|
Actually they're pretty useful.
|
# ? Jul 31, 2018 18:35 |
|
|
# ? Jun 5, 2024 08:48 |
|
man just had one of those games with four strangers where everyone gets in comms, we discuss picks and bans, then we coordinate in game and win by picking good team fights and keeping up macro pressure. really makes up for all the games where first pick slam nova says “gg fed” one minute into the game and proceeds to gently caress off into the ether
|
# ? Jul 31, 2018 18:37 |
|
No Wave posted:Actually they're pretty useful. What are win rates useful for? You need a bunch of other information before they can provide you with any useful insights. For example, Bloodlust, Pyroblast and Ultralisk both have higher win rates in Master League (and possibly lower as well). There's extremely heavy selection bias when you're looking at win rates, especially when one talent is much better than the other. https://www.hotslogs.com/Sitewide/HeroDetails?Hero=Rehgar https://www.hotslogs.com/Sitewide/HeroDetails?Hero=Kerrigan https://www.hotslogs.com/Sitewide/HeroDetails?Hero=Kael%27thas
|
# ? Jul 31, 2018 18:46 |
|
bamhand posted:What are win rates useful for? You need a bunch of other information before they can provide you with any useful insights. All the examples provided have incredibly marginal differences in win rate, so of course they don't give any valuable information.
|
# ? Jul 31, 2018 18:49 |
|
Pyroblast is absolutely a legitimate ultimate (as is ultralisk). NexEternal was the player who inspired Kerrigan's popularity with tempo storm in HGC last year and he swore by it. Like calling pyroblast a bad ult always just shows poor game knowledge. Now, rocket ride? That's a dogshit ult, with winrates to match.
|
# ? Jul 31, 2018 18:52 |
|
pyroblast, in situations where it’s good, is very good. it’s more narrow than Phoenix but if the enemy team has squishies without Ice Block and doesn’t have Medivh, it’s a very legitimate choice
|
# ? Jul 31, 2018 18:53 |
|
DAD LOST MY IPOD posted:pyroblast, in situations where it’s good, is very good. it’s more narrow than Phoenix but if the enemy team has squishies without Ice Block and doesn’t have Medivh, it’s a very legitimate choice Exactly, that skews the win rate a bunch. I'm not saying the talents are hands down bad but they are all situational picks and without a ton of additional information like your team comp, enemy team comp, map, etc. the win rate is not going to give you any useful information.
|
# ? Jul 31, 2018 18:55 |
|
Exactly what? Bloodlust is played 1/10 as much with a worse winrate, which means, "don't bother". Pyroblast has a similar win and pick rate, which means, "maybe bother".
|
# ? Jul 31, 2018 18:58 |
|
No Wave posted:Exactly what? Bloodlust is played 1/10 as much with a worse winrate, which means, "don't bother". Pyroblast has a similar win and pick rate, which means, "maybe bother". Phoenix is picked twice as often as Pyro if you filter for master but Pyro has a 6% higher win rate. e: Drakken Laser has a whopping 9% higher win rate on Tychus but Odin is picked 4x as often. And there are plenty of cases where the an undeniably good talent has the higher win rate than the "bad" talent. My point is there isn't a strong relationship, positive or negative, between win rate and how "good" talent is. Also I'm not saying pyro or laser is bad, only that there is a huge difference between their pick rate and their win rate. bamhand fucked around with this message at 19:07 on Jul 31, 2018 |
# ? Jul 31, 2018 19:01 |
|
There absolutely is. You've just created an insane definition of "good" that's context independent (and therefore worthless).
|
# ? Jul 31, 2018 19:08 |
|
bamhand posted:Phoenix is picked twice as often as Pyro if you filter for master but Pyro has a 6% higher win rate. Just looking at Win Rate is stupid, but if you look at Pick Rate and Win Rate together it'll generally tell you if a talent is Usually Good, Situationally Useful, or Trash. It won't tell you WHAT situation it's good in without further digging, but it gives you a basic breakdown.
|
# ? Jul 31, 2018 19:08 |
|
Relentless posted:Just looking at Win Rate is stupid, but if you look at Pick Rate and Win Rate together it'll generally tell you if a talent is Usually Good, Situationally Useful, or Trash. Ok that's fair. No Wave posted:There absolutely is. You've just created an insane definition of "good" that's context independent (and therefore worthless). And as a follow up to further digging, how do you gain context? Look around for pro opinions? Check out HGC games? Ask in some forums or reddit? At that point it's all these additional sources giving you info about whether and when a talent is good, the win rate itself really doesn't tell you too much. Like, how exactly does "this talent wins more games in specific situations, but we can't tell you what those situations are" help someone?
