BS Rankings

Took a minute to see what happened if I averaged all the rankings folks over on BigSoccer are generating (2 Elo rankings and 2 numeric systems in total), and I've got to say, it looks pretty good. Let's just call em BS Rankings. I may try to do this weekly.

Teams ranked first by mean rank, then by median rank if tied, and if still tied, ranked by standard deviation of ranks, lowest to highest. Click to embiggen.

I actually think the combined rankings look better than any of the individual ones. I was impressed. The craziest thing to me is the team tiering. i.e. RSL is clear #1, Seattle is clear #2 , LA and MIFC are interchangable, Colorado, SKC, RBNY and Portland are all interchangable, Dallas, Chicago, San Jose, New England and Philadelphia are all basically interchangable, then there's Houston, then Vancouver, then Columbus and Chivas are interchangable, then Toronto and DC.

[Clever_Title]

Numbers shaking out well this week, with the possible exception of New York. Impossible team to figure out, seems like they're playing well and then lose to Columbus, draw with Philly, lose to Chivas, then bounce right back with some wins. Still sandwhiched between New England and Chicago, but those two teams are trending down and NY is, once more, on the rise, so who knows what'll happen against Toronto next week. Kansas City back to winning ways, Colorado stops their slide and Seattle and Montreal jumping up the table. Click to embiggen.

How the mighty have fallen... behind.

Catching up again, I Swear.

Okay, so this is NOT the most recent power rankings to reflect games from this weekend, this is everything up through the weekend before last. I'm not skipping it because rankings are dependent on past results, so I would need to update the spreadsheet anyway, and it's trivial to add the rankings and graphics once that's all done. So, without further adu (not a Freddy joke) here it is.

Lots of weird movement in the middle this week. I think callups and the like made things screwy, combined with a more dynamic model and a lot of teams seemingly unable to figure out how they want to play this season.

Table.

Graph.

That is all.

Sleep when youre dead, which may be soon...

Okay, getting caught up again. Numbers through June 19th. No commentary other than whoever said that having your second kid is easier than having your first needs to get punched in the dick until they're infertile. Movement might look a little screwy given the weird scheduling. Next week may not be better...

Anyway, rankings...

Graph.

I'm going to bed.

It's a girl, buy some pink shit

So, apologies, I had another kiddo on Monday and it's kinda thrown my mojo off, as this isn't exactly a high priority item.

But anyway, I'm just gonna copy-pasta this from BigSoccer. Don't likey? Deal with it.

Okay, so here's last weeks ratings. Glad I'm not gonna fall too far behind, as the games next week are still gonna be part of Week 16. Had another daughter on Monday (second spawn). Home on Friday. Running on a good 10 hours of sleep in the last 72. Weeee! I love late night coffee.

Anyway, here's the numbers. Hard to take anything too seriously, as most teams "cruised" on their Week 14 rankings, but still, looks pretty good. The gulf between the top 2/3rds of the league and the bottom third (Chicago, San Jose, Chivas, Toronto and DC) is actually pretty impressive... Make of that what you will.

The table.

The graph.

And, for good measure, he's a picture of me and the new spawn. Because this is my blog, dammit.

What is this I dont even

Sort of a weird week this time around. Montreal just can't quite grab that #1 spot. I think it's the games in hand that's holding them back, believe it or not. Colorado my big surprise of the week, jumping up 3 places. I triple checked to make sure it wasn't an error... seems legit. This is what happens when you weight recent form more heavily, I guess. Kansas City continuuing the gradual fall, and Chivas only staying out of the basement thanks to Toronto and DC.

Table.

Graph.

I may well fall behind on this in the coming weeks, gonna have a second kid born on Wednesday(plus or minus a day or so) so who knows how much sleep I'll get... we'll see.

Lucky Week 13?

So not a lot of movement in this weeks rankings, which may be a good thing. I like it when teams settle down and trends get less violent. Of course that doesn't mean there wasn't plenty of action happening. Portland making some noise, a reality check for Seattle, Chivas has, perhaps, bottomed out, and Kansas City very luckily manages to creep up a notch to #6.

