jarguz and apEX taking over from Alex "⁠ALEX⁠" McMeekin in Vitality — is rarer in number one teams, with most adding a player to an already functioning system.

The same goes for double moves. There are only two examples, FaZe's blockbuster move for GuardiaN and olofmeister and Heroic adding Nikolaj "⁠niko⁠" Kristensen and René "⁠TeSeS⁠" Madsen, of a number one team making more than one swap to create their world-topping lineup.

Double moves, triple moves, and IGL swaps should be placed in the same category. They are sometimes necessary, but one should not expect a fast improvement from them. It is far more likely that they are landing platforms, signalling an era of transition for teams before they make a final, single, roster move to push that new system or core into tournament-winning form.

G2's ESL Pro League form shows that exceptions exist, but a case like Aleksib's addition to Ninjas in Pyjamas is more typical, in that the core's ceiling is probably still one roster move away.

The greater the change, the longer it takes for a lineup to reach their peak

For the vast majority of teams, it is not only easier but more effective to solve role clashes, firepower deficits, or an off-meta system with a roster move. The burst of motivation when a new player arrives at a bootcamp is another intangible that makes roster moves so tempting.

Teams like Virtus.pro, who could swap around roles every six months and take a charge at the No. 1 spot without even thinking about a roster change, are practically extinct. Players can develop while in a team; Spinx is far more of a star player now than he was when ENCE signed him. But it is Vitality who might benefit most from that development, by signing Spinx as both a ready-made product to improve their firepower and a lurker to help their role conflict.

This is ultimately a discussion of hypotheticals, but it is also a case of players making their own truth: By believing that a roster will not improve, they are making it very difficult for things to turn around, making a change inevitable. The number of roster moves around the top twenty may decline one day, when more teams find success sticking to their five for long periods of time and coaches are granted more control.

For now, history tells us that the best way to reach the very top is to take an already good core and add one final piece to the jigsaw. As long as teams are in the hunt to become the best, they will do whatever it takes. For good or ill, roster moves are the easiest, sharpest, and most direct way to take a squad to the next level, so don't expect off-seasons to quiet down any time soon.

For similar deep-dive articles, check out the links below:

Is the modern AWPer really too passive?

When do Counter-Strike players peak?

Expert take: Age and motivation in Counter-Strike

Why are modern IGLs so aggressive?

Is double AWPing worth it?

Who are CS:GO's map specialists?

What are CS:GO's easiest and hardest CT positions?

What are CS:GO's easiest and hardest T-sided positions?

We paired current and former academy players with their top tier doppelgängers

No. Less.

2022-09-19 21:49

26 replies

Should the hltv ranking formula be reworked ?

2022-09-19 21:50

12 replies

Yes

2022-09-19 21:51

4 replies

Should folks get money for free ?

2022-09-19 21:56

3 replies

2022-09-19 22:22

1 reply

what a question hltv........ ofc no😋

2022-09-20 09:11

2022-09-19 22:43

idk what i just read... they seem to have forgotten about how long astralis have been playing together to finally be able to take a major. and nowadays it's like "3 month trial time" and then "next roster swap" if you dont win a major in that time

2022-09-20 00:27

5 replies

The Core of Astralis had been playing together for a long time. But not the 5-man roster after the addition of Magisk. They only took 3 months to reach the top.

2022-09-20 01:25

Device/xyp/dupreeh have been playing together since late 2013, 7-8 months with kjerybye and added gla1ve 2-3 months before eleague atlanta in late january 2017 which they've won, and then later snake aka kjaerbye bye left for fc club money at the start of 2018, brought in magisk and started to win within 2-3 months.

2022-09-20 01:57

Hope NiP dont read this article. U can play terrible for 3 years and NiP wont bench u

2022-09-21 18:44

2 replies

ironic coming from a f0rest and friberg fan

2022-09-21 19:14

U dont kick legends. Its like bench krimz on fnatic. Plopski is the perfect example. Hampus should be the next, terrible igl, not enough fire power, es3tag is better support. I cant found a role for him on this team. Get a real awper and bench him

2022-09-21 19:21

Yes...movistar, astralis, and liquid are top 5 lolo... And Movistar didn't qualify for ESL or Blast...

