I don't know, Terry. It seems like no matter how poorly I perform I stay in Div. 4 or 5; I'm in 5 right now. If a winner of a game cranks off a 19 second round and I get a 42 it seems like I should drop like a stone. I haven't played recently because things seem to be weighted against me.

The idea of stratifying people into divisions is good, but the implementation could use some improvement. As CmdrK points out, there's a big difference between someone who regularly scores 10/19 and someone who regularly scores 10/42, yet the current algorithm seems to regularly put these people in the same division. Also, some sets are simply easier than others.

Instead of looking at relative position (17th out of 42, or 1st out of 42) or raw scores, I think you ought to look at performance relative to the median time for that set. A 10/25 might be a winning time for one set, but sub-par for another. Also, some players are capable of achieving times that are significantly better (at least one standard deviation) than the others. You might want to create extra divisions so that everyone in a division is within a standard deviation or two of each other. It's no fun to play a game where one has no chance of winning.

Maybe not for you, but not everybody is so focused on winning, many people play for fun and don't particularly care about their score.

You're absolutely right, samak, I should rephrase as: "Many people don't have fun ...". People who just play for fun will be happy under any setup, including no divisions at all. The whole purpose of divisional play is to cater to the people who do care about winning, by giving people more opportunities to do so.

"who regularly scores 10/19 and someone who regularly scores 10/42, yet the current algorithm seems to regularly put these people in the same division"

If those results place these two players in the top 20% of players, then yes, both of these players will be in Division 5 together.

If you could give me an example of 2 players who regularly play who are in the same division and really shouldn't be, I can go check the data to make sure there isn't a bug. I suspect, though, that when you're talking about Divison 5, there will naturally be a few players who are better than the others. We could stick them in a tiny little division by themselves, but I'm not sure that makes sense either.

This is why I'm mildly irritated at being catapulted straight into Division 5, where I'm instantly up against the top speed demons, against whom I stand zero chance of winning and advancing any further.

I assumed I'd advance slowly through the Divisions, allowing me to win occasionally and get a few advancement points, until I'd naturally reached the top. If the algorithm never drops me down, it's a bit pointless us "slower" players playing at all.

Stedman, it will drop you down. I got rocketed to 5 after a single game, but then moved down to 4 after a few more games, and now seem to move between 3 and 4, which feels about right to me. I don't actually look at my position, since I only try to play it at all if it is part of a Daily Challenge, or I see a category in which I think I might manage under 30 seconds, and possibly contribute to my team's score that day. In either case, it doesn't matter to me how other players went!

It feels about right to me as well. I tend to do well at Easy, less well at Intermediate and badly at Difficult and I normally stay at the same level. I assume this is what should be happening.

It is very fair. The top performing 20% of players end up in Division 5, and the lowest 20% of players ends up in Division 1. After 1 game play the system doesn't have enough info, so don't fret. Just play a few times and you will end up where you belong.

The reason we hear complaints from Division 5 players is that there can be a lot of difference in players in the top 20%. Like in any bell-curve, the top 3% will be significantly faster than those around the 18-20th percentile.

This seems to me to be a fact with no obvious solution. I could create a new division to throw the outliers into (say, the top 3%). But I'm not sure that having a division where 1-5 people play each other each hour is particularly interesting or fair to them either. It feels a little like "you're too fast, we don't really want you playing this game competitively". Who knows though, maybe these people wouldn't mind?

I seem to have jinxed myself by mentioning that I had a 42 second round. I'm capable of better than that (or used to be) and I've been laying back a little to see if my division would change. I played the 11:30 game last night. With 11 minutes to go, eyhung was in first with 19 seconds and I was last (around 28th) with a 42 (in my defense I did miss one question). I just looked and I'm now in division 4. That has happened before but after a game or two I find myself back in division 5. I'll try now to see what happens.

Hmm how about splitting division 5 into 2 smaller divisions.

Division 5 will be 2/3rds of current division 5.
Division 6 will be the top 1/3rd of current division 5, and will give 6 win points.

Divisions 1-4 will be larger, but easier.
Division 5 will be smaller, but against more difficult competition.
Division 6 will be smaller still, but against the very toughest.

Seems to be a better experience for everyone?

Sounds good to me Terry. I'm always in Division 5, and have no complaints, but I have little chance of beating the speed merchants in most games. I rarely get under 20 seconds.

Perhaps the solution is to return it to the way it was previously with levels and Champion divisions....it didn't seem broken to me.

Nope. Performance based divisions are much fairer, particularly for infrequent players. Also please keep in mind that you are in the top 1% of players in the easy game at the very least, so none of the problems surrounding levels and divisions effected you.

Giving the Division 5 split idea shot in both Easier and Fifty-Fifty.

Thanks Terry. I see that you've added Most Win Points & Highest Rated Players to the game pages which adds a nice little bit of spice to the game again. Thank you!

The splitting of 5 into 5 & 6 looks good too. It seems the previous "lower 5s" now have a chance at least of winning.

For a POC badge, could it be something like a score of 960 in 30 consecutive games played over say 72-96 hours with some sort of stipulation that X number of categories must be covered in the same bracket?

Also are we really going to measure down to two decimal places? Those close to the server geograpically are going to have a distinct advantage at that level. The speed of the network used is also going to have some bearing at this level of timing.

I would prefer to retain the status quo with zero significant figures and tie breaks to be settles by playing time within the hour. First played for equal scores, wions.
Thanks

I much prefer the measurement down to two decimal places and feel it makes for a better experience instead of giving the people who have the opportunity to play early in the hour a disadvantage. I wouldn't mind more precision in other games as well such as the Easy Fifty-Fifty, where it's almost all about speed.

I do appreciate your issue of playing from a distant country. I wonder if, instead of reducing precision, it would be better to have the server deduct .1 s from the submission of a client from a distant overseas country, such as Australia or China.

I think you're trying to pare this so thin there is only one side to it. In that case, people in my area who use CenturyLink internet should ask for a handicap since it's known to be slower than Cox internet, which the rest of us in the valley use.

No, we will not be doing anything like that.

I see each division in both the 50/50 and the POC now has its own 'nickname'. Very nice

I'm not sure what to think about the Division 5 nickname in Fifty-Fifty.

What is it?
Division 4 is "The Half Brains". Are players in Division 4 being called half-wits?

Division 5 is the Half Heroes.