Here's the full table.

And now the handy graphy-graph. Dallas on top once more, but Portland has very nearly caught them... Click to embiggen.

6 hours of no power!? Eat the children first.

Actually, no, I didn't lose power in the storm last night, though it sounds like a lot of people in the area did. We'll see how we fare tonight though...

Speaking of faring well, how about those Sounders, eh? Not too shabby.

The two midweek games threw this off a bit. Kansas City actually drops a little bit, despite picking up two road results (well, you can argue the DC one I guess). Mostly RSL's domination helped them up. LA moves up despite losing, again, helped by crushing Philadelphia midweek. Dallas still holding on to the top spot, Chivas droppping like a rock. With three Ps. How's that for not caring about spelling or typos?

Overall rankings look pretty good to me. SKC perhaps a bit low.

Here's the table.

And here's the graph. starting to see some real movement and trends here. Next week should be interesting for some teams.

I did what I swore I wouldn't...

So I told myself this season I was going to stick with the same methodology that I started with. And I thought I really meant it. But then I came up with something better. So I'm changing. I'm sorry. It's not you, it's me. Or something.

While the system had been working pretty well so far, it still didn't seem to capture 'momentum' shifts all that well. However, Itried to compensate that with adding a Flavor of the Week ranking syste. I've ultimatly decided that the 'mashup' system, using both systems and then combining the rankings, was probably the best, and when comparing it to how other sites rank teams, it looks like it's doing a better job.

Remember, these are still purely numerical, data driven rankings, that don't hvae any input from me other than results. I think it's become a very good way to creating "power rankings" that mirror what "experts" are saying and is backed up by results and not just hand waving mumbo jumbo.

Anyway, enough exposition. Let's check the numbers.

I'd note that I also did back calculate "mashup" rankings for each team over the course of the season. This will do away with the need for some of the extraneous info I had been providing earlier. From now on it'll just be teams rankings and movemnet form the last week and a chart showing ranking position over the course of the season.

Numbers. Yes. Here.

Dallas still at the top, New York with a big jump, and both LA teams with big falls (poor Chivas ;/ ).

Here's the graph for the full season. Check out Vancouver. What a crazy poor season they've had... Click to embiggen, I'm trying to go with a .png to see if that makes it prettier.

 Annnnd... that's all I have to say about that. More fun next week. Until then... courage.

Magic Bonus Update!

Magic bonus update!

For awhile I've thought about making a tweak to the formula to make it a little more sensitive to short term performance. I've sort of tried to account for that with the Flavor of the Week rankings, but today I thought I'd try to go ahead and combine the two and see how the average looks. i.e. the "mashup" system would be equally weighted between season long performance as well as last five game performance. I don't know if I like it more or less but I'll track it for a few weeks and see how it looks.

Edit - 

Huh. You know, after comparing the mashup numbers, I think I might like them more. They compare really well with what ESPN and MLS have while still maintaining some numerical defensibility and still identifying teams where the other rankings seem to be too knee jerk.

I think I'm gonna keep the mashup numbers and start including that weekly.

Snow? In May? I hate everything.

Plus I left my Kindle in Philadelphia and they're charging me $32 to ship it. $32!!! I'm even a Marriott Gold member! Insane. Crazy. Whatever.

Okay, so nothing too drastic this week except things starting to shake out like they should. Chivas in free fall, almost crashed out of the top 10, and LA comes back down to a reasonable rating. Below 200,000 is what I expect from just about everyone unless they're really flying, so we'll see what next week brings. San Jose the surprise poor performers of the year, and Kansas City keeps putting together the right result at the right time to stay in the top tier. Here's how the numbers looked this week.

Three folks tied on 9. I like how they're sorted right now, but if you think it should be different, just use the power of mental visualization to do whatever the hell you want, I don't care.

Here's the FOTW ranks. Looks pretty good to me as well.

And, last and least, the SHFR graph. I'm debating when I should switch it from log to regular, but for consistency, I'll keep it log for at least another week or two. Click to embiggen.

