Season Seven LAN Finals Preview

StarLadder started today, showing us some great action so far from 4 of the top EU Dota 2 teams. Please check out our LR thread or join us on IRC to discuss the games this week, I promise it will be fun. For this article we got one of our resident stats experts, K-poptosis, to explain the impact that playing offline or online events has on the results.

The games are already upon us, so sit back and watch some great EU Dota!





We have all heard it innumerable times before, "LANs are just different". According to many, players play faster and better without ping issues making skill shots easier to land, teams enjoy being in the same room, and the best teams always bring their A-games. Well, put a black beret on my head and horn-rimmed glasses on my face because I am now channeling the MythBusters to whether some points of this argument hold any water. Apologies in advance for the lack of explosions.

When the International 3’s qualifier teams were announced, many believed iCCup was chosen to compete over other teams, including Kaipi (now Speed Gaming), due to the fact that iCCup had successful LAN experience. By upsetting Na'Vi en-route to taking 2nd place at TechLabs March LAN, they had proven they could both successfully travel to a LAN, and play good Dota while offline as well. Whether or not that is the true reason they were selected , the popular opinion being what it was is very important, as it showcased just how much the community believes LANS are truly a different beast compared to playing online.

But is that truly the case? Sure, the fact that pings are reduced is not debatable, but do teams actually perform differently when playing at a LAN? One route I considered to analyze this was simply comparing teams records when on and off LAN, but there are multiple problems for doing so. The largest of the reasons being relatively small sample sizes and polarized win rates due to the higher concentrations of talented teams. For example, a team might have a bad record at a LAN simply because they weren't better than the competition, not because they perform poorly on LAN. Therefore, what I decided to do was dig a bit deeper to see if I could find a root cause for this perception and not just the outward display. Specifically, I wanted to see if one of the biggest reasons why LANs are placed on a pedestal - negligible ping - actually affects the accuracy of skill shots.

I believed that looking at abilities with long cast or travel times would give the largest disparity between on and offline play due to their heightened dependence upon timing and foresight. Consequently, I decided upon three abilities from regularly drafted heroes over the last year that showcase this reliance on having reading the other team quickly and with precision. They were Jakiro’s Ice Path, Clockwerk’s Hookshot, and Leshrac’s Split Earth. (Data prior to Jakiro’s Ice Path formation time buff was not included.)



As you can see, more skill shots do seem to be hit when on LAN, but the difference is negligible. According to this data Jakiro benefits most of the three from being played in low ping conditions, but even then only 1 in 62 Ice Paths that would miss online would hit on LAN, and that difference is not even statistically significant. This may come as a shock to many that would believe it to be much easier to hit skill-shots on LAN, but the possible explanation is a fairly obvious one: your enemy benefits from the decreased latency as much as you do, making juking easier. Furthermore, the difference in playing at 150ms ping versus 20ms may very well be too small for the human body to capitalize upon. This supposition is supported by data from the website HumanBenchmark, which through 6.5 million tests using it’s own reaction time test, determined that the average (median) reaction time is 215 milliseconds, with only 0.1% of testers reaching a reaction time higher than 130 ms.

So there you have it. According to the data, the ping differences for online versus offline competitions is simply too small in most cases to cause a significant change in gameplay that favors particular heroes or playstyles. That’s not to say that larger ping differences that are commonplace in EU vs US matchups are negligible though. Quite the opposite actually if the ping difference exceeds the normal human reaction time. However, performing on a LAN should not cause an increase in performance beyond a player’s normal faculties, it simply evens the playing field. If a team performs better on LAN than in online tournaments, I would consider it more likely due to outside-of-the-game factors such as team communication when in person, and ability to deal with jet lag and hotel food.



Okay, maybe I lied about no explosions.

---