EPTPM me please, also if I can join the garena clan or whatevers.
[HoN/DotA] Let's Play~!! - Page 582
| Forum Index > General Games |
|
Shikyo
Finland33997 Posts
PM me please, also if I can join the garena clan or whatevers. | ||
|
paper
13197 Posts
| ||
|
SK.Testie
Canada11084 Posts
| ||
|
Dark.Carnival
United States5095 Posts
| ||
|
JeeJee
Canada5652 Posts
| ||
|
Tadzio
3340 Posts
"What better way to figure out which heroes are truly overpowered," I thought to myself. Sure, it might be more important to know which heroes were being used in high PSR games, but fuck it, this'll satisfy the idle stats nerd dwelling within me  The Duds! (Heroes with less than 45% wins) Nighthound | 44.1% | 5.9 / 4.7 / 5.5 Dark Lady | 43.7% | 4.9 / 5.4 / 6.1 Warbeast | 43.6% | 4.1 / 4.6 / 4.3 Ophelia | 43.3% | 3.6 / 4.9 / 5.4 Wretched Hag | 42.1% | 5.1 / 5.2 / 7.1 Pandemonium | 41.7% | 3.9 / 5.9 / 6.3 The Dudliest! + Show Spoiler + Devourer | 40.7% | 4.4 / 5.3 / 6.3 The Ubers! (Heroes with more than 52% wins) PuppetMaster | 52.0% | 6.4 / 5.2 / 6.8 Andromeda | 52.3% | 4.5 / 5.3 / 7.6 Maliken | 52.6% | 5.6 / 5.6 / 7.7 Demented Shaman | 53.3% | 4.6 / 5.2 / 8.4 Jeraziah | 54.3% | 3.1 / 4.7 / 6.6 Tempest | 54.9% | 3.8 / 5.2 / 9.9 Sand Wraith | 57.7% | 5.9 / 4.5 / 10.7 Pestilence | 57.8% | 4.3 / 4.5 / 8.7 The Ublierest! + Show Spoiler + Plague Rider | 58.1% | 6.7 / 4.9 / 8.5 Anyway, I thought it was a fun collection of stats. For those that are curious, only 3 Agi carries had better than a 50% win rate. Sand Wraith is already on the Ubers list, and the other two were Arachna with 51.x% and Magebane with 50.x%. Flex your nerd powers and discuss. :D | ||
|
jpopidol
108 Posts
| ||
|
SK.Testie
Canada11084 Posts
| ||
|
Tadzio
3340 Posts
| ||
|
Renderer
United States30 Posts
On October 22 2009 06:36 Tadzio wrote: Just went through the hero stats for HoN available here and compared all of their winning and losing percentages. When I noticed they had the winning percentages for all heroes over the courses of 10s of thousands of samples (otherwise known as games) it struck me most error and bias would be removed from these stats. They didn't care about the skill of the player wielding the character, nor did they care whether the hero was a teamfighter, pusher, or a 1v1 hero. The only thing that mattered was whether the team they were on had won. Also, S2 had kindly ignored EM games when tallying the stats. "What better way to figure out which heroes are truly overpowered," I thought to myself. Sure, it might be more important to know which heroes were being used in high PSR games, but fuck it, this'll satisfy the idle stats nerd dwelling within me The Duds! (Heroes with less than 45% wins) Nighthound | 44.1% | 5.9 / 4.7 / 5.5 Dark Lady | 43.7% | 4.9 / 5.4 / 6.1 Warbeast | 43.6% | 4.1 / 4.6 / 4.3 Ophelia | 43.3% | 3.6 / 4.9 / 5.4 Wretched Hag | 42.1% | 5.1 / 5.2 / 7.1 Pandemonium | 41.7% | 3.9 / 5.9 / 6.3 The Dudliest! + Show Spoiler + Devourer | 40.7% | 4.4 / 5.3 / 6.3 The Ubers! (Heroes with more than 52% wins) PuppetMaster | 52.0% | 6.4 / 5.2 / 6.8 Andromeda | 52.3% | 4.5 / 5.3 / 7.6 Maliken | 52.6% | 5.6 / 5.6 / 7.7 Demented Shaman | 53.3% | 4.6 / 5.2 / 8.4 Jeraziah | 54.3% | 3.1 / 4.7 / 6.6 Tempest | 54.9% | 3.8 / 5.2 / 9.9 Sand Wraith | 57.7% | 5.9 / 4.5 / 10.7 Pestilence | 57.8% | 4.3 / 4.5 / 8.7 The Ublierest! + Show Spoiler + Plague Rider | 58.1% | 6.7 / 4.9 / 8.5 Anyway, I thought it was a fun collection of stats. For those that are curious, only 3 Agi carries had better than a 50% win rate. Sand Wraith is already on the Ubers list, and the other two were Arachna with 51.x% and Magebane with 50.x%. Flex your nerd powers and discuss. :D Well, unless you are running an actual multiple regression on the population data, you cannot know if there are hidden bias on these heroes. What this means is that you need econometrics and common sense to figure out whether hero preference by low-skilled players or popularity of a hero can affect these numbers and quantize the effect on the effectiveness of the hero ceteris paribus. Coming from an economics major, the data you have provided does not show the complete picture, and it does not answer the question completely: What heroes are uber (higher probability of win independent of the player's skill level)? However your data is interesting, and makes for a good thought exercise. | ||
|
Renderer
United States30 Posts
TL;DR: I need the spreadsheet with all the data for every game to run my nerdy formulas on. | ||
|
Dark.Carnival
United States5095 Posts
| ||
|
Renderer
United States30 Posts
u can't beat my logic. have some fish. http://mdn.mainichi.jp/mdnnews/odd/news/20091014p2g00m0in012000c.html | ||
|
Tadzio
3340 Posts
Unfortunately, we can't know if Pesti is always a good pick. We can be pretty sure, though, that he's a better pick than Pandemonium most of the time. | ||
|
paper
13197 Posts
does not really say anything other than what your average competent player already knows | ||
|
Manit0u
Poland17843 Posts
And Pesti is a good pick in most cases because he's very versatile. He can tank, he can dish damage, he can stun, he can reveal enemies + armor reduction and finally, he can run really fast. There's nothing wrong with this hero. | ||
|
Shikyo
Finland33997 Posts
HoN is a joke btw. | ||
|
Shade692003
Canada702 Posts
| ||
|
Dark.Carnival
United States5095 Posts
just look at its name. LoL. LoL. L O L. WHO WOULD DO THAT TO THEIR GAME? only jokesters. | ||
|
Shikyo
Finland33997 Posts
On October 22 2009 09:49 Shade692003 wrote: you're a joke ? It is super imbalanced and has only 3% of the heroes, its own heroes are terrible and its otherwise just a copy with some terrible names in there. LoL is of course a lot worse but there can be bad jokes and less bad jokes. | ||
| ||