2022-09-25 23:01

C9 respect

2022-09-19 21:55

So many charts I didn't undestand a single thing lol

2022-09-19 22:40

9 replies

2022-09-20 00:08

I think it is well written, maybe translates poorly or can’t be read as easily if 2nd language or something? What charts were the problem

2022-09-20 00:09

The way was made it its a little bit confusing and tons of types. It supposed to be helping to understand visualy and not this kind of mess. At least for me.

2022-09-20 01:07

I feel that, they were unnecessarily complex

2022-09-20 02:51

2022-09-20 05:40

The dots are awful, blue shading, i have to work to spot the difference and constantly check on the legend, but hltv is blue so it has to be blue to keep up with the theme, although we used orange and green later, just uhh, ignore that. 🤓🤓🤓

2022-09-20 01:59

Will agree the color choices were questionable, and some presentation choices like the blue dots should just be objectively terrible

How :D? No university for you, i guess

2022-09-23 18:24

I am graduated didn't said I'm not familiar with charts I just said its a mess and not well designed.

2022-09-25 23:10

Lol

2022-09-20 01:43

Say less

2022-09-30 15:02

yes

Yes please!

If they don't 16-0 first match, disband.

7 replies

+1 if they lose a single match scrap the roster, org, coach, country, city, town, village, planet.

2022-09-19 23:02

6 replies

*village town city country planet what is ur random order lmfao

2022-09-19 23:11

well everything is going so the order doesnt particularly matter thats what came to mind first.

2022-09-19 23:16

In that case, just start with the planet

That would be very efficient because you'll save yourself some work to do.

2022-09-20 09:05

flag checks out

2022-09-20 09:52

Thanks :]

2022-09-20 09:53

short answer : yes long answer : yeeeesssssss

2022-09-19 23:43

+1 lmao

2022-09-20 01:03

lol

2022-09-20 11:25

You expect me to read all that?

It's a short excerpt compared to the movistar riders article

2022-09-19 21:52

typical millenial with 5 secs attention span..bet you didnt go thru my entire comment

2022-09-20 05:05

2022-09-20 07:29

this is why i love nafany and co 4 players came together from lower tier and stay together (thanks to Gambit to create and grow them), i hope they will be more lucky on the next major

And I hope they stay together as long as possible, they are such a good team.

2022-09-19 22:19

Even if we love them if hobbit doesnt comeback to his 2021 he will need to go out he has been the best problem for Gambitxloud9 this year

Hobbit is not included in the 4 players, so the narrative will stay the same.

2022-09-20 00:49

Hobbit dropped off, so did nafany.

2022-09-20 02:02

Yeah but we weren't really ever expecting him to do such big heavy lifting yet he did, so we should be surprised he was able to be such a force for so long. While they were at no.1 Gambit basically had 3 stars, 1 strong support and a crazy agro IGL

2022-09-20 09:04

And that seems to be the required formula to make a top1 team hahaha this comment got me thinking hard about the shortcomings of a lot of teams nowadays

2022-09-20 17:04

Nice article

Cloud9 keeping it real

Evil Geniuses just needs more time guys

10 replies

no they just need more guys *

I dont agree, they need more strategic coaches, I feel malek need help to manage his strategic view of the current strategic position of the current meta in this strategic context of 15 man-rosters.

