MSWL UNITEDMSWL U2 TMBL MSWL The Manager
Wednesday, August 15th, 2018 - 10:16:20 PM (gmt)
 
ball MSWL UNITED ① - Landing
 
Home Auctions Blog Forum History Login Rules Scores Stats Tables Teams
 
Coaches Directory Donate Guest Rankings Schedule Updates Waitlist Wall
 

Join
MSWL
UNITED!

Recent Entries

Allan Sellers
3 Comments
Mike Parnaby
3 Comments
Brian Beerman
6 Comments
Tim Batth
1 Comment
Allan Sellers
1 Comment
Rob Baptiste
7 Comments
Allan Sellers
8 Comments
Brian Beerman
7 Comments
Allan Sellers
3 Comments
Paul Cockayne
3 Comments
Paul Cockayne
3 Comments
Dave Dowson
4 Comments
Roy Rolsten
2 Comments
Dave Dowson
4 Comments
Brian Beerman
1 Comment
Brian Beerman
1 Comment
Dave Dohm
2 Comments
Brian Beerman
3 Comments
Brian Beerman
6 Comments
Allan Sellers
4 Comments
Allan Sellers
5 Comments
Allan Sellers
2 Comments
Allan Sellers
2 Comments
Allan Sellers
6 Comments
Brian Hayes
1 Comment
Brian Beerman
2 Comments
Brian Beerman
1 Comment
Brian Beerman
3 Comments
Allan Sellers
5 Comments
Rob Baptiste
7 Comments
Allan Sellers
12 Comments
Allan Sellers
8 Comments
Allan Sellers
2 Comments
Andy Bate
3 Comments
Rob Peterson
6 Comments
Allan Sellers
4 Comments
Allan Sellers
4 Comments
Allan Sellers
8 Comments
Allan Sellers
1 Comment
Graham Wilkes
4 Comments
Brian Beerman
19 Comments
Brian Beerman
20 Comments
Allan Sellers
7 Comments
Andy Bate
1 Comment
Kevin Martin
1 Comment
Allan Sellers
21 Comments
Allan Sellers
14 Comments
Allan Sellers
12 Comments
Allan Sellers
6 Comments
Allan Sellers
11 Comments
Brian Beerman
9 Comments
Brian Beerman
3 Comments
Graham Wilkes
1 Comment
Jose Freitas
4 Comments
Allan Sellers
5 Comments
Dave Dohm
10 Comments
Brian Beerman
2 Comments
Rob Baptiste
8 Comments
Allan Sellers
5 Comments
Graham Wilkes
6 Comments
Graham Wilkes
5 Comments
Dave Dohm
11 Comments
Allan Sellers
11 Comments
Rob Peterson
5 Comments
Brian Beerman
11 Comments
John Holden
3 Comments
Brian Beerman
13 Comments
Allan Sellers
13 Comments
Kevin Martin
1 Comment
Allan Sellers
5 Comments
Allan Sellers
2 Comments
Kevin Martin
4 Comments
Allan Sellers
19 Comments
Allan Sellers
13 Comments
Allan Sellers
12 Comments
Allan Sellers
7 Comments
Allan Sellers
8 Comments
Rob Baptiste
2 Comments
Allan Sellers
16 Comments
Allan Sellers
7 Comments
Allan Sellers
11 Comments
Allan Sellers
14 Comments
Rob Baptiste
5 Comments
Mark Stretch
5 Comments
Jake Hanny
1 Comment
Andy Bate
6 Comments
Allan Sellers
1 Comment
Allan Sellers
25 Comments
Graham Wilkes
2 Comments
Brian Beerman
6 Comments
Brian Beerman
9 Comments
Allan Sellers
12 Comments
Allan Sellers
10 Comments
Allan Sellers
13 Comments
Allan Sellers
16 Comments
Brian Beerman
7 Comments
