Betting

Enter your bets for UAH’s September Reading!

58 Comments

The August betting pool was fun, so we are repeating for September. Enter your projection for the value Roy Spencer will report for UAH’s September Lower Troposphere reading.

You can enter until sometime late on Sunday September 20.
Results older than 2 months are not displayed avoid taking my site down.

Frequently asked questions:

  1. What’s a Quatloo? A quatloo is a monetary unit on the planet Triskelion.
  2. Can I change my bet? Yes. Just enter a new bet. The program will count your final bet.
  3. Could the script used for betting be screwed up? Yes. If it’s pretty new and not entirely debugged. Let me know if anything goes wrong.
  4. Will the script used for betting be improved over time? Yes. Sometime. When I’m not arguing about The Second Law of Thermodynamics.

58 thoughts on “Enter your bets for UAH’s September Reading!”

  1. With my newly patented and super secret method I have again made my fearless, and at this point most accurate predic.. err projection.

    I also expect that I shall again beat “you know who” no matter how many times she may stuff the ballot box.

    BTW, along the lines of improving the script, I am assuming that effort is being monitored by a truly independent party. No implication in that assumption, of course.

    I am still looking forward to buying that pony.

  3. Maybe my math is rusty, but it seems like it would be rather tough to double exponentially.

    Also, its a tad early, but I’d be happy to wager on sea ice next year being lower than 2005 :-p

  4. Zeke–
    Bets on 2010 sea ice won’t open for a while! If I had to bet today, I would also bet that 2010 summer sea ice will be lower than 2005 sea ice too.

  5. Of course you are right my bad. should be doubling. Lucia I am definitely willing to challenge that one. Curious why you did not do ice bets this year though (Sly ol me)

  8. You guys seen Tamino’s latest charts disputing the Arctic temperature charts. He won’t reveal the variables for his smoothing. They look a bit off.

  9. MikeN– I saw Tamino’s post. I agree that Tamino is evasive when he answers questions. He does some slippery things (like give “lessons” on how to do statistics, then apply t-test to data that is clearly bi-modal, but not show that etc. ) So, I tend not to trust him to make non-tendentious picks when doing “analysis”.

    But as far as I can tell, in this case, the only purpose of smoothing is to make pretty pictures. Does it matter what smoothing he used to make pretty pictures?

    But his unsmoothed graphs show warming.

    If you think his value are wrong, look up the data from the stations and compute the unsmoothed graphs. If you want, smooth it. See if you get something different. But if the underlying graphs show warming, smoothing does nothing but smooth. I doubt you’ll be able to make warming of the arctic go away by smoothing.

  10. MikeN–
    If you do want to have fun, it might be worth downloading the data Tamino looked and seeing if the residuals to the fit are normally distributed. Way back when, he did an analysis of data in Vaexjoe Sweden, applied a correction for autocorrelation, and applied a t-test. The t-test portion is based on the assumption the residuals were normally distributed.

    I took a look at the histogram,

    (Temperatures are in 10*C.

    But, as you know, Tamino doesn’t like questions about his inspired ‘simplifications’ . . .

  11. Arctic warming is pretty unambiguous, Will and Watts notwithstanding :-p

    What strikes me as funny is that all the discussion on Kaufman et al centers around recent temps rather than proxy reconstructions, where I imagine the real uncertainty (and novel findings) lie.

  12. Zeke–
    Real uncertainties are always in the proxies! The other real uncertainties are in the estimates of uncertainty intervals for “climate trends” — that’s why I brought up the histogram.

    Certainly, the artic has warmed. But if one is doing an statistical fit to the data, are Tamino’s error bounds appropriate for what might be a red noise process? Maybe they are; maybe they aren’t. Who knows? If they aren’t, how do we interpret the meaning of the brisk regional warming?

    That said: I do think it’s warmed. If Lucy or MikeN think the data say something else, it should be easy enough to download the data Tamino got and explain why they think the recent temperatures have not warmed.

