• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

DOTA2 |OT15| anime only please

W

waxinlyrical

Member
Volodja said:
The algorithm takes "skill" into account and it does use MMR even in unranked, just an hidden one.
Click to expand...

To expand on this more: I'm fairly confident the system equates MMR with skill. This is a problem as there are so many different types of skill in Dota: flat out knowledge (e.g. Do you know what all the heroes do?), itemization choices, AI manipulation (e.g. creep aggro, pulling camps, etc), ability to use trees to juke, team fight knowledge, map awareness, etc.

The system tries to collapse all of that into MMR, when all MMR really is is a measure of your ability to win games. People can literally do nothing helpful for their team but still be carried to a win, and their MMR will go up as a result. That, coupled with the loooong match lengths for Dota leads to people often feeling like they are in a trench, where they are clearly better than everyone else they are being matched with yet it's impossible to climb out.
 
D

Drkirby

Corporate Apologist
The system does use a bit more, but how many factors is unknown. One thing that is know, is it takes into account the number of games played, matching new players with other new players (unless the system determines you are too good to actually be a new player)
 
U

Uriah

Member
When your teammate picks jungle legion commander you know you're in for a good game. Of course the game ends with her running into the enemy team when we have megas and 2 heroes down and we lose.
 
shira

shira

Member
Qvoth said:
hopefully monkey king gameplay will be shown in boston :3
Click to expand...

All-Star game
drunk.gif
 
I

Imp the Dimp

Member
I would think skill and the ability to win correlate highly. Since there is a way to measure winrate and MMR but not skill, it's not possible to say for certain but I think everything we know highly suggests it.
 
K

kionedrik

Member
Imp the Dimp said:
I would think skill and the ability to win correlate highly. Since there is a way to measure winrate and MMR but not skill, it's not possible to say for certain but I think everything we know highly suggests it.
Click to expand...

We had this discussion a thousand times already. Skill and MMR are two completely different things.
You can be the most skilled player in the world but if you play non-scaling heroes you'll rarely ever win games simply because of the generalized inability of random pubs to finish games early and measure when they are stronger than the enemy or have reached their peak. Skill means that you understand what's going on and know how to take advantage of it but you'll always be dependent on the other 4 to effectively win the game, no matter what role you're playing (and the 1v9 mentality is the greatest cancer of the game and I'm willing to bet less than 0.01% of the games can be won this way).
MMR only shows how good you are at winning games and is by no means a measure of skill. Just look at all the hero spammers that take advantage of some OP bullshit hero like Sniper or Troll that even a braindead monkey can play to gain MMR. Those players usually have a very basic understand of the game and will instantly freak out if they need to play something else or need to adjust their item or skill build to the enemy heroes.

There is however a very small set of players that genuinely improve their skill over time and are able to gain MMR as a result but those adjustments are always minimal. When your able to gain 1k MMR in the span of a few weeks, either your initial calibration was completely off or you're just spamming heroes and not gaining any skill to help you in the higher MMR brackets.
 
S

Sigismundo

Member
In regards to the earlier Monkey King discussion I have the feeling he's not gonna be an illusion hero and that the monkey army is channeled CC. My guess is he has a lot of interaction with the trees.

https://www.youtube.com/watch?v=AgBpBSP6WME
https://www.youtube.com/watch?v=GfqJQhJqX7E
Apologies for Pokémon spoilers but these two songs are extremely good (although the first one probably needs context)

bbd23 said:
Hitman is a very good video game
Click to expand...
I wish I had either a better PC (still working on that) or a PS4 cause I'd love to play it. Games I can really sink my teeth into are right up my alley.
 
G

G.ZZZ

Member
kionedrik said:
We had this discussion a thousand times already. Skill and MMR are two completely different things.
You can be the most skilled player in the world but if you play non-scaling heroes you'll rarely ever win games simply because of the generalized inability of random pubs to finish games early and measure when they are stronger than the enemy or have reached their peak. Skill means that you understand what's going on and know how to take advantage of it but you'll always be dependent on the other 4 to effectively win the game, no matter what role you're playing (and the 1v9 mentality is the greatest cancer of the game and I'm willing to bet less than 0.01% of the games can be won this way).
MMR only shows how good you are at winning games and is by no means a measure of skill. Just look at all the hero spammers that take advantage of some OP bullshit hero like Sniper or Troll that even a braindead monkey can play to gain MMR. Those players usually have a very basic understand of the game and will instantly freak out if they need to play something else or need to adjust their item or skill build to the enemy heroes.