2022-09-20 01:00

-cerq +mantuu or any tier 1 awper on the market and they might actually start winning matches in tier 1 tournaments

2022-09-19 22:00

Why do people act like Brehze isn't washed up either? I also dont personally like the nealaN signing but that's neither here nor there

2022-09-19 22:01

Of course he also is a problem but cerq is an awper with rarely any impact and its not like NA has alot of talent. It would be either -cerq -brehze and +eu awper +na rifler or opposite. They want to play na rmr

2022-09-19 22:05

No, EG would never do this. Maybe +mantuu and also keep Cerq, maybe collect some more AWPers as well +mHL +everyone

2022-09-20 00:10

Evil Geniuses just need more money

2022-09-19 22:08

To invest in more rosters and hire more strategic coaches. Success is exponential in that case. Maybe, buy a whole region like all australian rosters and compete them in NA. If one of the regions fail miserably, act like they are the best region, like the NA vs SA. Sa is shit, but man, look how they beat NA.

The best investment Na orgs could do is move over the top 4 oce teams to compete in the region. Mix the talent we have here with what you guys got there after 12 months. You will have yourselves a revitalised scene with 3+ top 10 line-ups and a couple of 10-20 teams also. Even mix in a couple of CH+Asian lineups all these scenes are struggling to compete internationally in fps, its time to merge the talents in NA and mix it up.

2022-09-20 03:42

I was joking, but yeah, instead of always fighting NA against NA, they should mix more like before COVID. It wasn't perfect, but NA + SA + AU was enough to keep the level kinda high. Also, it merges many market together, so more potential viewers.

2022-09-20 03:55

2022-09-19 21:53

good question

Valde Porvalo

1 month is enough

absolutely not if you are hiring someone like botleksib. Rip NiP

rent free

2022-09-19 21:57

rent free fake flagger

2022-09-25 13:28

-aleksib +ztr

2022-09-20 00:19

HooXi smiling rn

banger

Yup unless it's really clear early on that the players/coach aren't getting along

Yeeess plss

2022-09-19 21:58

2022-09-19 21:59

-Hooxi asap

2022-09-19 22:02

who the fuck reads this articles? it's an fps game bro you know? bang bang pistols headshot noting more

2022-09-19 22:03

then why bother to comment

2022-09-19 22:31

2022-09-20 00:13

HLTV user questioning his own sanity (2022, colorized)

2022-09-20 00:11

2022-09-27 02:21

yes unless its very clear that coach and players (mainly igl) arent getting along like in c9 and EG with stewie and maleK

-Valde +BOT Matt

2022-09-19 22:04

well idk i have a new roster every match😎👍👍

I think 1 month is enough to see if roster has potential After 3 months u can see if this roster worth to be paid.

its not about time, its about maps played in official games.

2022-09-19 22:26

yes, if we count 2 tournament per month

2022-09-19 23:09

Not always. I still remember how ISSA was in his first month or so in HR. They decided to give him more time and after like 3 months he became one of the stars there. Guess coach and/or IGL saw his potential. But that's also the example of org that wasn't trying to be number 1 from the get go.

2022-09-20 09:29

nice plots

No Orgs should put more pressure on these lazy paycheck stealers and fixers with <150 hours/2weeks. They only tryhard when they are about to lose their "job". More pressure (unless you are shitting urself) = u learn faster and do better, its a well known fact.

2022-09-19 22:12

150 hours per 2 weeks is just being unreasonable. That's more than 10 hours every day. Nobody spends that much time on anything. Especially not for extended periods of time. Major coming around and you want to win? Go for that 100 in 2 weeks, sure. Random blast group stage? Don't sweat about doing 70-80 to get back into it.

2022-09-20 01:32

Even woth 100 you work so much more than a full-time job and your risk for RSI would go way up

2022-09-20 06:54

It's not uncommon for players to do 100 in the lead-up to the Major. But yeah, still crazy.

2022-09-20 10:46

150 hours in 2 weeks? I'd rather be a paycheck stealer than have no life

2022-09-20 02:15

loving these articles, keep em up!

2022-09-19 22:14

2022-09-21 19:17

you should make changes as quickly as you feel is right. I think there was times where Navi took way too long to kick Flamie and the same with G2 with Jackz. You cant just wait around for like a year when you have some of the best players in the world on your team in their prime. Kick the bots and replace them until you find the right piece

While I agree that the change should have happened sooner in both these cases, flamie was definitely the harder one of the two. Navi was the best team when covid hit with one of the best tournaments out of flamie for a long time in katowice 2020. There was no quarantee that b1t or anyone replacing flamie would have been as good on lan as flamie was before covid. And nobody thought that lack of lans would last so long. But its easy to say what the right decision would have been now in hindsight.