  13. Zeke Hausfather (Comment#19786)-What’s unambiguous is that during the mid century cooling period, the Arctic cooling 9 to 13 times faster than the globe, and in the first (natural) warming of the twentieth century it did so something like 4 or 5 times as rapidly. by comparison, the recent warming is something like 2 to 3 times as fast as globally. Clearly the Arctic can change naturally rather dramatically. Given that I know of no model which can explain the rate of global change from 1911-1941 (it wasn’t TSI since even assuming erroneous trends models don’t reproduce that warming) the situation in the Arctic is even worse. Me and Nick may be close to agreeing on the role of natural variability in the Arctic:

    http://www.climateaudit.org/phpBB3/viewtopic.php?f=3&t=778

    It seems that at least some Arctic warming is connected to ocean circulation changes not AGW.

  14. You know… I’m downloading data and looking at it in a simple minded way. Unsurprisingly, it has warmed in the arctic. But Tamino really is “the gift that keeps on giving”. Sheesh!

  16. Yea, I’m going to look at the data soon. The issue I had was with his smoothed data. He criticized Lucy for cutting off the last 6 years of data, but 4 of those 6 extra points were lower. There is warming, but the trends look a bit like the Rahmstorf smoothing.

  17. MikeN
    Indeed. That is part of the gift that keeps on giving. With respect to grid 0, it’s a bit ridiculous for him to plot temperatures since the 1910s. There were only two stations. Plus… when he writes something like this:
    “For those who want more than just the last 30 years, here are moving averages of the grid-wide average:”

    What are those? 5 year moving averages? 10 year?

    It’s not quite right to criticize Lucy for her choice of time periods and then picking a moving average that hides the most recent downtick.

    It’s true the trend is up. But what I find mystifying with Tamino is why all the bizarre machinations? Even with the recent downtick, the trend over 30 years is up. Is he afraid if ‘believers’ see 2008 happened to be sharply down, and that the data are noisy, they will somehow say that negates conclusion the general trend looks up?!

  18. Remember that AGW also needs the warming to be “unprecedented”, to show a current warming trend along with warming also in 30’s/40’s or during MWP does not help.

    Kaufman et al gives an extended cooling trend followed by an unprecedented warming trend – just what is required.

  20. The NYT says “tens of thousands”; the Daily Mail says “2 million”. Reminds me of the good ol’ crowd estimation days of the anti-war protests back in ’03 when I was a wee idealistic college student.

  21. Zeke–
    Clearly, they need to sell numbered tickets so we’ll all know….

    You’d think by now journalists would have some decent benchmarks to estimate the size of protest crowds in DC.

  22. lucia (Comment#19794)
    September 12th, 2009 at 1:46 pm
    “”MikeN Indeed. That is part of the gift that keeps on giving. It’s true the trend is up. But what I find mystifying with Tamino is why all the bizarre machinations? Even with the recent downtick, the trend over 30 years is up. Is he afraid if ‘believers’ see 2008 happened to be sharply down, and that the data are noisy, they will somehow say that negates conclusion the general trend looks up?!””

    Lucia, you have to remember that Hansen did it in 10 years. As an engineer, you know that life is path dependent. I also am a luke warmist. I do not deny warming, I just maintain “what is good for the goose is good for the gander.” Just as you select 2001, or Steig date selection of manned stations, because the selection is after the IPCC statements, one cannot (honestly) quibble if one uses the same criteria that the 10 year Hansen method did. Methodology rules. If it was correct for Hansen to use a 10 year starting at the bottom of a temperature rate, it is correct for Lucia to use a 10 year at the top of a temperature rate.

    The problem for Tamino, as you have shown with the “actual” temperature noise is that if the 10 year (2011) is less than the IPCC, if he/she/it were honest he/she/it would have to admit that the models are noiser, AND do not correctly predict CO2 vs temperature. The point of sensitivity/temperature over/under prediction is that the IPCC framed their claim with respect to a negative. The IPCC could not find a “natural” cause, therefore all, and I do mean all, fits are GHG’s. This is the paradigm the IPCC set up. If there is any proof that the senstivity is less (in the aggregate) than the IPCC stated, by definition, it has to be GHG’s are less than claimed. This is the point of their claim that since about 1950 it has not been natural but AGW. To retreat is to admit the argument is falsified. To admit it is falsified is to falsify their projections.