They have the plant, but we have the Power

Sorry I'm late. I'm out of town and working. I don't owe you anything. YOU CANT MAKE ME DO ANYTHING.

Also I'm tired.

No witty commentary. Here's numbers.

SHFR graph.

Week 9's flavor of the week. A little surprised SKC is still ranked so high. Two good weeks are keeping them up, they should be falling soon...

Where my boats at?

One more iteration in the books. Not a whole lot to say really. Midweek games throw a small little wrench into the system so Kansas City maybe "dinged" a bit more than they should after the cross country road trip while NY gets a bit of a bonus bump despite the midweek loss, but the rankings actually look pretty good to me. Chicago and Seattle almost out of the basement and DC and New England in need of some results, and quick. LA a bit vindicated in maintaining the top spot.

The numbers.

The handy dandy chart.

And the flavor of the week. A little surprised to see Chivas still so high, actually, but they'll likely drop next week.

Deluxe Apartment

A fair amount of mvoement this week in the rankings, though no one making major jumps. This is probably a good thing. Now that we've hit the "magic" five game mark I expect to see this sort of thing. Teams gliding up or down based on results and things starting to shake out. LA still at the #1 spot soley on the fact that they started the season so well, but if you check the chart yo ucan see they've been trending down hard.

Some how Chicago lost and still managed to gain points... essentially the system now sees a one goal loss for them as a positive, I guess (combined with some poor results for other basement dwelling teams). I think that's both funny and a little sad...

Anyway, here's how the table looks today.

Here's the graph. Note LA's trend if keeping them at #1 makes you uncomfortable.

And of course now that we're into the season enough, we can figure out a new flavor of the week. And no real surprise here, Dallas has been on a roll.

So there you have it, LA still on top, but just barely... will we have a new champion next week? Only one way to find out.

I'm flying to New York for work this afternoon and will be lucky enough to be in town for the SKC game, so I may not get around to updating this until the following week. Apologies in advance. Until then : courage.

BACKUP YOUR DATA!

Yeah, so I nearly lost everything on my laptop this week... fun stuff. Overnight data recovry worked though and now I've got everything backed up. I actually had it all backed up on my Google Drive accoun anyway (I love you Google Fiber), but was still a scary moment...

Anyway, enough about me, let's look at some numbers. A bit of movement this week. Unfortunatly LA and Seattle both have lagged a game or two behind eveyrone so their ratings still aren't the best, along with New England, but everything else is starting to take shape. Vancouver doing some downward movement and Chivas keeps climbing, along with Portland and Kansas City. LA at the number one spot but slipping a little bit. Still, as the only undefeated team, the math makes their rating look quite good. I imagine that'll settle down in the coming weeks. Montreal not quite out of their spot yet and Houston takes a bit of a hit as well. Chicago not quite doing enough to get out of the basement, though they'll likely jump Seattle next week unless something interesting happens. And what's up with New York?

Speaking of New York, I'll be traveling to the game next week in Harrison. Should be fun, last year was the first time I'd seen SKC win on the road.

Anyway, here's the table.

And now the graph, which is still a little hard to get a handle on, mostly due to LA. Still a log plot as well... Click to bigify.

And now we've got enough games to start doing a Flavor of the Week analysis. This is done in a similar way to the season long power rankigns except it only looks at the last five games. In that way it's more of a snap shot of how teams are performing at this instant and is more variable than the other rankings. Since it's early in the season this only impacts teams that have played six games. Essentially, it's ignoring the first game for those teams, so it's a little different than the season long rankings. It'll get more divergent as the season goes on.

Annnnd... that's all Ihave to say about that.

Goat Parade

So how bout them Chivasesessss huh? Pretty crazy stuff. I think, so far, they've got to be the surprise performance of the season, more so even than the poor start for San Jose. I love it.

Anyway, we've just about achieved normalacy as far as getting everyone to five games goes. Anything you still can't cope with is your own problem.. Next week things should start to look normal.

