The new forums will be named Coin Return (based on the most recent
vote)! You can check on the status and timeline of the transition to the new forums
here.
The Guiding Principles and New Rules
document is now in effect.
[Heroes of the Storm] There is no new hero, ONLY XUL.
Posts
If she can snipe, cloak and relocate, and then snipe again, she would be a much more fun and interesting character, and she would have more counterplay for the enemy team AND she would have a greater presence during team fights than she does now.
Right now Nova's presence is this: Delete a character and run like hell while waiting on cooldowns.
Take away her burst, let her re-cloak, and make her more of a ranged fighter than a one-trick pony.
Not a bad idea. This would add more interesting play with Decoy too.
Gotta remove GP for sure then though.
If Nova could recloak on command you completely negate the downside of GP and the talent becomes even more massively unbalanced.
3DS Friend Code: 3110-5393-4113
Steam profile
Diablo makes far too much sense to do this why does he use mana?
I really feel the orbital strike should be a nuke so she is more like a ghost in StarCraft as well as using mind tricks
The ability to cloak and uncloak would be beyond useful because as it was said she is not as powerful as you think and she could pop off a shot and then move
And you want to reduce it by another 10% while making GP mandatory. I mean, if GP is mandatory on any hero then you fucked up that talent tier and need to throw in a safer but less impactful alternative.
3DS Friend Code: 3110-5393-4113
Steam profile
But they didn't give her manual cloaking when they did that. I like the idea a lot. I can see a lot of potential in a sustain damage Nova who's constantly recloaking and re-positioning around the battlefield.
Switch: 6589-6405-3399
Abathur: Me - 56.5%, Diamond avg: 45.1% (46 games)
Sylvanas: Me - 68% Diamond avg: 45.6% (25 games)
And amusingly my Kael'Thas is at 40%, not so much because I suck at him (imho) as much as that I need to play tanks in HL to win and I've been using KT a couple times to middling success.
Switch: 6589-6405-3399
I went and looked this up on hotslogs and just .. wow. That front page says it all about the state of the game right now. Only 15 heroes above 50% and the top like 5 or 6 are at stupidly high win rates and pick rates.
Diablo
Tyrande
Raynor
Zagara
Kael
Just picked tons and blowing shit up most games.
Look, if 83% win rate with Li Li is wrong I don't want to be right!
Switch: 6589-6405-3399
This seems to point to a no-support meta. What do you think? Makes sense with all the burst. It's better to be able to kill fast than heal slowly. We've all played that meta in QM of course, but I played in a HL match the other day where the other team went support-less and they beat us. My team wasn't great, but I'm also wondering if there's something to it....
Switch: 6589-6405-3399
Then we could start talking about whether HotSlogs is worth looking at.
I always upload my matches and often someone beats me to it.
Every upload puts 10 people in the system.
I wouldn't use Hotslogs to tell people their hero's win rate is crap so they should uninstall but then again, I generally not an asshole.
Some people take statistically underwhelming characters and make them shine. Lunara is a great example. She is legit underpowered but some people can make her work well enough to even be considered good. Sure, some heroes work better with and against any comp but oh well.
I usually care about comp in hero league only. It will be nice when they fix their matchmaking. Even so, I still enjoy the game a lot.
I also just lost w/ support vs no support. Our tank wasn't the best, but still wouldn't have guessed that...
3DS Friend Code: 3110-5393-4113
Steam profile
It shouldn't really matter since the primary stats it considers are from hero league and you can only have a max of 2 people in a group, so even if it got its data from just a few people, you'd be collecting useful statistics on 8 other random people at the same time.
However, I'm sure if you asked on reddit or directly to the hotslogs author (who is very responsive to emails), he'd be able to answer that question. I've looked at enough people's match history to be satisfied that hotslogs covers a statistically significant amount of data, and due to match making I think it pretty much has to be accurate as well. The only way I could see it not being so is if there really was a strong trend of people not uploading loss matches, and also that the uploads only come from a small subsection of the user base.
Switch: 6589-6405-3399
which means that it is a good representation for the people in this thread that use it!
:P
Also, I'm tired of this argument. Someone should do a simulation to figure out how many % of people need to upload to hotslogs before it gets data on most players. I bet it's smaller than most people think.
Or maybe it's really high and we can be more confident that it's not super valuable data.
3DS Friend Code: 3110-5393-4113
Steam profile
That's not even remotely touching on what I am getting at when I say "Hotslogs Confirmation Bias"
Ehhh...if you have a 40% winrate with a champion after a bunch of games, it does mean that you objectively suck at that champion.
--for KT, it might mean that you have the correct instincts and mechanics to play him at HL 5 when your team is more with it, but that you just suck at playing him at HL 20. Still, losing games consistently=you are bad at that dude in that setting. There really no way around this. Insisting that you are good at a champion when the numbers don't line up impedes progress. (So, admit that what you are doing with him now isn't working, whether or not it might work with better teammates, and change your playstyle accordingly).
--...I suck at Lili now, which is embarrassing because how can that even be possible, but that is what my win% tells me. Part of analyzing and optimizing one's own play is facing uncomfortable numbers, unfortunately.
You don't need "most" data in order to have a good sample. If there's a problem with Hotslogs data, its because of the sampling method (self-selected volunteered data) and not the quantity.
I would say the best evidence that hotslogs data is reasonably accurate is simply that all of the characters that have been buffed / nerfed were correspondingly low / high in the hotslogs win rate list.
I think HotSLogs is probably sort of accurate for the highest level trends, i.e. global winrates. For MMR though it can't be trusted at all.
oh yeah I mean basically what I am saying is that i am pro and supagoat is nub
However, my Tyrande is 70% winrate, and Diablo is 100% winrate. So fair. So balanced.