    It is not a minor issue, IMO.

  24. John–
    2001 is not a “top year” for warming. It’s a “la nada” year in the El Nino cycle. I picked it because that’s right after the SRES were frozen.

    The models look bad if we start in 2000, 1990 or 1980 too.

  25. Yes, I know that it was not a top year; it is that you used a criteria just as Steig did. That you get catcalls and he got accollades was my point.

    The problem with Tamino is that he is moving the goal posts not you.

    Sorry for being unclear.

  26. I’m still on the previous chart of trends for each station.
    The later grid averages post feels like there is some manipulation going on, but it’s hard to say until I see the data.

  28. Lucia, one of the posters at reason stated that MSM was estimating the crowds from 1.5 million to about 2 million…the largest tax protest in American history.

  29. JohnF–
    Zeke said the NYT said “tens of thousands”. Given the dispararity in estimates, I thought I’d look for photos. Admittedly, I still don’t know because I don’t know if overflowing streets like that is “tens of thousands” or “hundreds of thousands”. I suspect you’d have to be in a helicopter way above the city to get a photo that shows “millions”.

  30. One observer who wrote the reason article stated it was in the 6 figure range at least. Others estimated the 1 million plus.

  31. MikeN–
    Something very odd happened with my spreadsheet.
    But it looks like with respect to grid 0:

    * He applied a 5 year average for the “moving average”. That choice hides the noise, including the recent drop, and also makes the more recent warming look more dramatic than, say 10 year smoothing.
    * He is making some judgments comparing current data based on 17 stations to 1920 data which is, in some cases, based on as few as 1 stations and I, if I recall correctly, never more than 3. (There were four stations in existence before 1930, but the measurements are spotty. (I’d be really, really cautious about drawing any conclusions about based on comparisons to any data before the 40s.)
    * It is definitely warmer now than in the 40s, 50s, 60s and so on.

    So, basically, it looks like this to me: It’s easy to see that grid 0 is warmer in the past 50 years. But as usual, Tamino stretched a number of points.

  32. The largest crowd I’ve personally ever seen was a 250k person anti-war protest in NYC in 2003. That one filled up Broadway for 30 continuous blocks. So I’d say the image above definitely suggest over 100k, but beyond that, who knows.

    The more relevant political question is the composition of the crowd. If its made almost entirely of people who didn’t vote for the current president/congress in the first place, the clout of the movement will be somewhat limited.

  33. Lucia,

    Getting “good” numbers about large-scale stuff is pretty dicey, as we are all well aware. Who wants to take the time and verify or even has the knowledge of how to get it done? Doesn’t matter what the subject is.

    Looks like lots of people in the pic, though! 😉

    Andrew

  34. Zeke

    The more relevant political question is the composition of the crowd. If its made almost entirely of people who didn’t vote for the current president/congress in the first place, the clout of the movement will be somewhat limited.

    Sure. To some extent whether or not these people voted for Obama or democrats is one of the relevant political questions. I’m not sure it’s “the more relevant” question, but it’s a relevant one.

    However, with any protest, there can be an effect of simply showing others that many people hold some particular view point and getting publicity for that view. I think that’s one of the reasons groups bother to organize these things.

    So,I think both the size of the crowd and its demographics matter and possibly equally.

    Also, I think whether or not the news agencies are fairly reporting numbers is an important political matter. If the NYT is reporting “10s of thousands”– a phrase which people are likely to interpret as being as low as 10 thousand, but not over 90 thousand, while MSN is reporting 1/2 million to a million….what does this mean?

    I know crowds are difficult to estimate, but there is a huge difference between “10 thousand” and “a million”.

