Skill Score Exposed

---

Ask Bob - What Do You Know About Skill Score

From time to time I'm asked Starhawk questions. Today's question came from a couple of my twitter followers. "What do you know about skill score?"

Well, LBI is very tight lipped about skill score. This is because they do not want people to pad their stats to get a better rank. That said, by playing there's a few things you can figure out. This is based on my own experience, LBI can change the calculation at any time, and your mileage may vary.

Winning Team Matters

Being on the winning team matters, a lot. You can take on the hardest server, be the top player on the losing side and your score will go down. I don't know how switching teams affects this; however, when you join the losing team does affect your score. If you start a game, leave and your team loses, your score goes down.

That said, if you join very late and your team loses you do not lose as many points. Like I said, when you join matters, but apparently when you leave does not.

Top/Bottom Player

Being the top player or bottom player can enhance or reduce the skill score change. If you're the top player on the winning team, expect a big bonus. If you're the bottom player on the losing team, well that sucks to be you. That is of course scaled based on when you joined. Top player on the losing team doesn't see their score go down as much, nor does the bottom player on the winning team see their score go up much.

The server difficulty does seem to factor into this change as well; however, winning in a cake server seems to always produce a net gain, where as losing in a hard core server will cost you.

Starting Score

By default everyone starts out at 3000 skill score points. Over time if you don't play your score will degrade till you hit 3000 as well. While I'd like to say the number of points you win come from a player on the losing side, I haven't confirmed that yet.

Basically, if you're under 3000 points you need to find easier servers.

Ranks

There are several ranks people can obtain. They are percentage based so that means you can change ranks one of two ways:

1. Get people to join
2. Improve your score

Here's how things break down: (For checking purposes, estimate approximately 100,000 players).

• In the top 2% Blue Splat
• In the top 11% Star
• In the top 21% Diamond
• In the top 51% Circle
• in the top 76% Square

Note, if your skill score is simply 3000, that puts you in the top 51%. That means if you see someone who doesn't have a circle or higher, give them a break they've had some rough rounds.

Stats Explained

---

The Rifle and Other Weapons

As noted the rifle appears to have way more kills than pickups. This is due to the rifle being a weapon when spawning. In addition the player gets a set of grenades. And indeed the hawk mech comes with swarm missles.


While I can normailze the graph (approximately) for these spawned weapons, I cannot do the same for the hawk weapons. The resulting graph is:



So there's a couple of things to note. First, as I said above the swarm stat is biased and should not be trusted. Second, the rifle drops in effectivness when compared to the previous graphs. This is because you spawn with it, so I've counted every kill as a pickup. The same with the grenade which is now in the sub 1% range.


What I don't have is how effective each weapon has been in total. For example, I often use up my grenades on structures instead of people. Also combination attacks are not taken into account (for example attacking with a rocket launcher to finish the player off with a rifle). As a result these stats should be taken with a nice helping of salt.

The source for these numbers comes from the Lighbox Interactive StarHawk Beta News page

More StarHawk Stats

---

New Numbers

Last night Lightbox Interactive updated their kill and pickup numbers. They now include everything from 11/15/11 to 11/21/11. I've updated my chart to show these new stats:



The first thing that jumps out, swarm missles have bested rifles. While it is not clear if troops or other hawks are being destroyed, it is clear that people are having a blast.


The new numbers show a total of 3029 kills using just the weapons above and 17514 pickups. That translates into a pickup to kill average of approximately 17.2% which is down from the last set of stats.


Mines jump out a clear winner as well as the shotgun, where as homing missles, grenades, rocket launchers and sniper rifles have a below average pickup to kill rate. As Warhawk players flood the private beta today it will be interesting to see how these numbers change.

Sweet Starhawk Stats

---

StarHawk Stats

In a sheer stroke of dumb luck, I stumbled across some of the first data from StarHawk's private beta. As a result I've compiled this graph for you.



Now I know you're asking what does it all mean. Good question. First let's start with the source data (which I'm not linking to, it should be out to everyone soon.) This is based on play data from November 15th to November 18th of 2011. In that time there were 5286 weapon pickups (excluding repair tool pickups) and 950 deaths (excluding 4x4, knife, etc). That means for every pickup the players had an ~18% chance of killing another player with it.


From there it is broken down further, every time a rifle was picked up there was a staggering 45% chance of a kill. Compare that to homing missiles at about 5% and you can see that StarHawk has a thriving ground game. Absent from the data was any pistol kills, which may mean the player doesn't get one, or it is so weak that no one uses it. Also the grenade numbers may be skewed as the player may spawn with a couple of grenades like in Warhawk (similarly with the swarm missiles).


Just based on this data, grab your rifle, shotgun or mines early, they are well worth it. The rocket launcher just doesn't kill as often, which may indicate that breaking locks is easy in StarHawk or they simply are hard to handle against your fellow troops.

There appears to be a few other secrets in the stats, I'll keep mining away.