LA ties and brings them down to earths somewhat. LA has actually been riding that first game four goal pasting to keep their #1 spot, so once that drops off they should start to settle down, unless they manage to do it again... Will they be better with Donovan? Who knows. Chicago hasn't budged from the cellar. Kansas City some how manages to put in a dominant performance and drop a spot (it makes sense mathamatically, just trust me) and Vancouver is starting to drop off after a hot start. Here's the full table.

I think we're just about at the point in the season where we can start making some general statements about teams. I generally think that 5 games is a trend and 10 games is a good enough sample to give you a good idea of what a team is capable of performance wise.

Oh, right, I owe you a graph too, don't I? Here it is.

Roll on, Oblivion

Okay, ratings starting to come together again. Still a group of teams lagging a bit at 3 games, which is throwing off the comparison somewhat, but I think it's starting to gel nicely. Columbus may feel a little hard done actually losing points even though they won, but no ranking drop, and its hard to argue the other teams don't belong up there as well.

Full table. Click to embiggen.

LA still flying high and Montreal is making their case. Chicago stuck down in the basement, and look at little Chivas go. GO CHIVAS, GO!

Seattle and New York still with plenty of time to figure it out, but definitely not how they wanted to start the season. And who thought Philly would get it together after their three goal first half meltdown against SKC?

Edit - And, what the hell, never too early in the season for a graph, right? As always, click for pops.

SHFR trends for the first four week. Note that the y axis is a log. LA's high rating makes a linear chart hard to read. Should clear up in a few weeks once LA levels out a bit (assuming they lose a game or two...).

GODDAMMIT ITS STILL NOT WORKING

Actually, it's staritng to shake out okay. The teams with only two games played still throwing a little wrench in everything though, hence LA still ranked #1 despite their bummer draw with Chivas. And by bummer draw, I mean hell yes Chivas, you beautiful bastard goats you.

Chivas is totally my #3 MLS team.

Anyway, as I said, numbers are starting to shake out well I think if you take the 5 teams with only two games played with a grain of salt. Here's how the table looks.

We'll see how LA looks after another game or two. Same deal with Seattle. Also, this in no way incorporates CCL performance into it, so don't ask.

Consistent Inconsitency

So I noticed one or two weird things creep into my spreadsheet when calculating rankings this round, and it's because of the way the formula does the fading averaging over five games when five games have not yet been played. In this sense, Houston, having only won one game, cruises a little unfairly on that, while on the other end Seattle's single loss has then unfairly punished as they're behind even Colorado, who have lost two. I won't get into details, but rest assured, this should be sorted out once every team is at five games. Well, maybe a little detail. Basically, the normal rating formula is calculated using 1A + 0.4B + 0.3C + 0.2D + 0.1E. With only one game played, Seattle has 1 loss/1 game played = 1. With two games played, Colorado has 1 + 0.4 = 1.4 losses / 2 games played = 0.7. So Seattles one loss looks bigger than Colorado's 2 losses. It's not presented here, but Seattles calculated rating is -0.333 while Colorado's is 0.231, which is -0.333 * 0.7. Note that it works out like this only because all the losses have been by one goal...

At any rate, here's how the tables look this iteration

LA still at the top, Candian teams climbing, with Seattle and Chicago in the basement. Too early to tell anything, obviously, but it's shaping up to be a wild March I think.

MEANINGLESS UNFRIENDLIES

So I hesitate to throw my hat into the ring in Week 1 with some power rankings since they're totally unrelated to the Week 0 ones from a bit ago, but I decided, for consitencies sake, I'd try to get in the habbit of doing this every week.

The reason I hesitate is that these ratings really don't get off the ground every team has played five game (they use a five game running average in the formula), so the numbers start off really meaningless and get better as the season goes on. But hell, even if the math is silly, let's just go for it.



Actually nothing too terrible there just yet... But like I said, this is all but meaningless, just based on one game and has no relation to what happened last season or evaluating overall team strength. Give me a few weeks and we might be able to hammer out something a bit better.

Until then?

Courage.