    Is either (or both) of the agencies being tendentious and exaggerating in the direction the “like”? Is there something weird in their reporting channels? Or is this just an issue of timing? Did the NYT write something at 9 am while groups were gathering and MSN write something after the crowd coalesced?

    (Now I’ll have to go see if the NYT still says “10s of thousands”!

  36. Estimating the crowd numbers also could be a topic for a bet 🙂
    But I’d like to stick to the ones related to climate.
    I have of course allready placed my bet for September 09.
    I am confident that our gracious hostess will open for an ice extent bet in due time for the next melting season.

    We should perhaps also find an agreement on what measure we are referring to (e.g. Sea Ice AREA or Sea Ice EXTENT), and what algorithm they should be calculated by.

    However, Zeke’s wager is almost irresistible to comment, and I am happy to go on record by a prediction of NH 2010 ice (extent and area :-)) LARGER than 2005.

    Cassanders
    In Cod we trust

  37. When I was young and doing a journalism degree, my lecturer, who was a political journalist, said that 10,000 demonstrators = 1 (non-form) letter writer in weighting in a politician’s mind. And this was in Australia, which has a much smaller population than the US, which makes it harder to gather 10,000 demonstrators.

  39. “Bet that 2009 NH minimum ice extent > 2005”: wasn’t this statement already wrong at the time at which vg posted it?

  40. David–
    Lots of letter have been written too!

    I’m not sure demonstrations are to influence congressmen. They may be to influence the public. People can be odd. Some believe things because they think the belief is fashionable, to fit in, or simply because they think if lots of people believe something it can’t be wrong.

  41. Lucia,

    True. People do like to tend to jump on bandwagons. We are very tribal creatures. And the article from fivethirtyeight tended to refute me somewhat.

  42. And I should clarify here that the whole jumping on bandwagons thing is regardless of whether the position is correct or not.

  43. Leave it to Nate Silver to preempt our discussion :-p

    Looks like, at least in this case, the Times was right, though I agree that “tens of thousands” tends to indicate closer to 10,000 than 100,000 on first glance.

    Apparently I’m not very good at estimating crowd size from pictures after all.

  44. Yep. Looks like the higher end of “10s of thousands”. I don’t think many people can estimate crowds from that picture. At best, you could get a lower bound because you can’t know how many people are on side streets etc.

    I had the same problem with estimating for the anti-congressional power plant protest. I started watching the video– people hadn’t arrived yet. I posted an estimate when the protest was supposed to “start”. But the cameras weren’t on the crowd who were evidently grouping together in a park, and planning to ‘arrive’.

    Later they arrived. It still didn’t look like a huge crowd– but it was a lot more than I’d noted in one of the opening comments when practically no one was in the video.

    Even at that, we knew we couldn’t see all participants in the video. Some were blocking other gates and so not captured in the video.

    To some extent, the same holds true here. Any video image is likely to tell us the lower bound on the count. Streets and sanitation, public transportation and the police are likely to give better estimates of the count.

    Looks like a lot of people to me — but not 2 million. Two million would have been three times the population of Milwaukee, four times the population of Seattle. There are a lot of people in the DC area, but that would have been one heck of a lot of protesters!

  45. It seems to me that the September anomalies thus far on UAH are exceptionally high, even for El Nino. It’s currently 1.3f or so above the 30 year mean. That seems pretty high for what is a mild El Nino to date …

    Does anyone have any thoughts?

  46. JK_- Do you mean the Channel 5 AMSU? Yes, they started rising again at the beginning of September. I didn’t look at them the past few days, but Channel 5 is recording historically high values.

    Bear in mind: UAH uses the aqua satellite because the AMSU drifts. The AMSU has been coming in high relative to UAH the past few months. I don’t know why, but presumably it’s drifted somewhere a little bit “warmer”.

  47. Thanks for the info – I was certainly shocked by the sudden jump in September. I guess we will see how Aqua comes in.