There is however a very small set of players that genuinely improve their skill over time and are able to gain MMR as a result but those adjustments are always minimal. When your able to gain 1k MMR in the span of a few weeks, either your initial calibration was completely off or you're just spamming heroes and not gaining any skill to help you in the higher MMR brackets.
Click to expand...


Pubs "inability to win" is on both sides. This argument is trite and old as shit. RTZ win 8k games playing Earth spirit. Solo has been top of EU ladder randoming and playing everything as a roamer support. I got my highest MMR playing as support (5.4k like 2 years ago), then started randoming every game cause all my favourite supps got meganerfed (WW, Sky, CM, Tusk, etc...) and fell to 4.5k.

Unless picks are heavily lopsided (aka no carry vs tri-core), games are decided by skill at all phases of the game. And this include the knowledge to not take fights that you know you'd lose and losing towers and raxes in the process if this improve your overall odd to actually win the game.

Pudge roamer is also super good
troll.gif
 
I

Imp the Dimp

Member
How would one even have such a strong opinion on something when the term "skill" is not easy to measure. Baffling really. Would you agree players with higher MMR have a higher skill on average? If so, you agree that MMR and skill correlate. Correlation ranges from -1 to 1 so any gray shades inbetween account for all the unexpected occurences that can happen in a game you have no control over. MMR has proven very accurate in pretty much all MMR ranges.

I mean, you can't even argue about this topic unless you offer an unambiguous definition for "skill". Leading your team to a win requires a wide range of skills, doesn't it? If you don't think hero spammers are as good as their MMR indicates, then allow me to ask under which exact conditions one is allowed to call oneself a true X MMR player. If you set arbitrary criteria, why stop at hero spamming? Someone who plays merely position 1 and 2 might not be considered a true 5k MMR player because by your dodgy definition they limited themselves to farming positions.
 
S

SecretMoblin

Member
I've been winning a lot as elder titan recently... still the king?

also apparently veno vaulted to the top of my most played list and I have a 70% winrate with it, I haven't been watching pro games but i assume veno is cancer there too? Veil is nasty.



Anyway enjoy y'all's thanksgivings, eat some pumpkin cheesecake aka the second greatest food known to man besides strawberry cheesecake
 
M

Monochromaticx

Member
SecretMoblin said:
I've been winning a lot as elder titan recently... still the king?

also apparently veno vaulted to the top of my most played list and I have a 70% winrate with it, I haven't been watching pro games but i assume veno is cancer there too? Veil is nasty.



Anyway enjoy y'all's thanksgivings, eat some pumpkin cheesecake aka the second greatest food known to man besides strawberry cheesecake
Click to expand...

he was getting picked a bit by zai and wings.
zai goes lvl1 jungle with him ....but wings counter picked it with treant. https://www.youtube.com/watch?v=HPEE_o4oud8
 
K

kionedrik

Member
Imp the Dimp said:
How would one even have such a strong opinion on something when the term "skill" is not easy to measure. Baffling really. Would you agree players with higher MMR have a higher skill on average? If so, you agree that MMR and skill correlate. Correlation ranges from -1 to 1 so any gray shades inbetween account for all the unexpected occurences that can happen in a game you have no control over. MMR has proven very accurate in pretty much all MMR ranges.

I mean, you can't even argue about this topic unless you offer an unambiguous definition for "skill". Leading your team to a win requires a wide range of skills, doesn't it? If you don't think hero spammers are as good as their MMR indicates, then allow me to ask under which exact conditions one is allowed to call oneself a true X MMR player. If you set arbitrary criteria, why stop at hero spamming? Someone who plays merely position 1 and 2 might not be considered a true 5k MMR player because by your dodgy definition they limited themselves to farming positions.
Click to expand...

Ah... the good ol' logical fallacy. Correlation does not mean causality.
You can say windmills are turning because there's wind but you can't say there's wind because it was caused by the windmills turning, the causality only has one direction even though there's correlation between both events. In our particular case, the higher the skill the higher the MMR will be but an higher MMR doesn't necessarily mean the player will be more skilled. It will only mean that he won more games than the ones he lost. Hero spammers are the most common and easiest to understand proof of this concept. Just because you can play 1 hero perfectly and gain MMR doing it, it doesn't mean you're highly skilled, all it shows is that you can game the system. There's also the copycat effect, where a lower skilled individual just copies what a higher skilled one does without fully understanding why that thing works and still manages to be successful with it (very common in CCGs for instance).