2022-09-20 00:57

Flamie was also shocking for almost the full 2010 season and continued into 2021 until they finally kicked him. Its not hindsight to say removing a bad player would improve them, its just common sense.

He was bad online. We will never know if he would have been as bad if the normal lan circuit would have continued. And everyone wanted lans back as soon as possible and thought it would happen.

2022-09-20 14:59

Give me sauce 2

2022-09-19 22:21

excuses kekw

2022-09-19 22:23

Choko top 1 ez

2022-09-19 22:29

Boltz GOAT SK went #1 quickly after his first shot

top 1 speedrun

2022-09-20 03:36

Not if you have GIGACHAD HOOXI In that case you have instant success.

2022-09-19 22:41

yes, way more!

2022-09-19 22:44

I love data like this, well done!

2022-09-19 22:45

#BRINGBACKNEXAANDJACKZ

2022-09-19 22:46

boltz case is interesting... SK was #3 when boltz was announced, one week later on his debut SK won EPICENTER (in one of the best bo5's ever, vs Polish VP) and climbed to #1.

2022-09-19 22:47

2022-09-19 22:50

rip old vp

2022-09-19 22:52

great article.

2022-09-19 22:58

I'd say it depends.

Dude, are you data scientist or what? This article is awesome!

2022-09-19 23:12

No. When Karrigan won Katowice, he reminded about when he said a year before after coming back to FaZe to "give him a year and he will start winning". And they won Katowice after a year (and then the rest). But let's not fool ourselves, ropz was a major signing. It made the roster supremely stacked and they started performing instantly. Ropz already got an MVP. Remember FaZe not making Stockholm playoffs? Dropping out of the top30? NAVI is no different. They replaced flamie with b1t (full-time) and NAVI won the first event they played. Of course, I don't know the situation behind the scenes at G2 but the truth is Aleksib wasn't working. With that level of talent on the roster, not even making Major playoffs was abysmal. Failing again vs FURIA at Dallas, vs Vitality in Lisbon, and another group stage exit at Cologne. The more stacked a roster is, the less time it should require. Time is needed when you have to make more with less so naturally more preparation is required.

2022-09-19 23:15

this current c9 roster is the roster that should never be touched. braindeads like maui saying to replace nafany (1/1 upcoming igls from the cis region with drive to actually be an igl) who is only 20 years old. you put ALL your resources into developing that guy and not changing him out

Hobbit 2021 needs to be back if they want to win

C9 will be either fantastic story of sticking together and persevering or massive missed opportunity. With 2 top 5 players on the roster, they underperforming a lot lately.

2022-09-20 00:50

I agree removing nafany will crush the whole roster, but if hobbit is still underperforming or interz. in this meta, even your support player should have decent stats.

2022-09-20 01:10

The only thing of note they've done this year was win IEM Dallas, the most scuffed tournament we had. The problem is that nafany hasn't shown that he can adapt his system while other teams surpassing them. Meanwhile HObbit is not doing well as a third star. They should probably wait until after the Major, but if they don't get a good result with clear improvement, they'll have to make a change. Maybe they just need new input, like a second strategic coach. Maybe they need a player change. But as it stands, the team is a fine team, but not a top 5 team. So it depends on what they want to achieve.

2022-09-20 01:39

I would not change nafany for at least some time simply because he, as a leader, was able to say "That's my fault, I played badly as an entry and my calls weren't good as IGL. If you're looking for someone to blame - that's me". That alone takes some huge balls to do when some players after reading some criticism go "You don't understand the game, structure and roles on pro level, stfu" or need some mental coach to deal with it. Also he's quite young. Having that much experience as an IGL in tier 1 at that age is huge. If he can keep up individually he has like at least 7 more years in tier1-2 teams even if this roster collapses.