    I also have not gotten a clear answer on how long the lag is between El Nino (say the readings in either Nino 3.4) and atmospheric temperatures.

    I’ve heard 6 months, which places us at about a 0-ish Nino 3.4 reading in March, with the current peak in 3.4 hitting about 3 months ago (early July-ish). Does that imply at least 3 more months of Nino-attributed atmospheric rises (assuming we level out in Nino from here)?

    Or is this 6 months that I’ve heard incorrect? Have you ever heard a clear answer?

  48. JK–I’ve heard 3 to 6 months lag also. It’s something determined empirically. The Nino 3.4 reading itself is expected to rise. If so, we expect warming to persist for a while. How much? Don’t know.

    That’s why we are watching.:)

  49. Watching, always watching. 🙂

    I found it curious, given ocean SST anomalies have been falling the past few weeks. Maybe we are seeing the big ramp up from June/July now, coupled with some drift. Even the ramp up (~.2-.3c difference in the overall SST anomaly) doesn’t seem large enough for the Channel 5 increase. But I guess we’ll find out in the coming weeks!

  51. Actually.. ice has taken a sharp upturn much earlier than 2005 so i am willing to bet a zillion quatloos as well that 2010 will be above 2005 . Meaning the 2009 15% ice is eaqual to the 2005 one and occurred earlier see previous post. http://ocean.dmi.dk/arctic/icecover.uk.php this is 30%, Do not confuse with JAXA 15% ice !
    Enjoy the coming ice age! At least you can have a scotch on the rocks without going to freezer!

  53. Yes, and my point was that if you go to http://www.ijis.iarc.uaf.edu/seaice/extent/plot.csv which is your preferred data source, the 2009 ice on September 11th, the day when you made your claim that NH minimum extent would be greater than 2005, was 5278594 km^3… compare that to the 2005 minimum extent of 5315156 km^3. I realize that you often exaggerate your belief that “things are about to get colder” and “the ice is recovering” just to get a rise out of warmers, but it is taking it to another level to make a claim that was wrong the day before you made it.

    Just to be clear: I’m no great shakes at predicting these things either – a prediction I made back in january was for “somewhere between the long-term trend and somewhat worse than 2007” with a best guess of “between the 2008 and 2009 levels”, and it looks to be obvious that even if ice has another few days of retreat left in it (and I’d give less than even odds for that at this point) that it isn’t going to get near 2008 levels, regardless of which dataset one uses, so I missed my best guess attempt (and the uncertainty bounds on my other estimate were… fairly large. And yet, probably not big enough – I love Granger Morgan’s 10 question test on trying to estimate 90% bounds on known numbers that shows most people greatly underestimate uncertainty).

  54. marcuis point taken was just eyeballing the graphs I suppose a few thousand may be relevant or not to the discussion

  55. Hi Lucia –

    I decided to bet this time after all. Partly, because it is only fair to give others a chance to get back some of my big pot and also because I want to try out another system of vaticination.

    This one is more touchy feely in nature, unlike the previous hard science, and clearly proven, biophysical model. 🙂

  57. BTW for Marcus as above. This posted on climate audit. This is what was meant me thinks previously. My apologies
    “Steve McIntyre:
    September 18th, 2009 at 9:51 am
    Revised 2009 is now above the 2005 minimum (which occurs in about 3 days). It’s up over 75,000 sq km over the past 4 days, 3of 4 being increases. Looks like the season is over.

    month day ice dd diff
    2002 9 17 5.844091 260 0.028338
    2003 9 17 6.033281 260 -0.008594
    2004 9 16 5.882813 260 -0.029375
    2005 9 17 5.422344 260 -0.025312
    2006 9 17 5.828281 260 0.022187
    2007 9 17 4.268750 260 0.001094
    2008 9 16 4.726250 260 -0.005625
    2009 9 17 5.326094 260 0.035000

  58. Put me down for +0.326C.

    I was waiting for last month’s AMO numbers to come out which were a little late.

Comments are closed.