David Blair
2 Comments
Brian Beerman
12 Comments
Brian Beerman
5 Comments
David Blair
8 Comments
Allan Sellers
18 Comments
Graham Wilkes
4 Comments
Allan Sellers
3 Comments
Mark Stretch
17 Comments
John Holden
16 Comments
Allan Sellers
5 Comments
Rob Peterson
1 Comment
Brian Beerman
11 Comments
Allan Sellers
11 Comments
Allan Sellers
1 Comment
Allan Sellers
25 Comments
Allan Sellers
30 Comments
Allan Sellers
6 Comments
Allan Sellers
7 Comments
Brian Beerman
9 Comments
Allan Sellers
8 Comments
Allan Sellers
5 Comments
Allan Sellers
10 Comments
Allan Sellers
9 Comments
Allan Sellers
6 Comments
Allan Sellers
9 Comments
Allan Sellers
12 Comments
Allan Sellers
15 Comments
Andy Bate
12 Comments
Allan Sellers
10 Comments
Mike Cabral
4 Comments
Andy Bate
2 Comments
Allan Sellers
1 Comment
Allan Sellers
14 Comments
Kevin Martin
7 Comments
Allan Sellers
26 Comments
Allan Sellers
6 Comments
Brian Beerman
3 Comments
Allan Sellers
11 Comments
Allan Sellers
10 Comments
Allan Sellers
23 Comments
Kevin Martin
6 Comments
Dave Dohm
4 Comments
Allan Sellers
4 Comments
Allan Sellers
4 Comments
Allan Sellers
2 Comments
Allan Sellers
7 Comments
Brian Beerman
4 Comments
Brian Beerman
14 Comments
Brian Beerman
2 Comments
Andy Bate
2 Comments
Andy Bate
2 Comments
Kevin Martin
3 Comments
Dave Dowson
2 Comments
Allan Sellers
14 Comments
Allan Sellers
12 Comments
John Holden
4 Comments
Mike Cabral
9 Comments
Andy Bate
5 Comments
Allan Sellers
6 Comments
Allan Sellers
2 Comments
Allan Sellers
2 Comments
Allan Sellers
23 Comments
Allan Sellers
6 Comments
Simon Compton
5 Comments
Allan Sellers
7 Comments
Allan Sellers
2 Comments
Abe Hamdali
1 Comment
Allan Sellers
4 Comments
Roy Rolsten
6 Comments
Andy Bate
5 Comments
Roy Rolsten
2 Comments
Andy Bate
5 Comments
Allan Sellers
20 Comments
Andy Bate
3 Comments
Allan Sellers
4 Comments
Allan Sellers
3 Comments
Andy Bate
7 Comments
Andy Bate
3 Comments
Andy Bate
2 Comments
Andy Lewis
5 Comments
Allan Sellers
4 Comments
Simon Compton
4 Comments
Kevin Martin
12 Comments
Simon Compton
6 Comments
Allan Sellers
5 Comments
Simon Compton
1 Comment
Simon Compton
1 Comment
Dave Dowson
2 Comments
Kevin Martin
7 Comments
Allan Sellers
11 Comments
Rene Wilkens
5 Comments
Trevor Taylor
3 Comments
Rob Peterson
17 Comments
Allan Sellers
16 Comments
Allan Sellers
9 Comments
Trevor Taylor
7 Comments
Trevor Taylor
2 Comments
Allan Sellers
6 Comments
Allan Sellers
3 Comments
Allan Sellers
6 Comments
Allan Sellers
27 Comments
Allan Sellers
6 Comments
Dan Fitzgerald
9 Comments
Allan Sellers
5 Comments
Allan Sellers
13 Comments
Alon Atie
5 Comments
Allan Sellers
9 Comments
Rob Peterson
10 Comments
Allan Sellers
8 Comments
 