2022-09-20 10:44

He can be as humble if he wants, if he doesn't fix his bad calls, he is just not good enough. I think the biggest reason you have to keep nafany is because you won't find another good Russian IGL at the moment.

2022-09-20 10:48

I mean, I can think of one OK-ish. And he even was somewhat of an entry as well on one black-and-yellow team, but... Is he really an upgrade? Not even talking about reputation. Edit: also, you don't need specifically russian IGL, you need someone speaking russian, if we go there. Which in terms of CS also gives you at least Ukraine, maybe Kazakhstan, but can't think of any IGLs there tbh. Maybe even some russian speaking latvians, lithuanians an so on if you find someone.

2022-09-20 10:58

2022-09-19 23:31

It really just depends on the roster, who is being removed and who is being added. I think FaZe could have got slightly better if they had kept Olof, but adding ropz made them immediately way better than they ever could. Same with Na Vi and flamie, or Vitality with the all-french lineup. Sometimes you have to risk it and make a sweeping change in order to improve.

2022-09-19 23:34

once again good work ner0, you have a future in data science once you are retired from cs

Complexity is a roster that shows changes are very much needed but then you have the flipside like Cloud 9 where they really don't need a change atm

2022-09-19 23:37

It depends on what C9's goal is. If they are fine being at the edge of top 10, sometimes making play-offs, then they don't need a change. But if they want to reach finals again they probably need to do something. sh1ro and Ax1Le are back in form, but the rest just isn't working out. The only thing they've done of note in 2022 was win IEM Dallas, which was the most scuffed event of the year. Aside from that, they got a play-offs in Katowice, and that's it. Not really top 5 material. Maybe the problem lies in nafany's calling style and they just need some new input from a secondary strategic coach. Maybe the need to change IGL. Maybe it's HObbit's drop-off. But as things are, unless they suddenly make a couple of deep runs at EPL and the Major, they probably will need a change.

I do miss the era of cores that had actually been together for multiple years

2022-09-19 23:39

yep

2022-09-19 23:45

Who ever is doing all this work to make these articles. Bravo. Amazing work. Data collection and giving the community scientific evidence to back up there never ending arguments. It seems that the most effective roster change is the 1 man change after testing out a roster for a long time. With a healthy core that have been playing together for a long time. (Look at astralis, navi and faze all at there peaks) It would be interesting as well to see data on the amount certain players have played together in relation to there success. Say players like ropz and karrigan. Or rain and karrigan. Dupreeh and magisk. S1mple and electronic. It seems like a good core of 1 igl and 2 other players that stick together and search for the 2 final pieces eventually go on to be great teams

Lies, damned lies, statistics. You should be a little bit careful with reports like these. The article doesn't really provide much explanation to the criteria by which the data was collected, or by which it was analyzed. For example, the integrity of the dataset should be questioned with the inclusion of roster moves that were involuntary (e.g. player being poached or taking a leave of absence) but also because of placing IGL swaps, double moves and triple moves in the same category. For the least, triple moves should never be included in the data, because by current definition such a roster shuffle would result in a new core, and therefore, an entirely new team. In addition, no reasoning was provided for the analytical choices. For example, why draw an arbitrary line at 100 days when comparing teams that stick together or make a roster move? What information can you acquire by making such a comparison, especially as the analysis was conducted (like the author states) as a cross-sectional study immediately in the aftermath of the summer player-break? While the hypothesizing and guessing presented in this article are interesting, the statistics don't seem to serve the purpose they should. With no sort of reasoning or guidelines provided, the statistics are only there to support the author's narrative, which is odd because there really is no narrative to begin with. I appreciate the effort and generally love to read these more in-depth reports, but this one doesn't quite hit the mark, does it?

