Proposed Changes to the 2015 UK Rankings

[OldManMyke]

Hi all

 

So I think that the rankings this year have been pretty successful and as we run into the last few events of the year the race for masters qualification is still very tight with a lot of new names looking to make it.

 

As the strength of UK events grows and we get more and more people playing the game (I think every event I enter has a couple of new folks) I think a couple of small tweaks to the rankings for 2015 are sensible and frankly needed.  So in an attempt to not be a unilateral dictator (a new approach for me) I thought I would stick down a proposal and see what the community makes of it, after all they are your rankings at the end of the day, I just facilitate them.

 

So proposals

 

1. Master qualifying ranked events will be Singles Events only.  We can have separate rankings for things like Teams and Doubles (or event Hardcore Henchman) but only singles events count for Masters
2. Minimum number of players for a ranking event will be 12 (up from the current 8)  The number of players for a 100 point event will remain at 28
3. Score calculation will change slightly so that the bottom player in an event will always score 5 (0 seems harsh)  This eliminates the problem of someone coming bottom at a smaller event scoring more rankings points than the person coming bottom at a larger (and probably tougher) event

The final discussion point is about large events.  Something like Vappa 2015 has currently 56 players coming but due to the fact that it is only 3 games the decision has been made to split it into effectively 2 simultaneous events.  I propose that each of these pools is counted as if it was a 28 person event rather than having say a 28 and a 26 due to drop outs.

 

Anyway thats what I'm thinking.  Let me know your views please and also if there are things that you would like to see on the rankings website that it doesn't do currently.

 

 

 

[OldManMyke]

So an example of how the scoring would work

 

In a 28 person event the winner gets 100, bottom gets 5.  The differential applied is 95/27 or 3.51 giving the following spread

 

1. 100
2. 96.48148
3. 92.96296
4. 89.44444
5. 85.92593
6. 82.40741
7. 78.88889
8. 75.37037
9. 71.85185
10. 68.33333
11. 64.81481
12. 61.2963
13. 57.77778
14. 54.25926
15. 50.74074
16. 47.22222
17. 43.7037
18. 40.18519
19. 36.66667
20. 33.14815
21. 29.62963
22. 26.11111
23. 22.59259
24. 19.07407
25. 15.55556
26. 12.03704
27. 8.518519
28. 5

[Cunning]

Sounds good to me Mike.

I'm not sure if this is already in place but I don't think story encounters should be ranked either.

[ProximoCoal]

This all sounds good to me

[Oshova]

Ok, I'm going to be honest here. I really dislike this form of rankings. It is a ranking system based off of quantity not quality.

If I go to larger events with low quality players, and do awesome. I would rank higher than someone doing equally well who has fought their way through the best of Britain. (in this case, as we are talking UK rankings)

To me, that doesn't seem right. A rankings system based off of quality would favour players who beat quality players. It doesn't matter at what tournament it is, or how big the tournament is. If a No0b comes in and stomps the number 1 with a 10-0, then that should be reflected on their ranking.

I would propose an ELO based system. (no not the band).

With an ELO system, your ranking score is compared to your opponents, and the winner and loser have a change of score based off of the difference in their rankings. So if the number 1 beats the number 100, then it will mean a small change for both players. But if it's vice versa, it's a big change for both.

Now, the common fear with ELO is that is generally linked to more competitive games (An example being chess). And Malifaux players generally try to steer clear of situations that may result in overly competitive play. But we have seen over the past years with rankings that isn't the case. We still have fun, and the rankings are an extension of that fun.

Another fear is over complication with statistics and obtaining data. But again, we already hand on the scores for the tournament to Mike for rankings. All that would be required would be a list of who played who each round. That is entered into a simple spreadsheet or database. Then the computer can do all the complicated stuff for you, and spit out the rankings at the end.

I honestly feel that this would create a more representative view of how players rank against each other than the current system.

[floso]

I like changes except,min 12 players to rank will suck for any 8 to 11 player tournaments.

[MariFaux]

I agree with floso, having moved now to a more remote location (tournament-wise), it seems harsh to people living far away from the midlands if 12 is the minimum number. That's what the relative scoring is there for.

 

Also, are there large events with "low quality players"? and what's wrong with quantity over quality? I hope we're all in for the game and the reason for going to as many tournaments as possible is (at least for me) that I like to play as often as possible with nice people, not for the one weekend I would play in the masters ('cause there will be an awesome side-event anyways) or a number in a list.

[Barnaberible]

Oshova

 

I don't disagree with your system, it has many merits but the existing geographical spread of ranked players may have an effect on this

 

At the minute the largest number of high ranked players are based in the North of England (excluding the squigs who are nice and central)

 

Now these guys, quite understandably, don't get down to many (if any) Southern tournaments so surely this system would have a detrimental (and increasing) effect on the Southern tournaments that don't, through no fault of their own, attract the top ranked players.

