Author Topic: ESPN Coverage (2012)  (Read 5964 times)

0 Members and 1 Guest are viewing this topic.

Offline The Chief

  • Posts: 30122
    • http://www.wnff.net
Re: ESPN Coverage (2012)
« Reply #175: April 19, 2012, 12:44:08 PM »
:|

^^best I could come up with for "blank stare"

Offline tomterp

  • Global Moderator
  • ****
  • Posts: 28210
  • Hell yes!
Re: ESPN Coverage (2012)
« Reply #176: April 19, 2012, 12:44:25 PM »
^^^So how does one explain that with popularity algorithms?

We should track the headlines day by day, and keep tallies by team of top 5/bottom 5.  Yesterday, while Guillen was #1, Yankees / Sox were 2-5.    Pretty standard it would seem that the Sox or Yankees are almost always going to be in the top 5 and usually more than once, no matter what happens with them on the field.

Offline MarquisDeSade

  • Posts: 13611
  • PK Subban will eat your soul.
Re: ESPN Coverage (2012)
« Reply #177: April 19, 2012, 12:45:14 PM »
And I do remember last season they had the Red Sox at about 99% with a month to go in the season.     :crackup:

And if luck wasn't the biggest driver in MLB some guy that placed two $500 prop bets on the Cardinals winning the NL and the WS wouldn't be sitting on freak you money right now.

Offline MarquisDeSade

  • Posts: 13611
  • PK Subban will eat your soul.
Re: ESPN Coverage (2012)
« Reply #178: April 19, 2012, 12:46:52 PM »
We should track the headlines day by day, and keep tallies by team of top 5/bottom 5.  Yesterday, while Guillen was #1, Yankees / Sox were 2-5.    Pretty standard it would seem that the Sox or Yankees are almost always going to be in the top 5 and usually more than once, no matter what happens with them on the field.

You do realize that those are often localized, right? 

Offline tomterp

  • Global Moderator
  • ****
  • Posts: 28210
  • Hell yes!
Re: ESPN Coverage (2012)
« Reply #179: April 19, 2012, 12:49:05 PM »
You do realize that those are often localized, right?

No, you mean based on IP addresses?

Those don't look very localized to me - not a single Nats / Orioles story in the top 10.

Offline MarquisDeSade

  • Posts: 13611
  • PK Subban will eat your soul.
Re: ESPN Coverage (2012)
« Reply #180: April 19, 2012, 12:54:44 PM »
No, you mean based on IP addresses?

Those don't look very localized to me - not a single Nats / Orioles story in the top 10.

Right, think about it this way - do you think more people in the DC vortex are interested in the Nats/Os or in their "hometown" teams like the Yankees and Red Sux.  The only acceptable answer, as of today, is the latter.

Online HalfSmokes

  • Posts: 15138
Re: ESPN Coverage (2012)
« Reply #181: April 19, 2012, 12:57:28 PM »
If it's variable, you would think user profile would influence it, and I've never seen ecac hockey on the front page

Offline MarquisDeSade

  • Posts: 13611
  • PK Subban will eat your soul.
Re: ESPN Coverage (2012)
« Reply #182: April 19, 2012, 12:59:10 PM »
If it's variable, you would think user profile would influence it, and I've never seen ecac hockey on the front page

Very little, most user profiles are garbage and are treated as such.  The referring data and passthroughs are what drive most of it. 

Offline tomterp

  • Global Moderator
  • ****
  • Posts: 28210
  • Hell yes!
Re: ESPN Coverage (2012)
« Reply #183: April 19, 2012, 01:02:34 PM »
Right, think about it this way - do you think more people in the DC vortex are interested in the Nats/Os or in their "hometown" teams like the Yankees and Red Sux.  The only acceptable answer, as of today, is the latter.

No, I think Nats/Orioles.  I see a lot more Nats swag around town.

Offline tomterp

  • Global Moderator
  • ****
  • Posts: 28210
  • Hell yes!
Re: ESPN Coverage (2012)
« Reply #184: April 19, 2012, 01:03:55 PM »
Very little, most user profiles are garbage and are treated as such.  The referring data and passthroughs are what drive most of it.

This would be easy to test, I would think.  Have Houstonnat, ericasnats or some other remote fans do screen shots at the same time some DC folks do.  I'm guessing they'll be identical.

Offline MarquisDeSade

  • Posts: 13611
  • PK Subban will eat your soul.
Re: ESPN Coverage (2012)
« Reply #185: April 19, 2012, 01:06:06 PM »
No, I think Nats/Orioles.  I see a lot more Nats swag around town.

Which is a non-testable observation and personal preference. 

Offline MarquisDeSade

  • Posts: 13611
  • PK Subban will eat your soul.
Re: ESPN Coverage (2012)
« Reply #186: April 19, 2012, 01:07:14 PM »
This would be easy to test, I would think.  Have Houstonnat, ericasnats or some other remote fans do screen shots at the same time some DC folks do.  I'm guessing they'll be identical.