As for "true MMR" (your expression), I do believe hero/position spammers tend to not be as good as the number indicates simply because they only employ a subset of all the skills in the game or only apply those skills in a very small pool of situations. Is it hard to imagine that a player that plays everything and manages to maintain a steady, albeit slow, increase in MMR is vastly more skilled than a fast climber that only uses fotm heroes/builds?


Dreams-Visions said:
I don't want to jinx it but...holy shit we have a whole page so far with nothing but talk about Dota.
zWVbK.gif
Click to expand...

That's an excellent excuse to post this:
zo2uq2g41azx.jpg
Click to expand...
 
U

Ultrabum

Member
kionedrik said:
Ah... the good ol' logical fallacy. Correlation does not mean causality.
You can say windmills are turning because there's wind but you can't say there's wind because it was caused by the windmills turning, the causality only has one direction even though there's correlation between both events. In our particular case, the higher the skill the higher the MMR will be but an higher MMR doesn't necessarily mean the player will be more skilled. It will only mean that he won more games than the ones he lost. Hero spammers are the most common and easiest to understand proof of this concept. Just because you can play 1 hero perfectly and gain MMR doing it, it doesn't mean you're highly skilled, all it shows is that you can game the system. There's also the copycat effect, where a lower skilled individual just copies what a higher skilled one does without fully understanding why that thing works and still manages to be successful with it (very common in CCGs for instance).

As for "true MMR" (your expression), I do believe hero/position spammers tend to not be as good as the number indicates simply because they only employ a subset of all the skills in the game or only apply those skills in a very small pool of situations. Is it hard to imagine that a player that plays everything and manages to maintain a steady, albeit slow, increase in MMR is vastly more skilled than a fast climber that only uses fotm heroes/builds?




That's an excellent excuse to post this:
Click to expand...

Knowing that picking the same hero over and over again will win you games is a skill.

MMR is the best measurement of skill.
 
S

SecretMoblin

Member
Monochromaticx said:
he was getting picked a bit by zai and wings.
zai goes lvl1 jungle with him ....but wings counter picked it with treant. https://www.youtube.com/watch?v=HPEE_o4oud8
Click to expand...

This video

omgggg wings is amazing

I typically play support veno though so as much as I may want to peace out and jungle I'm usually needed to be a shitstain in lane. 4-4-1 arcanes -> veil build for lyfe

NormalFish said:
Pumpkin cheesecake sounds vile. Moblin you need pie in your life.
Click to expand...

This pie you speak of.... it intrigues and interests me. If it comes in the strawberry rhubarb variety it may be worth examining further.

As it stands cheesecake remains the greatest creation on earth, and this has been proven through centuries of difficult and expensive science.
 
B

Brodnax

Member
Winning games means that you are picking, itemizing, rotating and communicating accordingly. So of course MMR it's a measure of skill, the issue is that there are people that while skilled enough for their mmr will still tilt if playing from behind or after losing a big lead.
Also there's a reason why boosting/selling accounts is a profitable practice so you end up having alot of players with these accounts that just feed MMR constantly to other unskilled players.
 
1

1.09

Low Tier
There's a pretty strong relation between MMR and skill, as long as said player plays an adequate amount of ranked games. I'll go more into this if this is still a topic later, but idk how people can say mmr has no bearing on a player's individual skill.
 
M

M.D

Member
Elimination mode is so boring compared to the previous one they had.

It can't be as exciting because every hero has been in the meta at some point compared to back then, but at least pick up some weird ass lineup.. they are playing it like its a normal tournament
 
C

Chris R

Member
M.D said:
Elimination mode is so boring compared to the previous one they had.

It can't be as exciting because every hero has been in the meta at some point compared to back then, but at least pick up some weird ass lineup.. they are playing it like its a normal tournament
Click to expand...

Thank you for this post, would have missed it otherwise since they took a two week break in the middle of the tourney for some reason
 
I

inkls

Member
In ranked players aren't static components either.

Mindset can vary based on factors like losestreak winstreak, being matched with the same player who flamed the entire game in the previous game.

Hero pool, how comfortable you are with a role as well. I mean, would you tell me !Attacker's mmr doesn't represent his skill level because he spammed Kunkka? Or Admiral Bulldog because he spammed Lone Druid and Nature's Prophet?

I mean, when a hero spammer can make it pro and win TI, then you either admit mmr is accurate or that the meta was soooo broken that he won just because of those heroes.
 
shira

shira

Member
M.D said:
Elimination mode is so boring compared to the previous one they had.