2022-09-20 00:46

The article clearly states that there is a difference between a single player move and multiple player/IGL moves. It even visualized and explained said difference. It's not like the article ignores the existence of it. That would make one question the integrity of the data. The 100 days line is indeed arbitrary, but it is to choose a reference point. Could you argue that picking 128 or 174 days (the medians including or excluding active roster respectively)? Sure. But you have to draw some line. With 58% of the rosters surviving longer than 100 days, it shows that a significant portion of teams make changes within just 3 months. It's not a crazy line to draw. To my understanding the article isn't trying to give a conclusive answer. It's more about giving the data and letting a discussion form from it, probably to boost numbers. But reading through the lines, the article is saying "If you aren't satisfied with your results after 6 months, you should probably look to make a player change."

2022-09-20 01:58

Yes, the difference between the various types of roster moves was explained. This makes it all the more questionable to cram every type of roster change into the same category when comparing ratings before and after making those changes. Plus, there's still the problem of including involuntary roster moves in the same analysis. What comes to drawing the line of comparison at 100 days, you're correct that it might not be that crazy. Emphasis on the 'might', because we don't know how badly the material is skewed due to including data that shouldn't be there. Anyway, why even bother with a statistical approach if there's no way to justify this threshold of 100 days?

2022-09-20 18:33

You should check the short-or-long term plan box plot. It literally does what you ask in regards to types of changes. To filter for 'involuntary' changes would make sense though, since the article is aimed at the active choice to switch players. The 100 days isn't a threshold. Teams with earlier changes are included as well. The article just considers anything larger than 100 to be 'long' while anything less is 'short'. It's an arbitrary choice, but it doesn't effect the statistical analysis at all. If you like, you can read the article and stats the same way but change the 100 to the median and nothing will change. All in all these complaints can be valid, but in my opinion they don't take away from the analysis itself.

2022-09-20 18:52

I agree with you that the 3 roster moves part of the data is a bit trash as then most of the team is completely different so is it even the same team anymore? The narrative is. Is it better to have a roster move or not? I think the conclusion is that it depends completely on what the roster move is and who is moved where. The graph of delta peak to old lineup is pretty much symmetrical so as much as the data is interesting it doesn’t tel the whole story. I personally think that rosters and certain members of the rosters such as IGL’s should be given longer to mature and develop and I don’t like seeing teams basically get picked apart when they have a small amount of success. Movistar riders for example. Stick together for another year and really see how far you can go. I dont think the idea of a roster move is inherently good or bad it really depends on what that move is and what it is trying to achieve. Look at the best teams in football (not best comparison but easy one to use) they have a core of players that are loyal to the team and very important to the team and they a few of the players change over time and get replaced by new upcoming players or players who are better suited to the style the manager is trying to achieve. An example of this could be Real Madrid, Barcelona under pep or Manchester United under Alex Ferguson. They had core players that would Be most important but then during transfer windows would find players to “better the squad” would t always work but sometimes it would. That’s the risk you have to take. You take a punt on a person you think they will take you to the next level and hopefully they prove you right. Sometimes it works and sometimes it doesn’t. But then if you have a 5man team that has been trash for 6 months and have no signs of changing then a roster could bethe best risk to take.

2022-09-20 02:06

2022-09-20 00:07

no, they must play like prime astralis did on the first day without any practice, or else disband!

2022-09-20 00:22

Generally yes, but it should be chosen individually.

Damn, thanks kjearbye for left Astralis

2022-09-20 00:31

I would actually say that especially in past few years certain orgs were WAY too slow to "do" things. In my mind anywhere from 3-6 months (one season) should be more than enough to see if some player struggles or not. In normal sport if someone is underperforming, he would be nowhere to be seen in month or two, because teams have substitutes (bench).

2022-09-20 00:55

Also, most players in sports who perform poorly get changed after a season as well. The season just takes longer, but player careers are also longer, thus making it similar.

2022-09-20 02:00

Absolutely useless article. So many flaws. For instance, Its not like teams swapping players in vacuum, some bench players who underperforming, other - for money, other - for structure change etc. Not all players are equal in terms of their value, so at least article should consider weighted values for players. So many other aspects not considered, I cant even start listing them. Trash article, as always.