That would all depend on what their history was and what amount of their browsing history they're sharing.  Like I said - a lot of the referral data could potentially drive that.

Offline houston-nat

  • Posts: 16385
Re: ESPN Coverage (2012)
« Reply #187: April 19, 2012, 01:07:33 PM »
Note they aren't placing too much weight on the performances to date due to small sample size.  But the longer we keep it up the longer we climb.

Didn't the Nats start the year at 12%?

Offline Coladar

  • Posts: 2820
Re: ESPN Coverage (2012)
« Reply #188: April 19, 2012, 05:08:08 PM »
Here are BP's odds, which are based on PECOTA projections for player performance through the end of the year, then run through a Monte Carlo simulation of the rest of the season being played out 1m times.  For the NL East -

Phillies - 58.5%
Marlins - 49.6
Braves - 42.4
Nats - 20.0
Mets  - 7.2

Note they aren't placing too much weight on the performances to date due to small sample size.  But the longer we keep it up the longer we climb.

And I do remember last season they had the Red Sox at about 99% with a month to go in the season.     :crackup:

This I find more acceptable than ESPNs. I just take issue with Philly being dead last yet running away with the playoffs playoffs 71%. With BP, and I'm not sure I like the Nats at 20% running away with the NL East at 10-3, but this is a much more logical percentage range factoring in we are only a couple weeks in. Philly at 71%, dead last, just irks me as being overly dependent on last year. Like I said, I'd rather everyone be bunched together which is what BPs, based on PECOTA, does.

In fact I just looked... Philly is 4.5 games back, dead last, at 5-7. 63.8% chance still. All the other three teams are deadlocked at 7-5 or 7-6, except the Nats, at 10-3, and we still aren't even 50-50, currently 48.9%. While the Mutts and Marlins are 16%, Atlanta 33%. Yet Philly is dead last, 1.5 games from not being dead last, and are still a near lock? Yeah, I'm obviously just not a fan of weighting it so much on last year's performance when a team like the Nats are running away with the division and yet a team in dead last still has such a significant advantage of playoff percentage likelihood.

But I do get its solely because of the small sample size for the current season, it's just as I said, I'd rather everyone be bunched together due to small sample than weighting so much on last season when a team is dead last, and another is running away with the division thus far. No way should Philly still be at 70% when the Nats still aren't even at 50-50.

Offline KnorrForYourMoney

  • Posts: 14322
  • Nats hitters = Maggie Lizer
Re: ESPN Coverage (2012)
« Reply #189: April 19, 2012, 05:21:47 PM »
No, I think Nats/Orioles.  I see a lot more Nats swag around town.

Those same people will trade out that swag for their Yankees/Red Sox gear during interleague.

Offline LostYudite

  • Posts: 758
  • Naaaa'aah-titude
Re: ESPN Coverage (2012)
« Reply #190: April 19, 2012, 05:23:36 PM »
This I find more acceptable than ESPNs. I just take issue with Philly being dead last yet running away with the playoffs playoffs 71%. With BP, and I'm not sure I like the Nats at 20% running away with the NL East at 10-3, but this is a much more logical percentage range factoring in we are only a couple weeks in. Philly at 71%, dead last, just irks me as being overly dependent on last year. Like I said, I'd rather everyone be bunched together which is what BPs, based on PECOTA, does.

In fact I just looked... Philly is 4.5 games back, dead last, at 5-7. 63.8% chance still. All the other three teams are deadlocked at 7-5 or 7-6, except the Nats, at 10-3, and we still aren't even 50-50, currently 48.9%. While the Mutts and Marlins are 16%, Atlanta 33%. Yet Philly is dead last, 1.5 games from not being dead last, and are still a near lock? Yeah, I'm obviously just not a fan of weighting it so much on last year's performance when a team like the Nats are running away with the division and yet a team in dead last still has such a significant advantage of playoff percentage likelihood.

But I do get its solely because of the small sample size for the current season, it's just as I said, I'd rather everyone be bunched together due to small sample than weighting so much on last season when a team is dead last, and another is running away with the division thus far. No way should Philly still be at 70% when the Nats still aren't even at 50-50.

Don't get me wrong - I love that the Nats are 4.5 up on the Phillies, but we've played <10% of the overall season.  Historically, teams don't change THAT much from one season to another, so last season's standings are still a much stronger indicator than current year performance as a metric for how things will look at the end of the year.  Over time that will shift, but at this point in the season it doesn't seem way out of whack to me.  4.5 games can be reversed in a week - just ask the Red Sox.

It's like the attendance thing - the signs are pointing up for the Nats, but it will take a long time still to get where we're going.

