Page 1 of 1

Why did I not get the same rating result as other players?

PostPosted: 16 Feb 2013, 21:21
by super_dipsy
As some of you may know, I don't like repeating myself :) . However, there ARE times when it is both necessary and a good thing. The new rating system is a case in point - it is a huge change for everyone, and therefore it is perfectly reasonable that people may take time to get to grips with the details.

This is an Addenda to the new rating system description/FAQs to cover questions such as
- Why did I not get the same as that player got for the same result?
- What is a solo worth? I think it is 400 points.
- What is a 2-way, 3-way etc draw worth? Isn't a 4-way worth 70?

While the following points have been covered in the previous guidance, I think it is worth gathering the relevant points together to see how they affect that key factor - what you get for your result. Hopefully this will help to increase understanding, and I make no apologies to those who complain I am just saying the same things again.

So, without further ado, here are the points that affect the score you get from a game.

1. When you play a game, your 'score' (the change to your rating) will depend on the ratings of all the other players in the game as well as your own. I think this one is reasonably well understood. If you play in a game with players lower than you in ratings terms, you will score less for a win or loss than if the other players are higher. You can think of it like an averaging process - if there are more higher than lower then you will still do better than if the reverse were true,

2. The scores used are those of the players in the game when the game finishes, not when it starts. So you may think - "but we were all 1000 ratings, so why didn't I get 400 for my solo like I have seen others get?". But in reality other players may have been in other games and some may have finished before yours and therefore the score will NOT be based on 7 1000 players.

3. If you join a surrendered position, your score will be rated as a % base don the percentage of turns you played that country. Although of course if your rating shield is active you are protected from getting a negative score. So if you solo but you only played half the turns for that country, you will only get half what you would have got if you had played all the game.

4. The rating system tries to find your 'average' playing level as quickly as possible. In order to achieve this, it is designed to swing more wildly (depending on your result) when it has little confidence in your rating, and less so when it has more confidence. The reason for this is that since everyone starts on 1000, it has no idea whether you are a top player destined for ratings of 2000+ or an average player or even a newbie. So it swings you about quite dramatically to try to get you were you belong. As you play more games, its confidence in its assessment of you grows, and it dampens down the swings to reflect that. What this means, of course, is that a result against a set of ratings that you achieve in your 100th game would not be the same as the same result against the same ratings in your first game, because the system has more confidence in your assessed playing level in the first case.

5. 400 is not the score for a solo. All players here in February 2013 were there at the big bang. You saw the stars starting to form. You saw everyone on 1000, when the system had no knowledge or confidence in anyone's ratings because noone had ever played any games. In this scenario, yes a solo was worth 400 if all the players in the game were still at 1000 when the game finished (see 2). This will never be true again. You may still get 400, or even more, or you may get less, or much less. The same is true for the draws scores. As a general guideline, a solo against 6 other players of the same rating as you as your very first game would score you 400 points, but once you have played enough games for the system to be very confident about you (50+) then the same result would get you 48. The way to think of it is that the 'base score' for soloing against players of the same level as you is 48 (you are beating 6 others of your standard), but that early on the system wants to hurry you into the right ballpark in terms of scores so you will be grouped with others of similar ability.

6. When you are in a game where more than one player has played a country (someone surrendered and someone else joined) the rating of the country for the purposes of calculating your result is an 'average' of the ratings of the players involved and the number of turns each of them played. So the player who started the game as Italy might have had a rating of 1000, but if he surrendered half way through that country's life and the new player had a rating of 800, when the game finishes Italy will be taken as being rated at 900.

7. Your rating result when you win a game of Ancient Med will not change as much as for the normal Diplomacy. This is because you are only playing against 4 players as opposed to 6, and therefore you have a better chance of winning in a set of random games (well, at least that is one way of looking at it :) ). So a win in Ancient Med against 4 other players of the same rating as you will be less than as if the game had been normal Diplomacy with 6 other 1000 rated players. An interesting point here for the future perhaps - the factor here is the number of players. So if we ever had Crowded for example (11 players) you would score a lot MORE for a win (beating 10 other players).

8. A minor point but worth mentioning for completeness, in the case of someone surrendering then that person is scored up as a loss at that point in time (whatever everyone's ratings were then). Also, when the final results are calculated at the game finish, the rating of the surrendered player will be taken as it was at the point the player surrendered before any adjustments are made. So an example to clarify: A player of rating 1000 surrenders, and receives a negative delta of -65, making a new rating of 935. This result is calculated based on everyone's ratings AT THAT POINT IN TIME. A new player of rating 1200 joins and plays the rest of the game, which turns out to be 50% of the moves for that country. At the end of the game, that country is assessed as being rated at 50%*1000 + 50%*1200 = 1100, NOT 50%*935 for the surrendered player.

9. Finally, a solo win is weighted more heavily than all other results. But this weighting is % based, so there will be no distortion when calculating a solo that results in a small gain compared to one calculated for a big one.

Since this is purely informational, I am going to lock this to avoid it getting long and confusing to people. But hopefully this helps to answer the question of why I did not get the same score as those players there got for THEIR results :D