Battle Reports

How's it going, Rob? Are the Player Ratings Ready?

published on

So the player ratings are set to be done by the end of Q3.  Yeah, that's eight days from now.  I'm reasonably certain I can get it done - mostly.  Chances are, the medals (like the level medal) won't be ready in time and that's okay.  The player ratings are generated at the time of the battle and are independent of the medals so I don't see a problem with applying the update before the medals are ready.

Medals?  Did you say Medals?

I did, actually.  When your player rating is generated you'll receive a medal that's proudly displayed on your player profile indicating which of 12+1 "ranks" you hold.  There's 12 ranks for rated players (any one who's reported more than 30 games) and a provisional rank.  They look something like this:

Medals for battle rating
Medals awarded for battle rating on iToysoldiers

I'm reasonably happy with them.

What's taking so long?

Well, as it turns out figuring in the impact of a player's faction required a lot more testing and tweaking than I originally thought.  My original thought was to modify the K value of the calculation through the use of a ratio between the player's faction battle ranking.  This didn't pan out exactly the way I wanted it to.

So I set myself to figuring out exactly where to apply the faction impact and I stumbled across a website that talked about how they determine EL0 for NBA teams.  One of the factors they introduced into their calculation is home team advantage.  Clearly I don't need that but HOW they did it was clever.  They essentially modified the rating of a team by a constant when they had a home team advantage.  "Ah ha!", says I.  That's slick.

So right now I'm messing around with the calculations with an aim to implementing a ratio that's applied to the player's rating before the rest of the calculation.  So consider a commander playing Eldar with a rating of 2000 against an Ork player with a rating of 1000, the calculation will actually set the Ork player's rating considerably lower and the Eldar player considerably higher.  Not that in this example it'd be necessary.   But if they both had a rating of 1200 then if the Eldar player won the change in scores wouldn't be nearly as dramatic as if the Ork player wins.

Please tell me you're not messing with anything else

Well...  Nothing major.  A couple of bugs.  The biggest of which is making it so that when you add an image from the modelling project page it'll actually get associated with the modelling project.  I've also done some work on the page layout for events so there's not quite so much empty space on the page.  Really, that's pretty much it.

Comments? Questions?  Post a comment here or visit the iToysoldiers Help Portal.

Rob @ iToysoldiers