Offline Coladar

  • Posts: 2820
Re: ESPN Coverage (2012)
« Reply #191: April 19, 2012, 05:32:37 PM »
Don't get me wrong - I love that the Nats are 4.5 up on the Phillies, but we've played <10% of the overall season.  Historically, teams don't change THAT much from one season to another, so last season's standings are still a much stronger indicator than current year performance as a metric for how things will look at the end of the year.  Over time that will shift, but at this point in the season it doesn't seem way out of whack to me.  4.5 games can be reversed in a week - just ask the Red Sox.

It's like the attendance thing - the signs are pointing up for the Nats, but it will take a long time still to get where we're going.

Yeah, like I said a couple days ago, I get the rationale behind it, but I'd just rather everyone be bunched together due to the small sample size than weighting it SO much on last year's performance. Philly at 70% still, dead last, when the Nats aren't even 50-50, just isn't the way I'd prefer they do playoff percentages, although I understand the logic behind how they figure it and why they do it that way. And while it's only 4.5 games, dead last, they've got to jump over four other teams. Philly has dug a hole that certainly isn't insurmountable this early in the season, but I don't think now I buy them as a near playoff lock. They've got as good a chance as anyone, sure, and that's what I've been saying. Bunch everyone together. But I'm not sure I buy now that they still have almost a 25% greater chance of being in the playoffs than the Nats at this point despite the early state of the season. I'd rather Philly be 55%, the Nats 52%, and everyone else 30-40% in the NL East. I think that's a more logical determination of it than weighting SO much on last year, but I do understand the why and how of ESPNs determination.

Offline JCA-CrystalCity

  • Global Moderator
  • ****
  • Posts: 18093
  • Platoon - not just a movie, a baseball obsession
Re: ESPN Coverage (2012)
« Reply #192: April 19, 2012, 05:37:57 PM »
The ESPN odds come from coolstandings. 

Quote
How do we calculate these statistics? Basically we simulate the rest of the season millions of times, based on every team's performance to date and its remaining schedule. We then look at how many "seasons" a team won its division or won the wildcard, and voila - we have our numbers.

The trick, of course, is to determine what chance each team has of beating every other team. Our method is to use simple team statistics (e.g. runs scored and runs against) to predict how each team will fare against all others. For those of you familiar with baseball prediction, we use a variation of the Bill James "Pythagorean Theorem" to predict results. Pretty smart, huh? That's why we call this prediction mode "Smart mode".

near the beginning of the year, they do include projections based off of last year's performance, accorrding to their FAQ.  This gradually fades away over time.

Offline The Chief

  • Posts: 30122
    • http://www.wnff.net
Re: ESPN Coverage (2012)
« Reply #193: April 19, 2012, 05:40:38 PM »
Almost seems like we overperformed against the division last year relative to the rest of our record.  9-9 vs ATL, 7-11 FLA, and 10-8 each vs NYM and PHI.

So....  oh, 36-36 overall against the East.

Maths :lol: :?

Offline spidernat

  • Posts: 54503
  • The Lerners are Cheap AND Crooked
Re: ESPN Coverage (2012)
« Reply #194: April 19, 2012, 07:05:18 PM »
Those same people will trade out that swag for their Yankees/Red Sox gear during interleague.

Sadly this is true.

Offline DPMOmaha

  • Posts: 18219
Re: ESPN Coverage (2012)
« Reply #195: April 19, 2012, 09:44:23 PM »
^^^So how does one explain that with popularity algorithms?

Surprise team reaches arbitrary round number first.  Not that hard.

Offline The Chief

  • Posts: 30122
    • http://www.wnff.net
Re: ESPN Coverage (2012)
« Reply #196: April 19, 2012, 10:07:31 PM »
Surprise team reaches arbitrary round number first.  Not that hard.

Given the way you spun it before, I'd call this majorly reaching, but okay.

Offline DPMOmaha

  • Posts: 18219
Re: ESPN Coverage (2012)
« Reply #197: April 19, 2012, 10:12:43 PM »
Given the way you spun it before, I'd call this majorly reaching, but okay.

Not really changing anything.  All I said was that a nice pitching performance wasn't that big of a deal relative to the other headlines that day. 

Offline The Chief

  • Posts: 30122
    • http://www.wnff.net
Re: ESPN Coverage (2012)
« Reply #198: April 19, 2012, 10:15:01 PM »
Not really changing anything.  All I said was that a nice pitching performance wasn't that big of a deal relative to the other headlines that day. 

You explained it via click generation.  If a nice pitching performance by "a historically bad team" (your words) doesn't move the needle (also your words) then neither does reaching an arbitrary number of wins in April first, either.  Just my opinion.

Shrug.

Offline DPMOmaha

  • Posts: 18219
Re: ESPN Coverage (2012)
« Reply #199: April 19, 2012, 10:18:35 PM »
Again, it's all relative to what happens that day.  If the Nats keep winning, they'll move the needle a little more, and they did something unique.  But a nice performance from Edwin Jackson the first couple of weeks of the season isn't going to do the trick.