Jungling: Farm Forever

The gold value of the jungle increased in yesterday's patch, matching the roughly 10-20% increase I initially expected with the advent of Season 3. As such, it's time to revisit an old theory of mine.

To summarize, a jungler has the potential to acquire more farm than any lane provided they continuously clear the jungle and occasionally cover a lane.

Before I dig into the math, I'd like to highlight the three biggest obstacles for this theory, in order of importance:

1. Smite on a carry is generally awkward.
2. It's much easier to predict and track a roaming mid than a jungler.
3. Machete represents a sunk cost of at least 180 gold.

Although #3 is a fairly simple hurdle and #2 can be dealt with, it's fairly obvious that #1 is a big problem. The only champion in the game for whom Smite is particularly useful outside of the jungle is Akali, and such a singular exception is not a robust solution to the issue. With the new jungle being as difficult as it is, jungling without Smite is not an option on any champions matching our needs.

While there are some practical problems facing the theory, mathematically speaking the new jungle is an insanely safe place to get a carry farmed. If we assume a jungler clears Wolves -> Blue -> Wraiths -> Wolves -> Red -> Golems, finishes at 4:10, and then begins a perpetual clearing cycle of the small camps, they will have collected 1550 gold by 7:40 game time (without accounting for scaling). By comparison, a lane who last-hits perfect will have farmed 1996 gold. The jungler, in this case, has only 77.66% of the lane's income.

However, all it takes for a jungler to match a perfect lane is to leech one creep wave every two minutes between cycles. Alternatively, a lane which misses 23% of their last-hits will match the jungler's income. Given the potential for lanes to build up creep waves, and the fallibility of players, it's suddenly easy to see how the jungler could be competitive.

For an example of farm in a real game, consider TSM versus M5 at IEM Kiev last year. 21 minutes into the first game the highest creep score of any champion in the game was 174. In an absolutely perfect lane that never took any jungle, the maximum possible creep score at that point in time would be 253 (40 creep waves with 13 cannon minion waves). In short, even the best players in the world are often only able to acquire 68.8% of their lane's farm (and that's assuming they never took any jungle by that point). Given that a jungler is absolutely assured of their farm barring incredible effort on the part of their opponents, there's obviously potential in this idea.

Some testing needs to be done to see exactly how much a jungler can farm in an actual game versus theory. There are myriads of extenuating factors which could kill this fledgling notion. With some help I aim to find out.