Official FOFC Hattrick Rankings for week 19.03

[FrogMan]

Another Hattrick week, another rankings... Sorry for posting kind of late, but HornedFrog and I were stuck in a "Frog only" meeting where we worked on some webpages for the upcoming FOFC cup :)

PLEASE NOTE: from now on, the rankings and request to post your game ratings will all happen in here, the new Hattrick board.

I'll get to them in a moment, for now, some explanations...
- TW stands for "This Week"
- LW stand for "Last Week"
- The first set of numbers (to the left of the team names) represents the relative ranking of every team. A blank in either column means that team wasn't ranked that week.
- The second set of numbers (to the right of the FOFC member name) represents the rankings points scored for the last two week.
- The rankings points for a given week are calculated using a weighted moving average of that week's NSI points and the score of the previous two weeks' NSI points.
- "NSI points" is the sum of all your position ratings from your game report, with midfield counting as triple its value
- Ratings are scored with the base value set at disastrous = 1, wretched = 2, and so on.
- If your team doesn't play its game because of a walkover, your rating for the game in question will be ignored in the weighted average and the remaining ratings will be used.


Below, you will find a text formatted version of the rankings, but they are available on the web, GO TO THE OFFICIAL FOFC HATTRICK RANKINGS HOME ON THE WEB (http://webhome.idirect.com/~stevegougeon/FOFCHR/)

Also, if you want to see who has done particularly well (or bad) this week, there's a "tops" page, on which you can also see every team's game ratings for week 19.03. Someday, there will be a link straight from the rankings page, but for now, it available only from here:
WEEK 19.03 IN A NUTSHELL (http://webhome.idirect.com/~stevegougeon/FOFCHR/tops.htm)

I won't make comments on the ratings anymore, since you can all see who's good and who isn't (yeah, yeah, the Pittbulls have moved down again)

That's about it for me, enjoy the rankings. As always, your comments are welcome and bragging is encouraged :D If you have any question about my very scientific methodology, feel free to ask :)