|
# ? Jul 31, 2018 19:18 |
|
rocket ride is way more fun tho
|
# ? Jul 31, 2018 19:18 |
|
bamhand posted:And as a follow up to further digging, how do you gain context? Look around for pro opinions? Check out HGC games? Ask in some forums or reddit? At that point it's all these additional sources giving you info about whether and when a talent is good, the win rate itself really doesn't tell you too much. Like, how exactly does "this talent wins more games in specific situations, but we can't tell you what those situations are" help someone?
|
# ? Jul 31, 2018 19:21 |
|
No Wave posted:Because you have a brain man. Should be pretty easy to determine when pyro or merc lord might be a good pick. Is it though? Obviously you don't use pyro vs people with invincibility or a disjoint. And phoenix is good for area denial. But what if you're on a map where you need to deny an objective (say Volskaya) but the enemy team also has a juicy Pyro target? A lot of situations have a ton of different factors going into it and the analysis is pretty complex. Pick rate will say as a general rule of thumb Phoenix will work better for you more often than not. Maybe not in your particular situation but if you want to play the odds Phoenix gives you the best odds.
|
# ? Jul 31, 2018 19:29 |
|
I rarely even pick talents, they seem kinda op.
|
# ? Jul 31, 2018 19:29 |
|
bamhand posted:e: Drakken Laser has a whopping 9% higher win rate on Tychus but Odin is picked 4x as often. I've been farting around with Tychus (and losing a lot because I keep forgetting that while Odin makes you look like a big stompy beast it doesn't actually add any HP, and I already have a tendency to dive too much. Is Drakken Laser particularly useful beyond maps with prolonged action on one point like Volskaya and Shrines? From what I've seen, it's generally best in the same situations that make Alexstrazsa appealing for Dragon Queen. I could see it being huge in a desperate defense of your core or pushing down a keep, but on most maps Odin seems to have more utility.
|
# ? Jul 31, 2018 19:30 |
|
As someone who is deep into sports analytics, watching this discussion unfold is so familiar. Metrics are great, and absolutely have to be applied correctly. A bit easier in sports, I will admit, where the number of complex and edge-case situations are much lower. Also don't be the KT who picked Pyro and used it on me (Murky) 3 times.
|
# ? Jul 31, 2018 19:32 |
|
I honestly don't know when laser is ever useful. Odin lets you attack people from a safe range so it's great in a lot of the situations you mentioned since you can fire safely from afar on the enemy team.Nybble posted:As someone who is deep into sports analytics, watching this discussion unfold is so familiar. Metrics are great, and absolutely have to be applied correctly. A bit easier in sports, I will admit, where the number of complex and edge-case situations are much lower. Yeah, I'm a statistician and making sure you're applying stats correctly is a huge, non trivial deal. A lot of stuff may seem super obvious but it really isn't. Relentless posted:Data mining! Context clues! Right all of those things come from somewhere besides the win rate you're looking up. In the end you're mostly relying on stuff outside of the statistic you're looking at and that win rate number really isn't telling you much. bamhand fucked around with this message at 19:36 on Jul 31, 2018 |
# ? Jul 31, 2018 19:33 |
|
bamhand posted:Ok that's fair. Data mining! Context clues! This game is big and complex enough there's no complete list. Ask around. Learn heroes one at a time. Guess and suffer for your mistakes.
|
# ? Jul 31, 2018 19:33 |
|
You're a statistician and you think there are no possible use for hots talent statistics? That is um unfortunate.
|
# ? Jul 31, 2018 19:37 |
|
No Wave posted:You're a statistician and you think there are no possible use for hots talent statistics? That is um unfortunate. Are you just purposefully trying to be an rear end? I've clearly said pick rate is a great rule of thumb to go by. I've given a ton of examples and context where win rate tells you very little, if anything. Pick rate is a statistic by the way, in case you weren't clear on that.
|
# ? Jul 31, 2018 19:40 |
|
Corsair Pool Boy posted:I've been farting around with Tychus (and losing a lot because I keep forgetting that while Odin makes you look like a big stompy beast it doesn't actually add any HP, and I already have a tendency to dive too much. Laser used to be a good pick when you could take In The Rhythm and That's The Stuff together. Nowadays, Odin is the only ult that matters unless your team has an enormous amount of lockdown.
|
# ? Jul 31, 2018 19:56 |
|
I really like bloodlust even though I know it's the 'wrong' choice that first post-20 teamfight is magical
|
# ? Jul 31, 2018 20:00 |
|
I like bloodlust a lot, the one that's a level 13 deckard talent.
