On February 25 2014 14:37 Ryder. wrote: Just wanted to bring this up for discussion to see what people think...
Whenever I see a post about people saying they are stuck in ELO hell, allies all suck, want to improve win rate etc I see people often give the advice to 'pick a snowball mid and carry the game from there'. But when I looked on dotabuff winrates some of the most common snowball mid heroes have winrates of; QoP: 43.02% Storm: 45.08% Invoker: 46.04% TA: 45.01% Puck: 44.51%
Whereas you get heaps of supports with win rates like Lich: 56.58% Venomancer: 53.58% CM: 53,29% Venge: 54.26% Omni: 53.98%
Yeah maybe if you are the top 1% you there's a decent chance you are the best player in each of your pub games and you are good enough to pick one of these mid heroes and snowball the game from there, but for your average joe wouldn't better advice be to pick a support/utility hero like the ones I listed?
The list is actually quite similar. Some of the supports I listed weren't top 15 on dotabuff either but they are definitely considerably higher than the solo mids I listed, and other supports (lich, omni, abba, tree, ES) are still right up there. Not a single snowball mid I mentioned is in top 15 for either...
I guess the moral of the story is that most people aren't as good as they think they are. I still think the generic advice of 'pick solomid and snowball' is inappropriate for 99% of players.
Incidentally my favourite position/hero is support Rubick, and I think at least with certain supports its possible you can still have a high impact if you can find some farm/levels around the map.
It's inappropriate advice for actually getting better, but actually says a lot about the player's skill if they are unable to win and supposedly the best player in the game.
On February 25 2014 16:17 rebdomine wrote: I usually just pick LD so that I can go any lane. Or even jungle if all the lanes have farmers. Quad core greedy lineup go go.
same with NP, offlane/jungle into early global medik (mek carrier)
On February 25 2014 16:15 Ryder. wrote: I guess the moral of the story is that most people aren't as good as they think they are. I still think the generic advice of 'pick solomid and snowball' is inappropriate for 99% of players.
It's really a relic of rating systems from DotA 1 that started people at a baseline rating and didn't have "placement" games, which meant that invariably you had to play a large number of games with/against players that actually were nowhere near as good as you. In that context, playing these heroes make sense because reducing the number of games required to reach your appropriate rating made a big difference.
With Valve's implementation of ranked, you start much closer to your appropriate MMR because you've played a ton of normal games before entering ranked, and placement games exist to iron out discrepancies. The majority of players are playing games very close to their skill level by the time their placement games are over, so there's no grind like there was in more archaic matchmaking systems.
On February 25 2014 16:15 Ryder. wrote: I guess the moral of the story is that most people aren't as good as they think they are. I still think the generic advice of 'pick solomid and snowball' is inappropriate for 99% of players.
It's really a relic of rating systems from DotA 1 that started people at a baseline rating and didn't have "placement" games, which meant that invariably you had to play a large number of games with/against players that actually were nowhere near as good as you. In that context, playing these heroes make sense because reducing the number of games required to reach your appropriate rating made a big difference.
With Valve's implementation of ranked, you start much closer to your appropriate MMR because you've played a ton of normal games before entering ranked, and placement games exist to iron out discrepancies. The majority of players are playing games very close to their skill level by the time their placement games are over, so there's no grind like there was in more archaic matchmaking systems.
even then it took less than 10 games for very high and like 20 games for smurfs to get to more or less their real ratings
so meepo is bugged as shit, like holy shit, so many bugs were introduced in this patch, some of them are actually really strong some of them are bad.
so all meepos have 100 percent stat sharing without aghs illusion runes create new real meepos that don't have boots and don't have stat share nor any abilities, if this meepo dies it crashes server if you die with gem, it does not drop the extra meepo you get when you get aghs does not have abilities but does have boots.
there are probably more but that's what Iv'e found so far.
On February 25 2014 16:15 Ryder. wrote: I guess the moral of the story is that most people aren't as good as they think they are. I still think the generic advice of 'pick solomid and snowball' is inappropriate for 99% of players.
It's really a relic of rating systems from DotA 1 that started people at a baseline rating and didn't have "placement" games, which meant that invariably you had to play a large number of games with/against players that actually were nowhere near as good as you. In that context, playing these heroes make sense because reducing the number of games required to reach your appropriate rating made a big difference.
With Valve's implementation of ranked, you start much closer to your appropriate MMR because you've played a ton of normal games before entering ranked, and placement games exist to iron out discrepancies. The majority of players are playing games very close to their skill level by the time their placement games are over, so there's no grind like there was in more archaic matchmaking systems.
Although this is kinda true sticking to one role and a limited pool of heroes "artificially" inflates your mmr by a lot, its possible to increase your rating by upwards to 1000 by doing that (have multiple people on my fl who did that from ) and most of those heroes are, in fact mid heroes because they can snowball a lot harder than others, one big example would be qojqva, he picks a ton of tinker/storm and has like 6.7k rating after starting at ~6k. There are also many instances on my friend list where someone got from ~4.5k to 5.5k by playing slark or (old) alchemist.
Obviously this also has to do with certain heroes beeing stronger than the average but overall mastering a small pool of heroes on a certain role will increase your rating.
I'd say that the supports have higher winrates because when someone picks a support in AP (which doesnt happen often enough), that team will atleast have a chance that someone wards and do other usefull stuff that helps to gain an advantage in AP. I have had plenty (if not most) of AP games with 5 mid/carry team. It maybe a good idea if the winrates could be made gamemode dependant, which will give the winrates on dotabuff and other sites more value.
The mid heroes with low win rates aren't that good in sidelanes splitting farm/exp which is what often happens when people random or you have multiple players picking mid heroes to carry the game. Add that to people not really knowing how to snowball when they do play mid and you get a 45% win rate etc..
On February 25 2014 20:17 rob.au wrote: The mid heroes with low win rates aren't that good in sidelanes splitting farm/exp which is what often happens when people random or you have multiple players picking mid heroes to carry the game. Add that to people not really knowing how to snowball when they do play mid and you get a 45% win rate etc..
I reckon it comes down to this honestly. Simple fact is that you're likely to end up with those particular heroes laning from time to time because of fights over the mid. Couple that with people being unable to snowball properly and you get a low win percentage.
HOWEVER for people who really are better than their team a snowball mid will tend to win them the game more than any other role.
Although I'd disagree on including Puck there. I always feel Puck is much more heavily reliant on the team than a lot of mids; because its more of an initiation mid role with Dream Coil and Waning Rift. As opposed to QOP, Storm or TA who all just go around killing people. Or Invoker who can take on a bunch of roles.
On February 25 2014 21:38 -Celestial- wrote: HOWEVER for people who really are better than their team a snowball mid will tend to win them the game more than any other role.
EPT