FrogMan


RANK POINTS
TW LW Team FOFC member TW LW
1 1 Airborne Penguins TargetPractice6 89.8 82.5
2 2 Screamin' Seamen dataking 67.5 68.7
2 3 Gargoyles Ravenhawk 67.5 66.0
4 5 Slumlords winking pirate 60.7 59.3
5 6 MoKan Mighty Bastards Meanstrosity 60.2 58.8
6 4 SouthPark Cows lytic 59.8 59.7
7 7 North County Raiders Ardent Enthusiast 58.2 55.3
8 10 Hamas Mad Bombers astralhaze 56.2 52.7
9 8 Ultra Nox Qwikshot 53.7 55.2
10 15 Morrisville Morrisvillians NevStar 52.7 50.0
11 14 Galesburg Avengers Coffee Warlord 52.0 50.8
11 20 Cowtown FC Mrushh 52.0 47.8
13 13 Cedar Pond Horned Toads HornedFrog Purple 51.0 51.5
14 16 Calibiri Daedalus 50.8 49.7
15 11 Aardvarks Aylmar 50.0 52.0
16 25 Louisiana Cyclones Doug5984 49.5 46.5
17 18 Ruskies The_herd 49.2 48.5
17 20 Chesapeake Cheapshots Quiksand 49.2 47.8
19 24 Arkham Avengers Nylarlahotep 48.7 46.7
20 Les Joujoux Ally 48.5
21 33 Negative Waves moriarty 48.2 44.0
22 9 Houston Hippopotami sterlingice 48.0 52.8
22 12 Long Island Pride NYFAN 48.0 51.7
24 22 Alf Trickers Alf 47.5 47.5
25 32 Slithy Toves Brillig 46.8 44.3
25 33 Isla Vista Vandals Vince 46.8 44.0
27 23 Blade Bombers Blade6119 46.3 47.3
27 39 Seagrove United auwillie 46.3 43.7
29 17 NE Oklahoma Sidekicks vexroid 46.0 48.7
29 28 Bees ausonny 46.0 45.5
31 33 Bull Dogs Eilim 45.8 44.0
32 52 Flying Condors condors 45.7 41.7
33 19 Midwest Ice Storm FC Pilotman 44.8 48.2
33 42 Toledo Drunkards Rich1033 44.8 43.0
33 44 G-Vegas FC Easy Mac 44.8 42.8
36 40 Tomahawks -Panther 44.7 43.2
37 33 Maniacal Misfitz Havok 44.5 44.0
37 42 Fantastic Flying Froggies fantastic flying froggies 44.5 43.0
37 47 Oklahoma City Rednecks Airhog 44.5 42.7
40 52 Lancre Wowhawks Katon 44.3 41.7
41 31 Eugene Emeralds JHandley 44.2 44.5
42 33 Bamas superbama 44.0 44.0
42 33 Reign HeavyReign 44.0 44.0
44 29 Complex Spikes Masked 43.8 44.8
44 30 Fairfax Admirals awm9a 43.8 44.7
44 40 Mississippi Strikers FC Ranger93C 43.8 43.2
44 50 Boston Wyckyd Sceptre RonnieDobbs 43.8 41.8
48 52 Witches johnnyveigns 43.5 41.7
49 59 AnalBumCovers AnalBumCover 43.3 41.2
49 63 Fighting Crawfish The Afoci 43.3 40.8
51 44 White Sox Pride Aesyrqwe 43.2 42.8
51 59 West Henrietta FC bigjdotcom01 43.2 41.2
51 72 Manhattan Marauders NAIWF 43.2 39.0
54 26 S. Boston Hooligans McSweeny 43.0 46.0
54 44 Spring Grove Fighting Eagles illifan999 43.0 42.8
56 70 West Fargo Stangs leverb66 42.8 39.4
57 Boston Bulldogs patman 42.7
58 57 goldurnears robbgmaier 42.3 41.5
58 58 Gasmasters Fonzie 42.3 41.3
58 65 White Bear FC Scholes 42.3 40.4
61 56 Golden Eagles FC GoldenEagle 42.2 41.6
62 49 McGees AccardoOutfit29 41.8 42.2
63 48 Scranton Slime TredWel 41.5 42.3
64 69 Else United rexallsc 41.2 39.5
65 62 booger ballers fflix 41.0 41.0
66 66 Racoons Barkeep49 40.8 40.3
66 67 Bbors bbor 40.8 40.0
68 50 High Tops Professor58 40.7 41.8
68 52 Goldfish MikeVic 40.7 41.7
68 77 Baltimore Blast Raven 40.7 37.7
71 74 Shadow Dragons Astott 40.5 38.2
72 87 Square Pegs mckerney 40.2 35.2
73 27 Aviators damnMikeBrown 39.5 45.8
73 59 Rochester Roosters dacman 39.5 41.2
73 63 Concord Assassins BishopMVP 39.5 40.8
73 68 FC Hartford samifan24 39.5 39.7
77 77 Newton Rams GrantDawg 38.8 37.7
77 79 Greenlight Superior FC Richards 38.8 37.0
79 81 Plowboys Calis 38.5 36.0
80 75 Boulder Blaze Bertogarce 38.0 38.0
81 Yukon Quest albionmoonlight 37.8
81 Port Angeles Colossal Squids WSUCougar 37.8
83 73 Shaw_Shanksters Superman=#54 37.7 38.5
84 70 Falmouth United Kevin 36.8 39.4
84 82 Pintendre Pittbulls FrogMan 36.8 35.7
86 80 Little Green Monsters SplitPersonality1 36.7 36.5
87 88 Vileurban FC jfbbis 36.5 33.8
88 76 Michigan Moon Shine DeToxRoxDVHStyle 35.3 37.8
88 83 North Plainfield Plague cthomer5000 35.3 35.5
90 Missouri Tigers Achilles 35.0
91 84 First Coast Conquistadors MylesKnight 34.7 35.4
92 Frankfort United DukeRulesMAB 34.0
93 85 IFK Bananskruv 3ric 33.3 35.3
94 85 Manchester Silver Bullets Sublime 33.2 35.3
95 Fairfax Fizz Bee 32.2
96 89 Averill Park Mercury RPI-Fan 30.7 30.8
97 94 ThunderThighs Neuqua 29.2 23.2
98 90 Rio Grandes Radii 28.7 29.5
99 91 Milford Money Shorty3281 25.7 26.4
100 92 Los Alamitos Fury heybrad 25.5 25.5
101 Clumps Marmel 24.2
102 93 Mountain's Manglers Mountain 22.3 24.0
103 Lexington Utd. ColtCrazy 22.0