|
# ? Jul 31, 2018 20:04 |
|
Bloodlust has been picked lately in HGC when the team has raynor so
|
# ? Jul 31, 2018 20:15 |
|
Star posted:Bloodlust has been picked lately in HGC when the team has raynor so
|
# ? Jul 31, 2018 20:27 |
|
Bogart posted:rocket ride is way more fun tho Yeah and when you're playing Sapper JR he shouts "FOR ALL MY HOMIES!" so it's basically an auto-pick.
|
# ? Jul 31, 2018 21:51 |
|
It is, by far, the worst heroic in the game (and rip tire is one of the best) Puckish scamp appears to be the only level 20 talent in the game with a sub-50% winrate, which is actually an amazing feat.
|
# ? Jul 31, 2018 21:57 |
|
bamhand posted:What are win rates useful for? You need a bunch of other information before they can provide you with any useful insights. This has been brought up before, and it's futile. Even the "LOL, you're saying winrates are no use whatsoever??" strawman has been used before. People loving love Hotslogs winrates.
|
# ? Jul 31, 2018 22:06 |
|
No Wave posted:It is, by far, the worst heroic in the game
|
# ? Jul 31, 2018 22:23 |
|
HitTheTargets posted:Quite a departure from Storm Shadow, that's for sure. It's about time for a female protoss to be in the game but I wish he hadn't called his Adept hero Selendis. She's a carrier to me. They could also use Raszagal or Vorazun and basically put any kind of kit on them they want.
|
# ? Aug 1, 2018 03:45 |
|
While I'm glad that I'm doing a rework post for someone who isn't the loving Lost Vikings this time, I also still want to die because these are so much work. At least half of the effort is tracking down appropriate images.
|
# ? Aug 1, 2018 04:03 |
|
Is there a blizzard issue with people dropping in draft? Had ~10 draft dodges today.
|
# ? Aug 1, 2018 07:17 |
|
Kith posted:While I'm glad that I'm doing a rework post for someone who isn't the loving Lost Vikings this time, I also still want to die because these are so much work. Speaking of Vikings, I've been lucking into playing with Diamond players in Team League lately. Someone was nice and offered me some constructive feedback to improve my play. First line: "No Lost Vikings, ever" I made a sad face. I get it though, I'm guessing that Vikings in leagues above Silver or maybe Gold are just way too easily outclassed unless the entire team is super duper on the ball. Probably just too much effort for what it's worth. (Still hoping for that Viking avatar, if you do have the time. My love for these bearded idiots will not stop) appropriatemetaphor posted:Is there a blizzard issue with people dropping in draft? Had ~10 draft dodges today. I don't know if this was just on his end, but someone in my group kept getting game crashes during draft yesterday.
|
# ? Aug 1, 2018 14:29 |
|
In my (admittedly limited) experience, Vikings have the potential to be good on the right map, but they just have exponentially more potential for gently caress-ups than any other character. Like, if I'm laning against a vike and I see 1-2 of the other ones getting into a team-fight in another lane, I know to push super-hard on the remaining one. Probably I'll straight up get a kill while the player is dealing with the fight, but at the very least I'll bully the guy out of lane and get all the xp while he freshens back up. It's just not worth the risk for the reward imo.
|
# ? Aug 1, 2018 14:39 |
|
i think a lot of players just don't like playing with TLV, like it's not actual hots. It should be free wins vs abathur on a lot of maps I don't think it's a power level question (and I've never seen a viking player as bad as most hanzos). It doesn't seem to get weaker at higher ranks given what I see on GM streams, it's just dumb as hell. No Wave fucked around with this message at 15:05 on Aug 1, 2018 |
# ? Aug 1, 2018 15:02 |
|
In open division this week, an underdog team smashed their opponent (who finished second place in their end) with a last pick vikings on towers of doom.
|
# ? Aug 1, 2018 15:54 |
|
In HL/TL people often don't know how to deal with Vikings and can lead to a bunch of easy wins for you. Still have to pick them on favorable maps though.
|
# ? Aug 1, 2018 15:55 |
|
|
# ? Jun 5, 2024 08:48 |
|
No Wave posted:i think a lot of players just don't like playing with TLV, like it's not actual hots. It should be free wins vs abathur on a lot of maps I don't think it's a power level question (and I've never seen a viking player as bad as most hanzos). Yeah the Vikings are just boring as gently caress is the problem
|
# ? Aug 1, 2018 17:10 |