Rankings revisited
Posted by Kevin Martin on Tuesday, May. 25th, 2010 at 5:35 PM

Lost in the reset was the original for this.  Reposting at Al's request as we look at the possibility of automating some form of ranking system for United.

Why rank out teams?  Why not just let the tables and Cup placings do that for you?  Because it's fun, for starters.  And enders.  I like tables.  I like numbers.  I like playing United.  So finding ways to make it all work together and give me some good blog/journal ammo as the seasons roll along is a good thing.  For the rest of you, hopefully these rankings can also allow inter-division comparison.  It can give a manager perspective on how the team is faring compared to every other team, especially in terms of whether or not the manager is gaining ground on the teams ahead in the tables.  Sometimes really good teams have runs of lousy luck and weak teams have things fall just right for them.  Looking at a combination of current strength, resources there for improvement, and total team potential gives a bigger picture about where the team is going to end up after a grueling season of play.  As the dust settles, the Champs are clear.  It's the rest of us that have lots of room for banter about who is going to fare better next season.

With that preface, I tried to come up with a system that would rank every team and allow comparison between them in terms of overall strength.  The idea was in building a predictive tool to take a best guess at future results.  Knowing that there is no way to count on fluke PK's (or deliberately Olmec-induced ones as the case may be for some managers), injury epidemics, a sudden desire to add in four more youth players, a team focusing on Cup wins over league play, unexpected NMRs due to life interfering in play, etc., I don't believe there is any better measure of "who can beat who" than the T11 scores.  Just compare them and the stronger team will be the more likely to win in one straight up match.

In terms of season-long results though, youth potential in development, T16 strength, Backroom Boy assistance, CP reserves, and cash for auctions/trades all come into play.  The temptation is to assign a straight value to each point.  For example, take whatever your T11 score is and multiply it by "X" value and get a rank.  The problem with that is that T11's are not static.  They change every session.  And the gap in scores is deceiving.  As I'm posting this, it is currently pre-session 1 of season 5.  The best T11 in the league is 12.64.  The worst is 5.82.  Is the best team in the league then a tad more than twice as good as the worst team?  Nope. Not even clsoe!  It's more like TWENTY times as good in terms of odds of winning a T11 match head-to-head on a neutral field.  A straight multiplier can't take that into account.  And creating a ranking system that utilizes a geometrical progression from a league-average baseline was more complicated than I cared to come up with at the time.  Maybe for Rankings Version 2.0...

So what I decided on is that some things have a set value regardless of the team: CP (because everyone earns and spends them identically), and Backroom Boys (because they have the same impact regardless of situation or team that has them).  So I settled on set values for those.  On the other quantifiable values, what matters is not how much you have as much as it matters how much MORE of it you have than anyone else.  If you have 1000k in the bank it means very little if everyone else has 1300+.  If your T11 is 9.00, that only means something if you compare it to another team or to the league average to see if "9.00" is a good thing or bad thing.  With that idea as the driving force then, all other values are comparative in nature.  In other words, the person who has the most gets the most points, the person with the least gets the least, and everyone else falls in the middle accordingly for that scoring area.

Finally, I wanted a system that spit out a final number that would be easily compared to everyone else's number and on a readily understood scale.  0-100 seemed like a good scale, so I shot for that and came real close.  I include the non-league teams for comparison sake, as until real recently we've always had a team that was worse than a non-league team.  This will also tell Al if he needs to beef those teams up a bit so whoever draws the non-league in a Cup isn't getting a "bye" match for all practical purposes.  Here's what I settled on:

1) Bank + Player Value.  Cash by itself is deceiving.  So what if you have a lot?  It matters what you do with it.  You are never guaranteed an auction win unless you're bidding more than anyone else has, and blowing your whole wad on one player isn't as helpful as getting 2-3 smart buys for role players, backups, or a critical T11 addition in a key position (which isn't a sure thing to even be available when you need it).  Similarly, Player Value scores are deceiving.  Because of the value formula [(7-age) x (SL-2) x 10k], an age I, SL 10 player with 0 Rsd is valued the same (480k) as an age III, SL 14 player with 3 Rsd already (480k). Which one is going to help you win more matches?  Yep, the old guy.  Good thinking.  Which one has the potential to be better?  The young guy, if CP'd up.  And who is going to help you win more matches from here forward through the next two seasons?  Definitely the young guy if he gets any attention at all.  So I figured the best measure is to lump the Cash and Player Value scores together, and consider that as a measure of current worth + short-term future potential.  The teams with the current top Player Values are all better in terms of player strength, while the teams with more cash tend to improve better over the course of the season, so it's worked out well the past two years in this regard.  Strength + potential is what we generally look at when picking the pre-season favorites, so this ranking value gets the full scale of 1-32.  Take the combined cash+PV score and rank them all from most to least.  Then assign points from the top down with 32 for the top team, 31 for 2nd, etc. through 1 for the worst.  If a tie, then all teams tied get the higher score, and the next one down gets the rank they would have if the teams didn't tie (i.e. if Teams A & B tied for 3rd best, then Team A 30 points, Team B 30 points, Team C 28 points - not 29 since that would have gone to Team B, Team D 27 points, etc.)

2) T11.  It matters a whole lot.  Rank the T11s top to bottom and give them full points, of 32 for the best down to 1 for the worst.  If tied, then all ties get the points for the higher slot.  This system keeps the T11 relative to every other team for the whole season, which is a good thing.  If your T11 doesn't improve and the teams around you in total strength all do, then you should slip in the ranks.  Holding steady in these leagues is a BAD thing, since if you're not improving then you're losing ground, because the teams above and below you are improving regardless of what you might be doing.