[TargetPractice6]

The ratings in a nutshell thing just leads to the ratings page...

[Doug5984]

wow- 16....I am very happy with 16....but I know there are some teams lower than me who do have better teams...oh well I get to ateast enjoy a week in the top 20 before I slip back down to around # 25

[DukeRulesMAB]

Good lord, there are teams that are going worse than mine? That can't be good. :)

[NAIWF]

Biggest mover this week, yay! I expect to drop next week, but it was nice to almost make the top 50 in my second week.

[FrogMan]

Originally posted by TargetPractice6
The ratings in a nutshell thing just leads to the ratings page...

yep, saw that. Link has since been corrected.

FrogMan

[Vince]

Woo-hoo :) Looks like I'm a top 25 FOFC Hattricker. And after this next week, my squad looks to possibly encroach upon the top 20...

[FrogMan]

Originally posted by NAIWF
Biggest mover this week, yay! I expect to drop next week, but it was nice to almost make the top 50 in my second week.

There's always worse. Poor ColtCrazy got his team the exact same day of the league game. Played his first league game without having the chance to change much on his roster...

FrogMan

[Doug5984]

A question regarding promotion- I am in a fairly week division, and have already beaten the team that I thought would give me the most trouble 5-0....There is a chance that I could get promoted to Div IV after this season- I managed to put up a 52 NSI, with an inadequate midfield, would this be decent enough to ateast compete in an average div IV. series? What would be the lowest recommended NSI rating a team should have before going to div 4....I know its very early in the season to be thinking about this....but just incase....

[AccardoOutfit29]

LOL, I am 62th. I know I have a much better team than that though.

[McSweeny]

damn... huge drop for me even though i won 4-1 this week

[mrushh]

It is AWESOME to be tied #11 for Cowtown, but I know when everyone is out of the cup and starts using their A-team and stops PIC in league, I will drop like a pass in a Dallas Cowboy receivers' hands. For now, it looks great. :)

[TredWel]

Urggh, down 15 to #65? This is what happens when you take a long-term view of things and let the B-team take the league game.

I'm aiming for a 50 rating next week, so look for the Slime to move up in the rankings then.

[Coffee Warlord]

Airbone Penguins 92

Galesburg Avengers 52

Okay, who the hell let TP in again? :)

[Neuqua]

Now granted I know I suck, but after winning 4-0 (scoring my first 4 goals of the season), how do I still move down 3 places?

Or is this one of those rankings where the higher the number you are, the better. Yeah, it's one of those isn't it.

I Rule!

Neuqua

[PilotMan]

I just needed to mention that this weeks fall to 33 has everything to do with training, and a need to put the best team on the field for my cup game this week. We will break the top 20 again once we get kicked out.

MISFC

[Raven]

dude, my name is Raven, not Ravenhawk! :)

I am the Baltimore Blast

[vex]

My guys will be in the top 12 3 weeks after we're out of the cup. Now to see if we can get back to #9 or better.

[vex]

A thought:

While it would take some time to look it up, how about having what place everyone is in their respective series?

[Raven Hawk]

Ummm, Frogman, we have that problem with me looking like I own the Baltimore Blast again. . .

Just thought I would mention it.