2022-09-20 01:15

Okay then. You go and write an article going through all these considerations you're thinking of. Let's start: how are you going to assign weighted values for players? By rating? Impact? Some other metric? Or just a subjective 'feeling'? You want to go through the history of every single player swap and filter by reason? For one, you are going to take a lot longer formulating that data. Two, you are going to be way to specific. All you can really do, is mention that the circumstances aren't always the same and when making individual decisions, these should be taken under consideration. Which the article did. The idea of articles like this is not to say "This is the ultimate solution", but to give a general overview of how things have been in the past. When doing statistical analysis you HAVE to make generalizations. Otherwise your data becomes completely useless and you can't draw conclusions from it. But if you are so convinced that you can write a more compelling article, please do. I'd love to read it! Probably won't happen though, will it?

2022-09-20 02:08

wow, relax there buddy! dont take it personally, ok? i agree with #106 completely, for me its feels wrong to generalize all roster changes, slap some ranking stats to it over some time spans and call an average on it. i like the effort and the data processing in general but this is just different per case, after all, you're dealing with humans. To me it feels like roster changes require custom solutions depending on the options you have and you just cannot and should not try to push that in graphs.

2022-09-20 02:39

I don't take it personally. I just disagree with him. If you aren't going to generalize, then you're just saying you shouldn't do statistical analysis. I can agree with that. Maybe a case study would've been a better fit for this topic. That being said, I think the choice of making it analytical is fine. The article never states "This is the right way to do things." It just shows a general trend and the results you can draw from it, which is EXACTLY the point of an analytical piece. And some people just like graphs to show them general trends and info. I personally enjoy these articles just for that reason. That doesn't mean I would take these statistics under consideration when making a roster move. That would be plain stupid.

2022-09-20 10:41

i can easily write one if hltv approaches me and pay me appropriate fare.

2022-09-20 03:18

You can at least try and retort my questions. Otherwise I can't believe you could. So how would you assign your weighted values you suggested? What distinctions would you add to transfers to make your article better?

mostly yes

2022-09-20 02:04

2022-09-20 02:45

"Astralis may not have become the greatest team of all time had Kjaerbye not chosen to leave the team" Lmao what

2022-09-20 03:49

Kinda true. With Kjaerbye got a little inconsistent post Atlanta and the biggest problem was Dupreeh and Kjaerbye wanting the same roles. That was temporary resolved when Dev1ce to took medical leave and Dupreeh Awped. With Kjaerbye gone, Dupreeh moved to Entry and they got an Anchor player in form of Magisk. They achieved role synergy when they Added magisk. Everybody knew what role they were playing and no one had complains about it.

2022-09-20 10:56

Shroud?

2022-09-20 04:07

If you are peaking after 6 months of the roster existing, it should last for at least another 6 months. I think that's mostly what Twistzz is getting at.

2022-09-20 05:59

All I know is, if TL decided to “have faith” in the line up with shox and missed out on even having the chance to have Yekindar, I would be pissed as a fan.

2022-09-20 08:06

yeah agreed. but was always obvious shox was washed even before they signed him, probably one of the worse moves turned into one of the best for yeki :D

2022-09-20 08:28

Facts.

2022-09-21 06:14

certain teams /players need more time, others do not

2022-09-20 08:27

G2 future top1 ok

2022-09-20 09:47

actually a good, thought out and thought provoking article, idk why g2 (or just niko tbh) wanted jacks out so badly, it is working for them but when you commit to wanting someone out like that you put them in a role, a role they can't come out of, and also the team goes into a role they can't come out of and it's no longer a flourishing team trying to work out what works, it's a team trying to avoid criticism, that changes the whole dynamic of EVERYTHING. That makes the team INEVITABLY spiral down, it becomes a self-fulfilling prophecy, that's what a self-fulfilling prophecy is. so of course when the change finally does come, the team is in such a bad place the only future result is up. Also, when the team dynamic is in a state of 'avoiding criticism' you force players to turn into people like baitf, who do everything they can to stay in the team, in his case milking every single eco, exit and bait frags he can, inflating his stats way out of proportion so when the team tries to work out the problem, they don't think it's him. Other players trying to stay in a team might play 'dirty politics'.

