Vainglory News 1

A Wishlist and Look at Tracking Competitive Stats

I recently found a spurt of inspiration in a Twitter conversation about the current state of tournaments in Vainglory. While not the only topic of conversation, the area of improvement that caught my interest was stats. Now I’m someone who loves to just open a new spreadsheet in Numbers and test out ideas for various builds, so the idea of being able to look at say the KDAs or creep scores of various teams, in say the LEV, and compare them with other teams is rather exciting to me.

Obviously this game is still in its infancy and there are higher priority things than the development of a stat system and API, but SEMC has also shown that they’re a fan of long-term thinking. I’m hoping that when we do get a stat system, we get a great one.

What We Can Do Now

Aside from what we can do with tools that would need to be provided by SEMC, what can we track right now?

KDA – This probably the easiest stat for us to track right now all you need is a screen shot of the results screen. There are also a number of ways to use this statistic. First, you can analyze the ratio for a match by dividing the (Kills + Assists)  by / Deaths. This way you’ll get an idea of how valuable that player was. A KDA around or below one generally signifies a player who did not or barely managed to carry their weight. Of course there are other factors in play, such as whether or not the other players on his team were supporting them enough or what role they were playing, but generally this is a pretty decent representation of the quality of play.

You can also use this stat to average kills, deaths, assists, and the KDA ratio over all of their matches. This is a particularly interesting stat because you get a look into how a team plays, which players tend to grab the assists and which go for the kill.

Gold – And, to a lesser extent, level. This is generally considered to be the deciding factor in games. The team who earns the most gold and spends it wisely almost always wins. So we can really only get insight into this stat from the end-game screen shot. So what we get is the amount of gold farmed throughout the entire game. We can use this to find a players average gold farmed overall, in a certain role, and as a certain hero. From there we can stack them up against other players in that role and even compare them to the average successful (winning team) farm for that hero/role. It can give players an idea of where they, or another player, fall on the spectrum and insight into whether they need to spend more time farming. 

Build Stats – There are two big things we can do with a player’s build. Firstly, we can look at the end of the match and see how various fights would’ve resulted at that point. Taking a look at DPS and EHP of players and teams to estimate the way a fight would have went. 

The next thing we can look at is how their build compares to the standard build of other players using that hero and the success rate of various builds. 

Other Stats – Since the data comes from screen shots right now, the only other big stat we have shown there is the match length. This data gives us insight into the length of a team’s losses compared to the length of their wins, as well as letting us analyze the length of their matches compared to other teams’.

What We Can Do with Spectator Mode

Spectator mode isn’t too far off now; we’ve gotten a look at its UI and have an idea of what features we’ll be getting with it. This is going to give us a lot more utility in tracking matches, that before, was only trackable on streamed matches. Calculus warning!

Gold Throughout the Match – You’re going to notice throughout this that the big difference with spectator mode is that it gives a look into the match as it happened rather than just giving us the end game stats. So being able to graph each team’s Gold/Minute will show us why the match went the way it went, as well as when it went that way. The coolest part of this for me is that by looking at the first and second derivatives of one of the graphs you’ll have insight into every turnaround in a match, and how it correlated with kills, turrets, and other objectives. Now, I’m getting a little ahead of myself since it’d be rather hard to graph derivatives without an API providing me with the exact data from the match. 

Once again taking this data and using it to compare a teams wins and loses could help in understanding what mistakes they are making in matches. 

Objectives and Kills throughout the Match – If you’ve ever used the stat website built into League of Legends, you may have noticed that it provides you with information on when kills occurred and when objectives were taken overlayed onto a graph of gold throughout the match. I can imagine that this would be popular in Vainglory especially with all of the jungle objectives in Vainglory. Being able to see gold payouts, mine captures, and turret take downs up against the gold farmed would give a strong idea of how a match snowballed, or failed to snowball for that matter. 

Scout Traps – A nifty feature of spectator mode is that it shows us all of the traps on the map. Because traps tend to be somewhat static, they can be tracked in matches. Allowing us to see how many were placed by a team, its individual players, how many were triggered, and perhaps even the average length that scout traps existed for. This can be stacked up against the averages of winning and losing teams to show how they correlate to success and failure. This can also be done with flares. 

Creep Score – While this is roughly the same stat as gold, it can be used to show the effect of various objectives on the amount of gold a team farmed from their creeps. 

What We Could Do with an API

This is the biggest thing with stat tracking; there is so much more we can track once SEMC opens the door with an API. For those who do not know, an API is essentially a tool that developers can use in sites and apps in order to interact with other services, apps, sites, etc. Most of the time when people refer to an API in Vainglory, they’re referring to a tool kit that SEMC would give so that sites could track data about matches.

In my opinion, it would be preferable for them to give us the tools to build the stats than it would be for them to do it themselves.  Not because I don’t trust SEMC to do a good job, but because the users are going to know what stats they want to see and having SEMC to do all of that work for us would be a waste of their time.

Depending on how the API works, it could also result in some less intensive version of spectator mode, similar to stat websites that report the status of matches like sports do.

Time Spent “X” – While this is technically a stat that can be gathered with spectator mode, it would a rather tedious one. It would be much easier to have the game report the status of certain players as say “fighting,” or “farming,” or “dead,” or “taking objectives,” or a combination of statuses. These statuses could be used to create stats like how much time was spent on farming, or how much time was spent fighting enemies.

Fight Breakdowns – Some of this can be tracked with spectator mode, but a lot of this really requires reporting from the game. Information recording all of the fights in a game, their length, deaths, and participants. From this information can be drawn about the results of how a team did in tough scenarios like 2v3s.

Damage Sources – How much did that Bonesaw help? How did fighting by the Gold Miner change the battle? This would be a pretty heavy amount of data, and is perhaps somewhat unrealistic, but I can’t help but dream to see breakdowns of how much each item in a build effected the final outcome of the game.

Objective Stats – How much gold did each team earn from the miner in comparison to the time spent fighting it? What level minions was your team spawning at what times? Helps in analyzing the decisions made by a team.

Damage Dealt – How much damage did each player deal to objectives, players, creeps, structures? This would be useful information in accessing contributions. Also a “Who Killed Who” would be nice.

Random Stats – How many times were Reflex Blocks used, how much damage did they block, CCs blocked, how many times were they wasted? What about boots or Atlas Pauldrons? How many potions were used? How many stuns were delivered? Silences? Slows? There’s so much potential for interesting stats here, like the damage blocked per gold spent of a Reflex Block in a match.

What stats do you guys want to see in competitive matches? How much of a priority do you think that this level of stat availability should be? Do you think some of these stats should never be available?

1 Comment

  • Reply
    AdyEndrus
    Mar 25, 2015 10:43 pm

    I want every stat in competitive matches, and that’s possibly why we’re still so far from them. I was discussing today how quickly a game changes when rank and elo become public knowledge. It was interesting that the consensus was that information about rank could potentially be harmful to the nature of the game.

    This article was great. It tickled my imagination on a lot of pieces of the Vainglory puzzle. I’ve begged SEMC for an API, even offered to build it for them, but it can’t be high on their to-do list. As long as the development team stays actively working on features the community wants then I’m satisfied because it means that these stat dreams that you wrote about will become a reality.

Leave a Reply

Your email address will not be published. Required fields are marked *