;)

[FrogMan]

To Raven and Raven Hawk:

Sorry guys, I was sure I had corrected it the last time. It has been edited in the ratings on this thread and will be corrected on the webpage tonight.

FrogMan

[Bee]

95 with a bullet baby!

[daedalus]

Woohoo! The Calibiri reversed our wonderfully spiralling ratings for the first time.

[lytic]

VIQ (Very imporatant question)

Next week, I will have disastrous across the board... Do you just want my next week cup results?

RavenHawk (the evil one) will also have disastrous across the board in the near future as well... So what will the ruling be?

[vex]

"- If your team doesn't play its game because of a walkover, your rating for the game in question will be ignored in the weighted average and the remaining ratings will be used."

[FrogMan]

Originally posted by lytic
VIQ (Very imporatant question)

Next week, I will have disastrous across the board... Do you just want my next week cup results?

RavenHawk (the evil one) will also have disastrous across the board in the near future as well... So what will the ruling be?

Nope, you'll have to live with your league ratings for 2-3 weeks. I did just that last week. Posted a 30, compared to my regular 40-42, it dropped my 39 spots while I posted a 49 in my cup game. Once I get back to normal, I should find my spot back in the rankingsm, at least the spot I know I belong to :)

If you mean that you'll have disastrous accross the board because of a walkover, as vexroid seems to imply, then that rating will be ignored and your rankings score (the average) will be ( (week 19.03)*2 + (week 19.02) ) / 3. I will simply drop the part (week 19.04)*3 from the equation... Why would you get a walkover? Any dead team in your series??

FrogMan

[dacman]

OK, let my first preface this by saying I appreciate all the work you have put into the rankings, FrogMan. However, I have a bone to pick with you, I think.

So, if I understand correctly, the formula is something like: (week 3)*3 + (week 2)*2 + week 1) / 6? That's not really a "moving average" as I understand it. 2 bad games in a row and your rating will drop significantly. A less volitile way of doing it would be something like: (previous rating)*6 + (week 3)*3 + (week 2)*2 + (week 1) / 12.

Let's use an example:

Week 18.14 - 49
Week 18.15 - 47
Week 18.16 - 49
Week 19.01 - 46
Week 19.02 - 38
Week 19.03 - 37

So maybe this team used its B team for 2 weeks in league games -- using the current formula as I understand it we get a rating of 38.83. Does this team really deserve a 38.83? I'd have to say no. Using what I proposed we get a rating of 41.98. You could weigh the previous rating even more to make it less volitile or vice versa. Either way, my understanding of a "moving average" is that the previous average bears the largest weight in computing the next one. Actually on second thought, I think what I'm describing would be more accurately called a "weighted average." Either way, the way it appears now, the previous rating really bears no weight at all outside your previous 3 weeks scores. I'm not really sure that's the best way of doing it.

Secondly, league games only in the score? Why should we not take the cup rating if its the better one? As it is now, the rankings are all over the place because many of us are using our B teams in league games and A in the cup. My team has dropped some 30+ spots in the last 2 weeks because my B team posted 38/37 in the last 2 league games, however my A team posted 47/49 in my last 2 cup games. Is my ranking right now reflective of how good my team really is? I'd have to say emphatically, no.

I realize changing the way it works may cause a lot more work for you, so if that's just not possible, I'll certainly be happy with what we have now -- its much better than nothing. However, I just think it could be improved.

[DataKing]

Originally posted by Doug5984
A question regarding promotion- I am in a fairly week division, and have already beaten the team that I thought would give me the most trouble 5-0....There is a chance that I could get promoted to Div IV after this season- I managed to put up a 52 NSI, with an inadequate midfield, would this be decent enough to ateast compete in an average div IV. series? What would be the lowest recommended NSI rating a team should have before going to div 4....I know its very early in the season to be thinking about this....but just incase....

I took a look on HatStats at the #32 rated Div IV. series from this past week. Series IV.5 had an average rating of 54.6 (with an average of 26.2 stars). So, if you consider this to be the average Div. IV series, I'd say you should be able to compete.