 

The North would get stronger and stronger as their players are continually playing high ranked players in tournaments and the South would get weaker as they cant attract high ranked players.

 

Re minimum numbers

 

Everyone loves the rankings (right?) and it would seem a shame to exclude tournaments completely. Would there be a way of making smaller tournaments still ranked but making them worth less ranking points that they are now?

[Oshova]

@Marifaux

I wasn't saying there was tournaments like that, it was an example of something that could happen, not something that does happen.

 

With quality being the basis, it means you have to be on top of your game at every tournament, and every game. (Or at least as much as possible). Which creates a more representative view of rankings. And as a lover of stats, it is something I am interested in seeing...

 

 @Barnaberible

 

Actually, with the way ELO works, you would get a spread within each microsystem of rankings. The North would end up with high and low rankings, and the South would end up with high an low rankings (using your example). This in itself can be a problem. However, every time that someone visits both microsystems it actually begins to equalise the 2 against each other. So to further your example. If The Squigs went to both sets of tournaments across the year, going back and forth, then the rankings system would essentially use The Squigs as a basis to rank the 2 systems against each other. Obviously, this requires people mixing between different systems to equalise it. But with the current system you end up with the same problem, but with no medium to equalise the two against each other.

[Joel]

How many events this year have failed to get to 12 players? I can think of a couple off the top of my head - and one of those did have more than 12 before dropouts on day 2...

...I'd like to see the number of players that start an event be the number of players that the event is scored from, with DNFs getting 0 ranking points, but their dropping out not harming the results of others.

[Joel]

@Marifaux

I wasn't saying there was tournaments like that, it was an example of something that could happen, not something that does happen.

 

With quality being the basis, it means you have to be on top of your game at every tournament, and every game. (Or at least as much as possible). Which creates a more representative view of rankings. And as a lover of stats, it is something I am interested in seeing...

 

 @Barnaberible

 

Actually, with the way ELO works, you would get a spread within each microsystem of rankings. The North would end up with high and low rankings, and the South would end up with high an low rankings (using your example). This in itself can be a problem. However, every time that someone visits both microsystems it actually begins to equalise the 2 against each other. So to further your example. If The Squigs went to both sets of tournaments across the year, going back and forth, then the rankings system would essentially use The Squigs as a basis to rank the 2 systems against each other. Obviously, this requires people mixing between different systems to equalise it. But with the current system you end up with the same problem, but with no medium to equalise the two against each other.

Ah, the SQUIGs at the heart of Malifaux.....

[MariFaux]

In my opinion the system works as it is for small tournaments. Winning an eight player tournament only scores you 80 points and it goes rapidly down after that. If we want to devalue them more than that, we might as well kick 'em out completely.

[OldManMyke]

As I mentioned earlier on Twitter I had actually considered an ELO system but didn't want to introduce it for 2015 as it is a radical change.  What I propose is that we run it alongside the existing system for 2015 and see what the results look like.  We can then reconsider it for 2016.

 

With regards 8 vs. 10 vs. 12 I always thought this would be the bone of contention amongst folks as you are all correct, we don't want to exclude events.  Perhaps 10 would be a sensible compromise

[mythicFOX]

Mike - Any chance we can get the exact formulas?  Just cause I’m a nerd and I’d like to play with them.

 

As the game grows the system needs to move with it and so think it’s right the system continues to evolve. The above changes will work and I welcome them.

 

IMHO the biggest issue we face with the current (2014) system is the way the maximum points calculation works in practice. Two day events are far harder to win than one day events but as they tend to attract slightly less players tend to be worth less points.  On that basis I’d like to propose a further change (in addition to Mike’s suggestion) to the calculation used for an event to scoring max points;

 

• Two day events with 5 or more rounds count as having a bonus X players at the bottom of the table (I’d recommend 8) for the purposes of calculation.
• One day events score as they do now, you can still score max points from a one day event.

 

In practice this would mean you need less people at a two day event to score maximum points but also that players below the top spot will score slightly better at a two day event than a one day event. This naturally scales so that players at the bottom of the table get a bigger boost than those at the top, <1pt to the player in second but around 15 points to the player in twentieth.

 

This should encourage more players to attend two day events and better reward players for winning and participating in them, without penalising players who can only make one day events.

 

Thoughts?

[Soul Puppet]

My opinion hasn't really changed since the last time we discussed the Rankings. I still feel that as well as the more players = more rankings points, we should have more rounds in the tournament = more ranking points. I guess this links a bit to what Mr Doxey is talking about.

 

Of course the biggest issue I have is that I should be qualifying for the Masters and I'm not currently. If you could come up with a system that guarantees me a place every year, that would be great, thanks!

[Barnaberible]

 

 

 

Actually, with the way ELO works, you would get a spread within each microsystem of rankings. 

 

OK so maybe I think I dont quite understand it properly.