It can't be as exciting because every hero has been in the meta at some point compared to back then, but at least pick up some weird ass lineup.. they are playing it like its a normal tournament
Click to expand...

Next Moonduck tournament: Twitch chat plays Dota
 
G

G.ZZZ

Member
Ultrabum said:
Knowing that picking the same hero over and over again will win you games is a skill.

MMR is the best measurement of skill.
Click to expand...

I agree. We're not talking about how good is someone at playing every position and in every condition. I'm talking mostly about the ability to play a game of dota, recognize your strenght, your weaknesses, recognize the general conditions of the game (heroes picked, lanes, form an overall strategy etc....), and win accordingly. Being an hero spammer can be seen as extreme but it's just another form of skill, that can however be neutralized to a certain point (now with the AP ban and someone else taking the hero before you, or a team composition that don't allow for the hero in question). Those games where you get countered in said way, will lower you MMR accordingly and will account for the flaw in your play stile. Just like every other game.

Just think of any other sport really. Being a good goal keeper , defender, midfielder, wing, or striker don't mean that you're not "really" good at soccer in general. That's stupid.
 
K

kionedrik

Member
It's not a question of a hero spammer being skilled or not, it's a question of if he's more skilled than a non-spammer just because his MMR is arbitrarily higher due to it. He might have specialized in a couple of heroes but, in my eyes, he is not automatically more skilled than someone that's 1k MMR below him.
My point is, MMR is not the be all and end all. Skill plays a part, sure, but the system can be gamed with, more or less, cheap tactics. Remember Zeus and Oracle abuse for instance.
 
I

inkls

Member
kionedrik said:
It's not a question of a hero spammer being skilled or not, it's a question of if he's more skilled than a non-spammer just because his MMR is arbitrarily higher due to it. He might have specialized in a couple of heroes but, in my eyes, he is not automatically more skilled than someone that's 1k MMR below him.
My point is, MMR is not the be all and end all. Skill plays a part, sure, but the system can be gamed with, more or less, cheap tactics. Remember Zeus and Oracle abuse for instance.
Click to expand...

what about remembering when the zeus and oracle stuff was removed?

You can prove if hero spammers deserve their mmr by looking at how attacker and bulldog lost 1k mmr after their respective heroes were nerfed right?
 
S

SecretMoblin

Member
idk 1k is actually kind of a lot, if two players with a lot of games played have that big of a difference then I'd certainly consider skill as being the major distinguishing factor between the two of them
 
K

kionedrik

Member
inkls said:
what about remembering when the zeus and oracle stuff was removed?

You can prove if hero spammers deserve their mmr by looking at how attacker and bulldog lost 1k mmr after their respective heroes were nerfed right?
Click to expand...

I'm guessing they didn't lose that much MMR or any at all by the way you said it (I don't follow any of them, or any player for that matter). No I can't prove they don't deserve to be there nor can I prove that they do. But I also recall clearly when the troll era ended seeing some massive MMR drops being posted around (mainly reddit) because those that rose with OP hero spamming were unable to compete with those that were rightly in that bracket. I have no idea what's more common, keep the MMR or lose it after the fotm fever has passed. I believe it might be more common to lose a significant part of that MMR but I can also be completely wrong, what I'm certain though is that a 2 person sample doesn't prove anything.

1.09 said:
1k mmr is alot lol
Click to expand...

I might have exaggerated, sorry, but is the difference that noticeable in the highest brackets? Can you really tell a difference between a 7k and a 6k or a 7k and a 8k?
The lower we go I have no doubt that it's perfectly noticeable.
 
I

Imp the Dimp

Member
Been losing a lot lately. I should stick to playing PS4 Pro probably. Really happy about the FFXV release next week. I think the rumours that it does tighten up at some point as opposed to just being a ten million square miles map are in fact good. Open worlds the way Xenoblade and FFX12 handled them are still acceptable and fun to me, but exploration should absolutely not turn into a second job. Having a predisposition for 100% completion ruined multiple open world games for me. I can't do anything about it.
 
You must log in or register to reply here.

Similar threads

Zenless Zone Zero | Review Thread
Analysis Review 2
82
8K
LordOfChaos replied
A Diablo III Help Topic.
Community
44
2K
FreeY$L replied
Kotaku: Skull And Bones Is Less The Pirate Life And More The Division (Beta Review)
Analysis Review 2
52
8K
deeptech replied
Fairy Tail 2 | Review Thread
Analysis Review
12
1K
graywolf323 replied
The 12 days of Gafmas |OT| The year in thread popularity!
28
1K
jm89 replied
Top Bottom