Just pray you don't land in IV.49. :D

[DataKing]

Dola...

FrogMan: Raven Hawk and lytic do have a dead team in their series. The Center City Gooners are an automatic walkover this season for everyone in their series (IV.19).

[FrogMan]

Originally posted by dacman
OK, let my first preface this by saying I appreciate all the work you have put into the rankings, FrogMan. However, I have a bone to pick with you, I think.

So, if I understand correctly, the formula is something like: (week 3)*3 + (week 2)*2 + week 1) / 6? That's not really a "moving average" as I understand it. 2 bad games in a row and your rating will drop significantly. A less volitile way of doing it would be something like: (previous rating)*6 + (week 3)*3 + (week 2)*2 + (week 1) / 12.


I appreciate the tone of your comment, dacman, and we'll pick the bone together ;)

I agree this is not a typical moving average and I have always called it a "weighted" moving average, in a sense that more weight is placed on the more recent games. That's my point of view that the more recent games should be a better representation of a team's performance than a game that happened, say, five or six weeks ago, and thus should be weighted more in the average. You have the right to disagree, though.

The reason I decided to go with a 3 weeks average was because I thought that, given the length of the HT schedule, it was a reasonable number of games to gauge a team's performance. I could have gone to 6, 8 games, or could have gone the same way HATStats is doing it, i.e. only the last game. I thought that only using the last game was unfair to teams who sustained an injury and were out of a league game. That 3 weeks average was my way of making it less volatile...


Let's use an example:

Week 18.14 - 49
Week 18.15 - 47
Week 18.16 - 49
Week 19.01 - 46
Week 19.02 - 38
Week 19.03 - 37

So maybe this team used its B team for 2 weeks in league games -- using the current formula as I understand it we get a rating of 38.83. Does this team really deserve a 38.83? I'd have to say no. Using what I proposed we get a rating of 41.98. You could weigh the previous rating even more to make it less volitile or vice versa. Either way, my understanding of a "moving average" is that the previous average bears the largest weight in computing the next one. Actually on second thought, I think what I'm describing would be more accurately called a "weighted average." Either way, the way it appears now, the previous rating really bears no weight at all outside your previous 3 weeks scores. I'm not really sure that's the best way of doing it.


This is where we differ the most. A regular moving average would simply average the last "x" scores. My way of doing it, I put more emphasis on the newest score, instead of putting more importance on something that has happened many weeks in the past. As I said, I decided to draw the line at 3 weeks because I thought it was reasonable. That's why a game played before the last three games has no bearing on this week's rankings.

You see it one way, I see it another way. If many other FOFC brethren think differently than I do, we could change the formula, otherwise, I don't know what to do...

Now comes the problem with teams not fielding their "A" team on Sunday...


Secondly, league games only in the score? Why should we not take the cup rating if its the better one? As it is now, the rankings are all over the place because many of us are using our B teams in league games and A in the cup. My team has dropped some 30+ spots in the last 2 weeks because my B team posted 38/37 in the last 2 league games, however my A team posted 47/49 in my last 2 cup games. Is my ranking right now reflective of how good my team really is? I'd have to say emphatically, no.


Now, I gotta admit, I had not foreseen this when I started computing the ratings... I also wanted to concentrate my effort to one day of the week (Sunday night/Monday).

That being said, we could go on a voluntary basis as to who wants his cup match ratings to count instead of his league match. Yeah I see that this could be done. But as I said, I would have to be notified (probably through the "post your ratings" thread) that a team doesn't want it's league ratings to be used and would want to use the next Wednesday cup match. I see the week as starting with the league match, with the cup/friendly being the second match of the with. Reason I want to go with voluntary is that most of the guys who care enough are already posting their ratings in the thread and I don't want to start hunting down ratings in the cup (or tracking who's in or out of the cup) unnecessarily on Wednesday/Thursday.

Also this option would either delay the production of the ratings or ask me to present some updated ratings midweek...


I realize changing the way it works may cause a lot more work for you, so if that's just not possible, I'll certainly be happy with what we have now -- its much better than nothing. However, I just think it could be improved.

Playing with formulas wouldn't be that big a deal, just a matter of thinking them out and finding a way of doing them in excel.

As I said, I did the formula that way because I thought it was fair. I know I have never consulted with anybody around here before posting the rankings. If anybody else think the formula is unfair or could be done in a differen way that would be better, let yourself be heard. By better, you need to understand that retrieving stars, and/or team position within a league is a nightmare when done manually. There are 103 teams on the list right now...

I agree I never thought about the cup vs league game problem, nor the walkover problem. My way of dealing with the walkover was a simplistic way for me to deal with it and it shouldn't.

Finally I gotta say I didn't think people would take the rankings that seriously. I'd like to see my team in the top 30 (realistic for me, heh) but I know I'm picing my games right now and I knew that by playing my "B" team last week, it would be about 2-3 weeks before I have a shot at that goal, but I'll live with it.

FrogMan

[Airhog]

I wonder if it would be possible to write a program that would retrieve the ratings for you automatically?

[FrogMan]

Originally posted by Airhog
I wonder if it would be possible to write a program that would retrieve the ratings for you automatically?

It's certainly possible, but it's not permitted by the Hattrick developpers. That interdiction is in the rules. HATStats had to get a special permission to do just that.

FrogMan

[Airhog]

yeah, but they did get permission to do it. I think it would be great for groups of people like us that have alot of memebers, and would like to maintain ranking, but not as a system wide program...

[DataKing]

If we weren't able to pull the data directly from Hattrick, perhaps we could get it second-hand from HatStats? Just an idea.

[Airhog]

well I know its possible, since HAM pulls basically the same information....

[HornedFrog Purple]

Maybe we could come up with a compromise (if FrogMan wanted to do it) to include both League and Cup games for the 1st 3 weeks of the season. That seems to be the average length everyone here lasted. Those that didn't, get their friendly in for those weeks.

I think hatstats is stars based though isn't it?

[dacman]

Frogman -- I don't think we differ much. By including the last rating as the largest weighted portion of the formula, you just "dampen" the changes a little. The weights for each weeks ratings would be (using the same formula I originally stated):

Most recent game 3/12 = 25%
1 week ago 7/24 = ~29%
2 weeks ago 11/48 = ~23%
3 weeks ago 11/96 = ~11.5%
All other previous games total = 11/96 = ~11.5%

So 88.5% of the rating is still the last 4 games. Clearly the above would need tweaked a bit to give more weight to the most recent game and maybe less to the game 1 week ago, but the jist of the idea is there.

[dacman]

Originally posted by HornedFrog Purple
I think hatstats is stars based though isn't it?

Hatstats has both stars and the rating we refer to as the NSI.

[Airhog]

well, its never a bad thing to have several different ranking systems, kind of like college football :D

[dacman]

I've mirrored a team page (Bbors) here so you can see what I'm talking about or maybe even see hatstats for the first time! - Hatstats team page for Bbors (http://mywebpages.comcast.net/fofchattrick/Bbors.htm)

The "rating" column is exactly what we use and call the NSI.

bbor - if you have a problem with this I'll take it down immediately.

[FrogMan]

okay many replies here:

Airhog I do know it's possible, but I've also read in the conferences all the talk about the HattrickPoli guy not being able to get the permission to pull game ratings automatically. Only HAM has that ability for now, as far as I know.

DataKing Maybe we could get the data that HATStats pulls from the system, I don't know, but probably. I have no knowledge of SQL or that type of stuff and am not ready to get immersed in something like that. I started doing the rankings out of fun and because I was able to do it with a software that I like to play with (excel)...

HornedFrog That's basically what I was offering, but to do it for people who expressly would know that their league game wil give them lower ratings than their cup game. I'm not sure I want to retroactively redo the ratings for the last three weeks. If we are still doing the rankings that way next season, this would probably be the way to go, take the best of the league game and the cup game in a given week...