2022-09-20 10:03

konfig and blamef into astralis turned astralis into another a complexity with lackluster tournament results

2022-09-29 08:57

Depends on how many tournament they participate and the results, but generally yes a roster should have given more time, like 6 months at least

2022-09-20 10:06

Thank you for the analysis. HLTV rankings may have problems, but it is the best indicator available at the moment. What about double peaking? I think the common narrative is that a change will occur when the team is in a slump and after we will observe a short honeymoon period (the "hype era"), followed by a decline in the team's performance. Now say the team started working on changes in their system and a steady increase may follow, perhaps peaking much later higher than initially. This would show in the scatterplot as a 'late peaker'. Does the honeymoon peak exist, and if so, does it correlate with the team's later performance (if kept together)? Although more difficult to analyze, I would appreciate such an analysis of the 'local' minima/maxima. Another interesting aspect would be to look into the change in the coaching staff, although sample sizes may be too small to make any meaningful conclusions.

2022-09-20 10:34

No. But give bymas more time and he will be top1

2022-09-20 10:47

That's some great analysis.

2022-09-20 10:51

That really depends, but for example Ence was playing really well, even tho hades wasn't the best awper. Raiders were doing really well also, but now Ence took Sunpayus and let hades go, Ence is playing bad atm and Raiders are also gonna be playing bad. Basically 2 teams who were doing really good got destroyed because of this changes. If your team is in a good phase don't change it even if you think it's best to get a better player there is always more at stakes

2022-09-20 10:55

"Cloud9 are the ultimate outliers, picking up Abay "⁠HObbit⁠" Khasenov all the way back in July 2020 when they were still Gambit Youngsters. " Bring Supra back!

2022-09-20 10:59

how did faze win everything as soon as ropz joined?

2022-09-20 11:15

A good post. My answer is “yes” and in my opinion G2 made a huge mistake by changing Aleksib for HooXi.

2022-09-20 11:18

yep, HooXi for the legend aleksi, totally crazy decision.

2022-09-20 12:34

hooxi>>>>>>>>>>>>>>>>>>>>>>>>>

2022-09-21 00:43

This work is actually good

2022-09-20 14:16

Someone has to lose

2022-09-20 15:12

NEROcs, great job!

2022-09-20 16:59

Teams doing the 10 day contract like the nba scckss

2022-09-20 20:56

simple answer : Yes they should give roster more time. some form of synergy comes after some time while other teams have that level of synergy you are competing against. for some rosters its hard to know early on if that level of synergy can be obtained. which begs the question to cut your loses early on and start with a new addition or two to a roster. one will have to judge that per case.” basis.

2022-09-21 12:53

Just realice NiP not relevant at all since 2017, and they gave players LOT of time until benched them

2022-09-21 18:42

Survivorship bias. The teams that already work well only make few changes so it looks like teams with few change get better results.

2022-09-22 23:37

even hltv writers favor aleksib what is this

2022-09-23 17:14

Some teams won't work no matter how great they are on paper, and a team of otherwise mediocre players might do just fine if the chemistry is there.

2022-09-25 15:18

No. Bc those teams who changes rosters every event should be penalized.

2022-09-25 23:00

NIP xaxaxaxaxaxa

tldr, yes or no

2022-09-27 02:25

I get the point, and it does seem a bit insane how much rosters will change in short periods of time. But really results aren't everything, you could have a team where 3 members of the team don't get along very well, argue a lot, yell at teammates out of frustration etc. Sometimes those players can cope fine with the negativity and play great together and win lots. But some teams may still be able to win lots but can't actually handle the negativity and want a way out of that negativity, whether they want to leave themselves or have someone booted. if stress and negativity are causing the team to fail I totally expect to see roster changes.

2022-10-01 14:03