 

I did google it but got put off by the number of y's, x's and brackets

 

so lets say at tournament a a player ranker 10 beats a player ranked 1 and at tournament b a player ranked 100 beats a player ranked 90 the ranking points they achieve would be the same, or not?

[Barnaberible]

 

If you could come up with a system that guarantees me a place every year, that would be great, thanks!

 

There already is one, its called winning the GT, no problem for a man of your caliber 

[Oshova]

Yeah, it is definitely not a simple equation... 

 

Ok, so it's not your ranking per se that matters. It's your ranking score.

 

Basically, if you win or lose, your score is changed by a set number (let's say 10). However, the difference in your scores is used as a kind of multiplyer.

 

So in a simplistic form (and totally not correct...):

 

Player A has a score of 100

Player B is new, and so has a score of 0

Player C has a score of 30

Player D has a score of 25

 

So Player B beats player A, and they get a large bump to their score. Come on, they just beat a high ranked player (in this example) in their first game. Also, player C beats player D, there is a minimal change, as they are so closely ranked.

 

However, if Player A beats Player B, then they get a negligible change in their score as they just stomped a no0b.

 

Hopefully that makes a bit more sense... =]

[MalifauxMartin]

Maybe I'm missing something, but what is the benefit of raising the qualifying number of players in an event?

[ProximoCoal]

Interesting topic. Thanks everyone for the feedback so far.

 

 First of all, the ELO system (I have no idea what that stands for by the way, I am just working off what people have told me)

 

 I am not sure I like it. It makes the rankings far more flexible, which I can see why people might think that's a good thing but for me it undermines what I think is one of the coolest part of the current system; that is being able to set yourself little goals. For example, at the moment I haven't set myself getting into the Masters as a goal because it's far too dependent on other players, but getting the big hat has an obvious number to strive for. To try and put it another way it's like having a race based on how far ahead of other players you are rather than when you reach the finish line, it makes it much harder for the runner!

 

I agree with Doxey's suggestion, that has been something in mind for a bit now. I'd suggest that the extra players were based on the proportion of entries into the event rather than a static number (so if you did 20%, for example, a 10 man tournament would get the 12 to qualify, but a 25 man tournament would be pushed up to 30 instead) but otherwise support the motion.

 

One of the things I was worried about with the minimum number of players was the Scottish tournament scene, which is sparser than Britain. I don't know what the best solution would be though. 

 

On Twitter, @derbymalal suggested that there should be some form of benefit based on the number of wins you got etc, which I suppose links to the ELO thing. I think if you do keep it the same, however, that I can't think of any effective change to take that into account. Giving some form of bonus to the number of wins and how close you are to first would not only be complex but also skew things in favour of three round events. which is counter to what I think should be done if anything. I can appreciate the motive behind it, but I don't think it would work well in practice.

 

I hope this is of some use

[theJudge]

a minimum of 12 people is too high.

 

I think 8 works well but if it has to be raised then 10 would be the highest I think it should go.  We here in Northern Ireland can hit 10 pretty easily but getting those 2 extra players to hit 12 would be tough.

[mythicFOX]

 

I agree with Doxey's suggestion, that has been something in mind for a bit now. I'd suggest that the extra players were based on the proportion of entries into the event rather than a static number (so if you did 20%, for example, a 10 man tournament would get the 12 to qualify, but a 25 man tournament would be pushed up to 30 instead) but otherwise support the motion.

 

Thanks

 

I like that, although I'd push it up to 40% right now and look to reduce it if we see the desired uplift in two day attendance. 

 

Developing the idea further I'd uplift events with five or more rounds, rather than two days.  Therefore if you wanted to a marathon five round one day it gets the uplift but a three round two day doesn't.

[ProximoCoal]

You could just do the percentage increase based on the number of rounds. 

 

I.e. 3 rounds is 100% of players, 4 rounds 110%, 5 rounds 120% etc. 

[Oshova]

@ProximoCoal

 

You compared the rankings to a race. With the current system being like the 100m sprint. Well, how about if you look at it like the Decathlon. In that, it is your effort over all of the events that make up your total score.

 

Also, I kind of like the suggestion of different scores depending on the amount of rounds in a tournament. 

 

I actually had some ideas about differences to the score for Elo dependent upon how large the wins were, and actually for the purposes of rankings, determining a winner in draws. So winning 10-0 would give you a larger score, but also drawing with a higher ranked player would count as a win (of sorts...). It increases the complexity a lot... but it also gives a more realistic representation of the scene.

[ProximoCoal]

 @Oshava I can totally see where you are coming from, but I think where the contention lies is what you use the rankings for.

 

 For me the rankings is best used not to prove who are the best players in the country but for banter and setting personal goals. Elo might show more accurately who is top and not, but the difficulty of making goals makes it counter to purpose in my minds. I suppose it depends on how seriously you take it.