FrogMan

[FrogMan]

Originally posted by dacman
I've mirrored a team page (Bbors) here so you can see what I'm talking about or maybe even see hatstats for the first time! - Hatstats team page for Bbors (http://mywebpages.comcast.net/fofchattrick/Bbors.htm)

The "rating" column is exactly what we use and call the NSI.

bbor - if you have a problem with this I'll take it down immediately.

I assume this is in response to HFP, right. I know that HATStats has both stars and NSI ratings, but they use only your last game...

FrogMan

[dacman]

Did you look at the page, FrogMan? Bbors ratings go back to week 1 of season 18! 17 league games worth.

[FrogMan]

Originally posted by dacman
Frogman -- I don't think we differ much. By including the last rating as the largest weighted portion of the formula, you just "dampen" the changes a little. The weights for each weeks ratings would be (using the same formula I originally stated):

Most recent game 3/12 = 25%
1 week ago 7/24 = ~29%
2 weeks ago 11/48 = ~23%
3 weeks ago 11/96 = ~11.5%
All other previous games total = 11/96 = ~11.5%

So 88.5% of the rating is still the last 4 games. Clearly the above would need tweaked a bit to give more weight to the most recent game and maybe less to the game 1 week ago, but the jist of the idea is there.

I see what your saying. What do you mean by "All other previous games total"?? Do you mean an average of all the games that team has ever played? I think that's just too much stats to keep around. If 4 weeks is better, then 4 weeks could do it, but I simply don't think we should carry around all the games a team has played in the past and show that as being representative of how this team is currently performing...

If we always take the best between cup and league, this would mean that every team would always have its "A" team scored in the rankings and then should always be going upward. You can see it in the rankings, most of the team have an increase for a week to the other. The other way around (a decrease) should not be seen as a bad week (other than because you played your B team in league, but we've taken care of that with the max between cup and league) and should not be compensated by the fact that your team has played well in the past... In my opinion, that would be wrong...

FrogMan

[FrogMan]

Originally posted by dacman
Did you look at the page, FrogMan? Bbors ratings go back to week 1 of season 18! 17 league games worth.

I know that, I've seen that, but I don't see how his rating for week 1 of season 18 will tell how good his team is performin RIGHT NOW...

FrogMan

[FrogMan]

dola, and when I said that HATStats was using only the last week, I was referring to their list of best ratings, which could be seen as similar to our rankings... I wasn't referring to the fact that Hatstats didn't have many weeks of data.

FrogMan

[HornedFrog Purple]

Actually what I was suggesting was take both cup and league games from weeks 1 and 2 of the season and the cup game right before the season.

Since the majority of people here last to round 3 of the country cup you would not have to differentiate which game to use just use them both. If a team is out of the cup, then use the friendly that week instead.

After week 3, just revert back to the league game. That way you would not have to mess with special circumstances since a majority of the teams are out anyways.

[dacman]

Originally posted by FrogMan
I see what your saying. What do you mean by "All other previous games total"?? Do you mean an average of all the games that team has ever played?

Yes, but once you already have last week's rating computed that's the ONLY number you need to go back that far -- its just works out that way in the formula. You don't need to keep numbers any farther back than the extra weighted weeks.

Originally posted by FrogMan
...should not be compensated by the fact that your team has played well in the past... In my opinion, that would be wrong...

So are the Lakers a horrible team because Minnesota beat them by nearly 30 last night? 1 game does not make the team, but it kinda seems that way with the way the rankings are now.

[dacman]

Originally posted by FrogMan
dola, and when I said that HATStats was using only the last week, I was referring to their list of best ratings, which could be seen as similar to our rankings... I wasn't referring to the fact that Hatstats didn't have many weeks of data.

FrogMan

Gotcha, but you made it sound like the ONLY rating they kept was last weeks.

