![]() ![]() I don't work on data sets like this (time series are always a bit funky to me) however I do data analysis in R daily for my work and have an MS in the topic if you'd like to spit ball some other ideas I'd be happy to help. ![]() So if you are ONLY looking at win/loss then 20 is likely fine, but if you are interested in a multiple regression model with interaction terms then your sample size might need to be larger (I didn't do the math). if a radiant T1 falls first, how likely is it that dire will win? If dire takes two towers before radiant takes 1 how likely is it for dire to win? The sample size necessary for an analysis is entirely based on the standard deviation of your sample. Given the research question I listed in (3), you could look at this as just categorical data (14 wins for Radiant vs 6 wins for dire), or go into the interactions which is what I think you are starting to do with the percentages listed. (4) This is going to be a short rant on power given how many comments are saying you need more data. In this case I would guess that the main question is does radiant win more often than dire? It really comes down to what the research question is however. Melee Weapon Attack: +4 to hit, reach 5 ft., one target. The wight can have no more than twelve zombies under its control at one time. (3) From a stats standpoint there are a lot of interaction terms which could be relevant for how the data should be analyzed. A humanoid slain by this attack rises 24 hours later as a zombie under the wights control, unless the humanoid is restored to life or its body is destroyed. (2) I would block "Top 1, Bot 1, and Mid 1" together, followed by 2's, then 3's so that it is more intuitive in a time frame. (1) I see how you formatted the first 4 games in a double row style showing when each building falls, but why not continue that for every game? This is neat, however I have some questions and suggestions: 50% on dire.ģ- When the first set of rax are down, there is an 75% chance of losing for radiant, while dire has 100% of losing)Ĥ- All games have been won within 10 minute interval (50:22 to 59:28) with the average being 52:45.Ħ- If Radiant's 45 minute pushes are dead, there is %75 chance of losing for Radiant.ĮDIT: Removed Tower Awards because it turns out they are all shitty towers, except Radiant's tier 3 mid, that tower is op. However, mid rax rarely fall (only twice in 20 games, and both were dire's).Ģ- Hero (Anti-Mage) has no observable effect on how the match goes, 50% of the time he was on radiant. Here are some basic data about trails with game ID'sġ- Mid towers always fall first. Pinnacle Cup: Malta Vibes #3 5h 7m LVLUP MarsbetĬontinuing what we started before, we have 20 samples so far (will probably go for 75-100 in the next few days) Pinnacle Cup: Malta Vibes #3 2h 7m NAVI J RSĮuropean Pro League Season 12 5h 7m KZ SR HIGHLIGHTS: 9PANDAS vs BetBoom Team - DREAMLEAGUE S20 135 ESL_DOTA2 Live! jikroy & IG.zeal vs SPAWN.DH QUỶ 186 DotaTV247 Common offlaners are heroes such as Batrider, Beastmaster, Dark Seer, and Tidehunter.Livestreams rose looking for jack 329 febbydoto ![]() These heroes are usually referred to as "offlaners". Certain heroes are suited to this solo role, due to having escape mechanisms in addition to spells that can be used to farm. When running a Jungler or a tri-lane, often the Hard Lane is left to a solo hero. TobiWanKenobi and Purge ) use opposite definitions of these terms, referring to the distance between the natural creep equilibrium (where the creeps initially meet up, assuming no blocks) and the tier 1 tower (where safe lane would be short lane). The terms long lane and short lane historically represented the distance between the tier 1 tower and the ancient (where safe lane would be long lane), however many people (e.g.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |