EPT[Patch 4.4] CUDDLY INCOMING! ヽ(*・ω・)ﾉ - Page 74
| Forum Index > LoL General |
|
FinestHour
United States18466 Posts
| ||
|
cLutZ
United States19574 Posts
| ||
|
Sufficiency
Canada23833 Posts
On March 25 2014 10:22 cLutZ wrote: So, I think Nasus vs. Trundle is the most boring matchup in the game. I just was part of such an abomination. Try Trundle vs Teemo  | ||
|
nafta
Bulgaria18893 Posts
| ||
|
FinestHour
United States18466 Posts
| ||
|
Sufficiency
Canada23833 Posts
On March 25 2014 10:27 nafta wrote: You probably didn't play back when you could stack gold items.Every top lane was worse than that. Well, I think top lane in S2 was much better actually. When you play things such as (say) Irelia vs Jax, trading is actually meaningful. Trading as Trundle vs Nasus? Well... | ||
|
DiracMonopole
United States1555 Posts
On March 25 2014 10:27 nafta wrote: You probably didn't play back when you could stack gold items that also gave a bucketload of HP and armor.Every top lane was worse than that. Fixed for you. RIP HOG | ||
|
cLutZ
United States19574 Posts
On March 25 2014 10:29 FinestHour wrote: neither one is resourceless, it cant be that bad Thats actually the bad part. Half the time we were both out of mana (or close) and couldn't do much other than auto creeps or each other. Also I screwed up because I was trundle and had TP selected before Nasus came out, and was afk during countdown because picks went way faster than I expected. I think I could have gotten kills at lvl 2 and 6 with ignite. | ||
|
Ryuu314
United States12679 Posts
On March 25 2014 10:29 Sufficiency wrote: Well, I think top lane in S2 was much better actually. When you play things such as (say) Irelia vs Jax, trading is actually meaningful. Trading as Trundle vs Nasus? Well... every time I see someone go off all nostalgic about S2 top lane I find that absolutely hilarious. S2 top lane had even less pick diversity/strategy than the current meta, you had super sustain champs in top making it even more of an island than it is now. In terms of build paths, virtually everyone went Force of Atmogs with maybe a Trinity Force for damage. Now, we actually see some semblance of counterpicking and there's more pick diversity. 2v1s are a thing, as are early level 3 dives. Build paths are still relatively uniform across all top laners, though. Also, perhaps most importantly, Yorick is no longer played ever. Personally, I think S3 top lane meta was the best. | ||
|
Sufficiency
Canada23833 Posts
On March 25 2014 10:39 Ryuu314 wrote: every time I see someone go off all nostalgic about S2 top lane I find that absolutely hilarious. S2 top lane had even less pick diversity/strategy than the current meta, you had super sustain champs in top making it even more of an island than it is now. In terms of build paths, virtually everyone went Force of Atmogs with maybe a Trinity Force for damage. Now, we actually see some semblance of counterpicking and there's more pick diversity. 2v1s are a thing, as are early level 3 dives. Build paths are still relatively uniform across all top laners, though. Also, perhaps most importantly, Yorick is no longer played ever. You have a good point, but the playing style of the current fotm top laners are really not that exciting. | ||
|
FinestHour
United States18466 Posts
clearly someone at riot games development team thinks it is exciting | ||
|
Sufficiency
Canada23833 Posts
On March 25 2014 10:49 FinestHour wrote: thats only a matter of opinion clearly someone at riot games development team thinks it is exciting Surely a Trundle and a Nasus wrecking each other for 0 damage is exciting! | ||
|
Kupon3ss
時の回廊10066 Posts
| ||
|
Sufficiency
Canada23833 Posts
https://github.com/jlinn/riot-api-java It works really well. It took me about 1 hour to write a rudimentary program to mine Riot's data. But my algorithm is so bad that it took me 2.5 hours to mine only 20k games (theoretical limit is 75k in 2.5 hours). I'll fix things up as I go. In my defense though, I am tossing out all non-summoner's rift games. Out of the games left, I am getting around 80% ranked soloQ games, 1% ranked team, and the rest normals. I also need to add a few more things to log (currently it only logs for the games, not finer stats such as damage done, kills, etc.). It would be very interesting to also do that and find out (for example) which champion has the highest number of pentakills. | ||
|
cLutZ
United States19574 Posts
You don't want Renekton, mundo, trundle, shyvana all fotm. You want one of those (or shen), jax/irelia, kennen/rumble, singed, etc. Its more interesting to play or watch when that is the case. The issue is that it is also harder to balance in that manner. | ||
|
killerdog
Denmark6522 Posts
On March 25 2014 11:11 Sufficiency wrote: Btw, I am trying this set of libraries: https://github.com/jlinn/riot-api-java It works really well. It took me about 1 hour to write a rudimentary program to mine Riot's data. But my algorithm is so bad that it took me 2.5 hours to mine only 20k games (theoretical limit is 75k in 2.5 hours). I'll fix things up as I go. In my defense though, I am tossing out all non-summoner's rift games. Out of the games left, I am getting around 80% ranked soloQ games, 1% ranked team, and the rest normals. I also need to add a few more things to log (currently it only logs for the games, not finer stats such as damage done, kills, etc.). It would be very interesting to also do that and find out (for example) which champion has the highest number of pentakills. Wait, does that mean that there are 4 times as many ranked games as there are normal games? (80% vs 19%) or is the data you're mining not representative of the overall quantity of games of each type (or am i misunderstanding.) Because that would really surprise me if it's true. | ||
|
AsianEcksDragon
United States1036 Posts
| ||
|
Amui
Canada10567 Posts
RoS comes out in an hour | ||
yamato77
11589 Posts
lulzy | ||
|
Sufficiency
Canada23833 Posts
On March 25 2014 11:42 killerdog wrote: Wait, does that mean that there are 4 times as many ranked games as there are normal games? (80% vs 19%) or is the data you're mining not representative of the overall quantity of games of each type (or am i misunderstanding.) Because that would really surprise me if it's true. Hard to say. The way the algorithm works is as follows. 0. Pick an initial player (I picked myself, because why not), P_0 1. For player P_i, find his 10 recent games. Save them. 2. Pick a player P_i+1 who was an opponent of P_i in P_i's last 10 games. *** 3. Repeat 1-2. So, since my history is full of ranked games, I will be more likely to find players who also play ranked games. In my earlier versions of the program, I actually found a whole bunch of bot games after 50 iterations from myself, which crashed the program entirely. I "fixed" it, by not recording the bot games. For better or worse, the program crashes every 40 minutes or so. I am not 100% sure if it's my bad coding or there is something else funny going on. When it crashes, the program automatically restarts.... from me..... for a fresh run. | ||
| ||