Now that I have most teams linked in the team list, looking up the best rating for a particular week for a particular team is now much more feasable.

edit: spelling

[FrogMan]

Originally posted by dacman
Yes, but once you already have last week's rating computed that's the ONLY number you need to go back that far -- its just works out that way in the formula. You don't need to keep numbers any farther back than the extra weighted weeks.



So are the Lakers a horrible team because Minnesota beat them by nearly 30 last night? 1 game does not make the team, but it kinda seems that way with the way the rankings are now.

I'll have to think about that first statement a little, I don't get it just now.

About the Lakers comment, it is not that way. The thirty point loss would be the last game used in the rankings. I agree that one game does not make the team, that's basically why I decided to go for three game. I just thought, and still do, that the latest game is the best representation of how the team is currently performing which is what I thought that the rankings should be showing, how a team is currently performing...

Taking another pro sport example, in some NFL power rankings, it's not uncommon that teams will move quite a few spot from one week to the next. Hattrick is more like that, in a sense that we only play once a week, not like basketball (or hockey for that matter), where teams' performance can vary more from a game to another but would be better represented on a weekly basis...

FrogMan

[DataKing]

Originally posted by HornedFrog Purple
Actually what I was suggesting was take both cup and league games from weeks 1 and 2 of the season and the cup game right before the season.

Since the majority of people here last to round 3 of the country cup you would not have to differentiate which game to use just use them both. If a team is out of the cup, then use the friendly that week instead.

After week 3, just revert back to the league game. That way you would not have to mess with special circumstances since a majority of the teams are out anyways.

Using this sort of system would give an (undeserved) boost to those teams who do not maintain two full squads. If you only have 16 players on your roster, then odds are your ratings will be better than if you played 22 players, since that "b-team" will likely have lower ratings than those players on the 16-player club that pulled double-duty.

Also, ratings from the first two weeks of cup matches aren't worth much, IMHO, as they were for the most part against Div VI bot teams. I'm sure that precious few of the team owners around here bothered to play their starters against bots.

Personally, I think that the ratings are fine the way that they are, although I am sure you all think I feel that way because I am #2 right behind TP (and he doesn't really count :p ). I look at the ratings as a week-to-week guage of performance, not an overall indication of team strength per se. The last thing I want is some sort of BCS nonsense here in Hattrick. The FOFC Cup itself is a much better measure of team strength, I think.

And let's all please bear in mind that this is something FrogMan does for fun. He could alter the ratings to declare himself the #1 team "just because he feels like it" if he wanted to, and we would all then have the right to pay attention to or disregard the ratings as we saw fit.

OK, rant over. I just don't want to see people take this too seriously, I guess. :)

[Nyarlahotep]

Guess I should add my two cents:

1. I am happy with the current ranking formula.

2. I am in favor of teams being able to decide which ratings get used for the first three weeks of the season. The main reason for this is that the rankings from the third week are used to decide FOFC Cup seeding. I would also like to note that this will not take any longer than the current rankings do. Since the national cups start before the season, people that wanted to use their cup ratings would just have to post their info by the normal Monday deadline.

[dacman]

Just to clarify -- I'm not unhappy with the ratings. I just think the seeding for FOFC Cup III could've been better represented by a tweak in the way rankings are computed. Counting the best game of either cup or league might be the only change we really need to make if at all.

[Nyarlahotep]

Originally posted by dacman
Just to clarify -- I'm not unhappy with the ratings. I just think the seeding for FOFC Cup III could've been better represented by a tweak in the way rankings are computed. Counting the best game of either cup or league might be the only change we really need to make if at all.


Agreed, my seeding took a hit, but I don't want anyone to do any extra work. If I was that worried about seeding I would have used my starers in both matches.

[TargetPractice6]

Man, that hurts DK :p I'm not inviting you to my birthday party!

[DataKing]

Originally posted by TargetPractice6
Man, that hurts DK :p I'm not inviting you to my birthday party!

:D

[lytic]

Originally posted by vexroid
"-Lytic is blind when on percocet."

Thanks Vex... I guess I'm blind.:rolleyes: