Top-lane strength over time

tl;dr Top: short games - Wukong, medium games - Wukong, long games - Wukong.

Hello friends. The LeagueMath.com analysis of champion power over time per-lane continues. Two weeks ago I looked at marksman strength over time and last week I looked at mid-lane strength over time. Today I have the summary for solo top lane picks.

The plots in today's post draw data from 1,008,906 matches (Yes, we broke the 1M barrier this week). These games are all NA ranked solo-queue matches played on patches 5.1-5.7 by players of all skill levels.

Each plot below was made in the following way. I first grouped the matches by which champion was solo in the top lane. Then I binned each of those groups by game-length into two-minute buckets from 20-60 minutes. Then for each bucket I calculated an 'advantage' value for that champion in that bin. The advantage is 1 if the team with that champion won all the games in that bucket, -1 if they lost every such game, and interpolated between otherwise.

Disclaimer: These data and plots are in no way meant to suggest that having a certain top-lane champion on your team at a specific time during the game is good or bad. League of Legends is a team game, and the process of making plots like these is fraught with untold statistical, social, and other difficulties. If you are the type of person who likes to get upset and argue about the quality of statistics you find in random posts on the internet, then the plots here are basically guaranteed to provoke your ire. The rest of us can put on a kettle and settle in for some insightful fun.

Here are four plots for champions that do better in games that end earlier rather than later:

All four of these have a distinct downward trend, but Sion is the sharpest of the four for sure. I am a little surprised to see Nidalee here. Remember, these plots are solo top-lane data only. There are no mid or jungle Nidalee games in this analysis. Last week we saw Nidalee's plot was pretty flat when she was picked for mid-lane. This week, it looks like I would prefer not to take Nidalee top into the late game.

Here are four more tanky/bruiser champions with similar plots:

These four are all fairly flat until about 35 minutes, and then take a small dive. Poor Renekton does not do well in the season 5 meta; in this dataset he has a negative advantage at every game length. Hecarim does better, but there does seem to be a bit of a dip for him in longer games. That last bar on his plot should probably be ignored. It is only based on 80 games worth of data. Though, the team with Hecarim did win a juicy fraction of those 80 games. Ha!

Wukong is better than Renekton is bad right now:

He seems to have a sizable advantage at all points in the game. This data agrees with other sources. For instance, at the time of this writing, the excellent site champion.gg has Wukong as the winningest top laner.

Here are three plots of champions that do best when the game does not end too early or too late.

Funny coincidence, their names all start with G. Gnar is the most popular of these three (as measured by number of games played in the dataset), but his win rate is by far the worst. He seems to have his best luck when games go about 40 minutes. Otherwise, the advantage over time for these three does not vary too much by game length.

Here are three plots of champions that either need to win early, or have the game go super-late.

I like that these three plots are so similar, because it matches well with how I think of these three champions. They can each win, either psychologically or actually, in the early game but they have a definite lull around the 40 minute mark. Then, if the game goes long enough, these are three champions that can really affect the late game.

Here are 5 popular top-lane picks that unfortunately defy this type of analysis:

These five are a grab bag of winning and losing picks that do not have a clear regression of strength to any of short, medium, or long games. Take what you want from these, as I know you will. Riven and Irelia seem to be the most reliable of these picks, based on having the most bars with advantage above zero out of this group.

Now we begin to head on in to the late game:

These two are a bit flat in the beginning, but have definite trends up as games go past 40 or so minutes. Lissandra top is waining in popularity, and perhaps rightly so given these sort-of middling results. The current meta leans in the direction of tanks, and so we are seeing more Maokai being picked. This plot suggests that it may be a good idea not to give up on him too early.

And finally, some real late-game powerhouses:

Fizz and Jax I would expect to see here, but Kennen is a bit of a surprise. In his current state, he does need a few items to really get off the ground. That may start to explain his poor early performance. I like Jax's plot; ok guys, the game has gone to 50 minutes, I am Jax, we can win now.

Why not end the day on Teemo?

This plot is hilarious. Those first three bars are based on about five thousand games. You can just feel the frustration of the teammates. "Blasted Teemo top again, I give up." That feeling when. Ha!

Phew, we made it. All the way to the bottom. Thank you for being here with me today, I genuinely appreciate it.

These power over time analyses are going well, and I am having fun. There are more popular picks for top-lane than there are for mid or adc, so there was a lot to look at today. I bet dollars to donuts that I missed something here, so let me know what you think on the ol' twitterino.

See you next week, I will probably be doing a similar analysis for jungle picks. Until then, head back home and catch up on some old posts. Then rss it up so you can get real-time notification for future posts.

Until then, you know where to come for League insights and math insights. That's right, LeagueMath.com.

No loyal subjects of the monkey king were harmed in the making of this article.

Peace.