A Proposal for Gameplay-based Statistical Analysis of the Impacts of Player Behaviour

Hey guys, this is hopefully gonna catch the attention (and hopefully interest) of at least a few of you - and since I have nothing better to do I'm hoping to produce results that will be useful to the playerbase and maybe even to Riot (if I end up following through and finishing up this will probably end up on reddit, if allowed).

So, we've all seen the load screen messages. x% more games are won by non-toxic, than by toxic players. But how truly significant is that figure? This is what I will (hopefully) be able to find out.

My proposed methodology (which I have already begun carrying out) is this:

One condition will see players presented with a friendly teammate, who will use a scripted set of small jokes (puns of damage) to break ice in champion select, a scripted opener in all chat, and throughout the game, scripted encouragement towards teammates for doing well, or to offset poor in-game performance.

The other (which I haven't tried yet, as will become obvious) is a 'toxic' encounter. Teammates will be presented with a teammate who will be rather difficult to work with. I haven't started testing this mostly because it's difficult to devise a scripted 'toxic' dialogue, and also because I haven't really felt like being a total asshat to my teammates recently. The script will have to involve:

• Something mildly offputting to teammates in champion select
• A slightly 'salty' opening line
• Something to say to teammates that is offputting when doing well (an arrogant statement or something about the jungler not being present enough on my lane perhaps?)
• Something to say when teammates do poorly (an indirect statement such as "omg" is likely, for reasons discussed below)

At the end of the 'toxic' condition, players will be informed of the aim of this little pseudo-study and any further questions will be answered. The main problem with this condition is that I either need to find things that are off-putting to players but still adhere to the Summoners Code or seek additional guidance/permission from Riot (one of the reasons I'm here, actually).

All players in the nice condition will be 'debriefed' so to speak at the end also, and hopefully they don't have too much of a problem there!

The reason for using scripted chat messages is mostly to reduce variation with a condition, because that basically reduces the validity of any results.

Since the matchmaking is random within skill levels and I'm a low-skill player, it's highly likely that players encountered will be different, so the samples are unrelated. As a result, either an un-paired t-test or Mann-Whitney U test will be used to analyze the results, and the key findings will be released (probably on reddit).

So yeah, if anyone is interested or has any questions (or if any kind, lovely Rioters want to give some guidance on that elusive 'toxic' condition), let me know. Even better, if you can offer any corrections or improvements, feel free to give me some pointers (mostly because my stats work is awful because I'm a terrible student).

Thanks, for reading and hope you guys have a good one!

Also have a poll because why not.