3) T16.  Doesn't matter as much as T11, yet still important.  When suspensions/injuries pop up (not if.  When. Especially when playing Dr. Death, Dr. Doom, El Presidente, El Vice-Presidente, The Bostonian Hitman, The Imp Lord, or other hard-nosed manager), the T16 helps to cover holes.  T16 is even more important during a four-match session if a team wants to have any hopes of surviving in the league while advancing in a Cup or two.  Still, it's not as important as the T11 in winning any one given match, so half-value seems good.  Again, rank each team in T16 score, and assign points top to bottom.  Only for this one the top two teams get 16, the next two 15, etc. down to the bottom two teams getting 1 point.  Ties are assigned the higher value as with the other scores.

4) CP.  A team's total reflects more what it has already done (i.e. wins) than what can be done.  However, the more CP a team gets then the stronger the team can become.  So it does matter, just not nearly as much as what's already there in terms of team SL.  So about a third as important worked out to 1 point given for each 5 CP a team has, rounded down.  5 is good because that's also what you get for a draw, or 10 for a win, so going by 5's matches the style of the resource in the first place.  The theoretical CP minimum for a team is 0, if they had all youth players or all the older players were injured and they lost all their matches that session while using all CP from the session before.  By the '1 point per 5' system, you could have up to 4 CP and still get no points.  The actual lowest value ever seen so far is 8.  The highest in theory is 85 CP.  That would require 4 wins in a 4-game session (which has been done), plus carrying over 15 CP from the last session (also been done), and carry the full 30 age I+ players on the roster with no youth and no player injured below EL -2 (never been done).  Typically, totals over 60 are rare and the high 40's to low 50's are the usual mark for most teams who are doing very well.  That means the CP totals will add between 1 and 14 points to a team's total.  Up to 17 is possible, as is a 0, however those will probably never be seen.  In tracking this for the past two seasons, most teams fluctuate between 30-45 CP any given week, with lower and higher numbers rare except for the best and worst 3-4 teams in the league.

5) Backroom Boys.  All things are not equal.  There is no way that you can convince me that a Groundskeeper BRB is the same worth as a Youth Coach BRB.  An extra two area total points to distribute with the groundskeeper, only on home matches (guaranteed 9 games per season, max 18 possible if every FA Cup and then Shield match and every AP Cup match is played at home), does not equal being able to get critical QNL for an injured youth player.  Youth Coaches get 10 matches minimum and maximum playing time, so a hair better on the low end and trailing well behind on the high end compared to the Groundskeeper.  However, equally important is that if the Youth Coach is helping a Schoolboy for the session, then you can play a player in that spot at a far higher SL instead.  Potentially, this could easily be 7+ SL difference being added to an area for the match.  This contribution of helping the Youth develop AND helping the team win at the same time can't be matched by any other BRB.  So I have them down as being twice as valuable (more actually, however the points worked out well at 2 each).  A physio only helps if an injury occurs, and in some cases seems to do way too little as they can only keep a player on the field if you luck out and get a "0" on the injury table.  A -6 to -5 can be helpful, yet the player is still gone for two sessions in most cases, so they also get just 1 point.  Coaches are nice in saving some CP, yet most teams rarely coach up more than 3 players per session due to CP totals.  So that saves just 3 CP per session for most teams, or 30 for the season.  That's just 2 SL-worth of CP for a T11 player.  Being able to pour that 30 CP into fitness helps more, however that doesn't carry over from season to season.  So the net worth is 2 SL per season spread out over 10-15 matches, or about what the groundskeeper adds in area totals over the course of the year.  So 1 point for a Coach as well.  To sum up, 2 points for each Youth Coach (4 max) and 1 point each for a Coach, Groundskeeper, and Physio (max 1 each, 3 total) for 7 points possible total.  That's about 1/4 of the value that a T11 score can give, which seems about right.

So the final tallies are:
Bank + Player Value (1-32 points possible)
T11 (1-32 points possible)
T16 (1-16 points possible)
BRBs (0-7 points possible)
CP (0-17 points possible)

Totals: 3-104.
So subtract 2 from the total and you have the final rank score of 1-102.
A "1" would mean that the team is absolutely the worst in the league and really can't get any worse.  Just shoot them.  A 102 would be "perfection," though again that would require 85 CP in the bank.  A 4-win session (40 CP) and 22 healthy players (8 youth on the roster for the Youth Cup matches) and some CP carried over (8-12 points while saving up to add an SL to a T11 player) would just break 70, and would give a 99 score if that team were also tops in the other categories.  For practical purposes the pinnacle of achievement is thus 99.  1-99 is a pretty good scale, as the wide ranking differences then clearly distinguish between the best teams (90's), the league averages (50's to 60's), and those teams in pure rebuilding stages (20's and below).  In case of ties on final points, better T11 is 1st tie break, then better T16, then more Cash+PV, then more BRB points.  If still tied after that, flip a coin.

While a straight ranking just shows 16th to 17th place, the 1-99 system gives a good feel for whether that was incredibly narrow (1-2 points) or whether there was a significant talent/potential gap there (10 points or more) between the teams.  Just because a team is right above you in the rankings doesn't mean you are anywhere close to imminently overtaking them.

Here is what this ranking system would like for the start of season 5:
1. NOR, 94 points
2. TRA 93
3. CRD 89
4. QPR 83
5. BRE 81
6. HER 79 (T11 - 11.36)
7. PNE 79 (T11 - 10.91)
8. PLY 75
9. WAT 73
10. LEI 67
11. SHU 66
12. CHF 63
13. PV 62
14. LIN 61
15. FOR 56
16. SWI 55
17. BHA 54
18. BRI 51
19. SU 50
20. LEW 48
21. CRE 35 (T11 - 8.64)
22. DAG 35 (T11 - 8.27)
23. DER 34
24. WYC 33
25. WES 29 (T11 - 8.09)
26. LEE 29 (T11 - 7.55)
27. DOR 18
28. COV 17
29. BLY 16
30. OLA 15
31. SWA 12
32. CAR 7
The average value is 52.

If you had to rank the teams in the league and predict division results, is that at all close to how you'd have them set out?  Within a place or two in each division at least?  If so, then it seems pretty good.  If not, let me know how you'd make it better!

Readers Comments

I think this is the best method to rank the league so far.  It is strange to see LEI ranked at #10 while the team currently resides in D3.

Wish I had the time and abilities to make suggestions on this, but I don't.  Then again, I am happy with the method being used.

Brian Beerman on Wednesday, May. 26th, 2010 at 11:02 AM
 

Leicester is a great example of the comparative ranking benefit that you don't get in a straight division preview.  If you look at their overall stats, there is no way this team isn't getting promoted this season barring disasterous luck on a scale yet unseen.

6th best player value total at 7570.  Only 5 Div 1 teams are ahead of them.  That's right, they've got more potential on this squad than the entire 2nd Division thanks to a whopping seven age I players at SL 9 or 10 to start the season.

12th best T11, with only Watford and Chesterfield (2nd Division expected champ and promotion favorite) outside of the first division having a better top line for now.

They improve to 10th best at T16, being better than two Div 1 teams and having just Watford better outside the first division.

They have 7 legitimate T11 players, and any of those 7 age I players can be a great addition to any lineup if there is enough CP to go around for them all.  Likely, that won't be the case and Fran will need to figure out which two he needs the least and focus on maxing the others.  Likely the Gk, Sw, and top Df, Mf, Fw get attention and the other two become backups.

The only way LEI doesn't storm the third division with a shot at going undefeated is if all six of the quality apprentices get max attention to.  With the players above them, I think that would be a mistake.  Max out just the top at each position, or perhaps both at forward as that's the thinnest area on the team right now.  The 2nd App at Df and Mf just need to be SL 5 to survive aging and earn CP for two seasons before going away.  Focusing on getting playing time for all those guys will cost this team games, which takes away 10 CP per shot which is needed to max out all those age I guys.  By focusing on just 3 or 4 Apps, LEI can be nearly unbeatable in the 3rd Division and rack up enough CP to march into the 2nd Division as the favorites there with a very potent T11 next season.

If you meet them in a Cup match, do not take them lightly!

Kevin Martin on Wednesday, May. 26th, 2010 at 5:09 PM
 

Yeah, Jake was 'delighted' to see that the AP Cup draw had managed to survive the rollback!

Andy Bate on Thursday, May. 27th, 2010 at 10:15 AM
 

Great ideas from Kevin on this.   I now have this implemented.  Click Rankings at the top!

Al

Allan Sellers on Monday, May. 31st, 2010 at 12:58 AM
 

Liking the rankings!

Rob Peterson on Monday, May. 31st, 2010 at 8:45 PM
 
 
 
Terms and Conditions