TWD Player Ratings Recap

[qan]

Before we begin a new season of TWD using the new player ratings system and ladder, I wanted to break down the ratings system so it's a little easier to understand the details. Ogron, please feel free to correct me if there's anything in error here. (At this point it should all be implemented according to plan, as we discussed over the last year, with the exception of Ladder being multiplied by 10 to represent a slightly finer degree of accuracy. Let me know if there are any other surprises/discrepancies. The individual ratings themselves are of course all done according to your thorough and excellent blueprint.)


All leagues

• You'll receive a rating at the end of the round, as usual. This rating has been carefully crafted to better represent overall skill than the original TWD ratings formulas. Average rating is 100.
• For each round played, the round rating is averaged into your current overall rating.
• Your Ladder rating is different than this overall rating. Ladder is simply the overall rating multiplied by Confidence, times 10. Average Ladder rating, when at 100% Confidence, is 1000.
• Confidence is the number of rounds you've played, compared to the average of the top 25 most active players in the league. If the average number of rounds played by the top 25 most active is 50, and you've played 50 or more rounds, confidence is 100%. So if you have a rating of 120, and 100% confidence (1.00), your rating will be 120 * 1.00 * 10, or 1200. If you've played only 25 rounds, meaning you're at 50% confidence (you've played half as many rounds as the 25 most active players) your ladder rating would be 120 * 0.50 * 10, or 600. Confidence keeps players from simply getting a great rating and sitting on it, as they need to continue to play rounds while continuing to perform well to maintain their position.
• If you're struggling to understand confidence, you can think of it as a guarantee of overall rating. The rating system's confidence in the ladder rating it gives you is extremely strong when you've played many more rounds compared to other players, and weak when you've played fewer. The more rounds played, the less chance there is of a statistical anomaly.
• Ladder and cumulative rating are available on the overall Statistics page (which is sorted by Ladder rating) that can be seen on the far right nav bar of the TWD site, (coming soon) on an individual player's Statistics page, and on the Top 5 players per league (appearing on the TWD front page at the start of the new season). Note that before the new season starts, the ladder ratings will not yet be accurate -- in particular BD.

League specifics

DD/JD (and also SD/FD, if anyone ever plays those)

• Non-basing leagues will use Star Rating. At the beginning of the season, each player's Star Rating will be loaded from TWDTD/TWDTJ. Where a star rating can't be found, it will be set to 6.
• Minimum Star Rating is 3, and max is 10.
• After each round played, Star Rating is adjusted slightly, based on Rating for the round.
• How is Star Rating used? It adjusts all player's round ratings. The star ratings of all players on a team are combined, and this forms a general idea of how well the team will do. Basically, the difference between the two teams' star ratings are taken into account in player ratings. If one team has a combined 27 Star Rating and the other has a combined 20 Star Rating, the expectation is that the team with 27 will win most of the time. Effectively, Star Rating is sort of a stand-in rating for when a season starts, to account for the already-well-known skill of many players. If a highly-star-rated team performs poorly against a low-rated one, their personal ratings will be skewed by this factor, and vice versa. This prevents vets from making big ladder gains by stomping newbies.

BD

• BD doesn't use Star Ratings, due to the fact that there are many ships people can play, and ship-switching is common.
• Rating for BD is determined on a per-ship basis. How well you play in Spider doesn't affect how you're rated in Terr.
• Your Ladder rating for BD overall is just your best Ladder rating of any ship.

If you have any questions I'll do my best to answer them.

[Ricko]

How is star rating determined for new players/players aliasing when it comes to all this stuff? A 9* rated jav could alias to gain more points from a worse team?

[qan]

How is star rating determined for new players/players aliasing when it comes to all this stuff? A 9* rated jav could alias to gain more points from a worse team?

Unfortunately, that's true. The silver lining is that it would only work for a short time. Their star rating would rise over time as they continued to play well, and eventually it would even out. ATM star rating just works to represent perceived rating. If they're on an alias, technically there is no perception of their rating, so it's fairly accurate. (Unknowns rate as 6 star to start.)

If they feel like they need to alias just to ladder well then they probably aren't a 9 star, though, because let's be honest, most people at that level are all about the personal glory, which includes their name. Everyone will figure out pretty quick that they're aliasing just to game the ladder. Not exactly something that inspires respect.

[qan]

Here's a quick and dirty illustrated guide to confidence for people who don't quite get how ladder rating works.

The tl;dr of the tl;dr is: keep playing TWD or your personal ladder rating will fall.

Click to view.



EDIT: Janky forum system killed the attachment. I'll have to upload from the original PNG later. Here's a JPG with painful text for your viewing pleasure.

[Jones]

Nice work!

I don't care if people alias just to get better (star) ratings. It just proves they still enjoy this game which is great!

[Rough]

LOL zidane is going break the confidence meter to 200% and his star rating...he never logs off and loves this game

Also, good work to the staff that implemented this...I am sure it wasn't easy at all.