Willis on why Piers Corbyn claims such a high success rate

Piers Corbyn (@Piers_Corbyn) says:
July 15, 2012 at 9:37 pm

John says he expects to arrive home between around 6pm & 6.15pm with probably a copy of the New York Times. He arrives at 5.55pm with the Washington Post. His friend Miss Right says “Well your estimate of what you said you would do is wrong so as far as I am concerned you are not even here”. He says “What’s up with that? ah, I forgot your first name was ‘Always’ ! ”

He is certainly there, but his estimate of what he said he would do is wrong,

Firstly the innuendo conveyed by Willis’ title “Why Piers Corbyn claims such a high success rate (for this period ~13-15th).
1. What success rate did I claim? Tell me? Come on? What ‘rate’ and why ‘Claim’?
Answer: I didn’t claim any specific rate, I noted some events which confirmed our forecast. What’s Up with that? Of course your innuendo implies something doesn’t it? To what end?

So you are going to start out with distortion? The title of the piece is

Willis on why Piers Corbyn claims such a high success rate

not

Why Piers Corbyn claims such a high success rate (for this period ~13-15th)

While you may have inferred that I was talking about this period, I did not imply it, nor did the title. In fact, I was referring to your claimed success rate of 85% which you made here, saying

Of 49 extreme events Trial forecast statements March to Sept 2008 The WeatherAction SWT forecast was confirmed on 42 occasions and there were forecast errors on 7 occasions (See Table for details)
Success rate 42/49 = 86% (ie better than 85%)

That was the context of the title.

2. Trickery from WE / Martin Gordon – portraying a chat on Facebook about one part of USA as ‘The assessment’.
Please get real and even fair. We have plenty more to say about the forecast on our website and more which we deal with internally. Why do you have to be so misleading?

Piers, you are the one who claimed success based on a Facebook report. It is certainly valid to object to that kind of claim.

3. Fortunately plenty of readers are not as stupid as W.E. etc appear to assume.

I assume our readers are the brightest of the lot, well educated, and thinking for themselves, and I am rarely disappointed.

4. No forecast is entirely right and some will be more wrong than others. For this period the three main thunder areas were all confirmed – see that Accuweather video – in the approx times and regions forecast (Pacific NW, North Centre and N/E). As a comparison of a forecast one might make by random choosing of past pressure maps I looked at all the past maps for 1-15th July 2011 and none resembled what happened as closely as our forecast. What does that tell us?

Piers, I fear that predicting “thunder” is meaningless. I’ve heard thunder when I couldn’t even see the clouds. Thunder is so common that it is not even reported in the severe storm reports.

5. Thank you Tom Rude, Brian D and Ian W and others for pointing out verifications of some specific events / regions around the US. One has to ask what is the chance of getting these three main thunder areas verified by luck , (along with any misses)?

Tom Rude? I find no trace of a “Tom Rude”. Brian D reports one lonely tornado the day before your forecast. Ian W. said nothing about the US, just about the UK.
Piers, I would hold up that last paragraph as a good example of how you claim success based on … well … nothing. That does not verify the “three main thunder areas” in any way, particularly since you also forecast “giant hail” and tornados.

6. What we get from W.E. is constantly flying to detail (hail size, precise locations, timings..) as a way of making us wrong – and we have some comments on that on our web:
http://www.weatheraction.com/displayarticle.asp?a=472&c=5 –
He shows a crass misunderstanding of our long range forecasts and their purpose (and note we define that not W.E.) and what others* say about them and worse still an unwillingness to learn. [*success measures come from users and auditors not us].
If W.E. really wants a short range forecast he only need look out of the window and not complain that I cannot can tell him two weeks ahead what detail he will see.
We have gone OTT with hail sizes in some areas (and UTT in Seattle but we have a reason) but SO WHAT? The point is
(i) We got the thunder regions and there was some hail.
(ii) There was even some large/giant (??) hail I don’t know how giant, certainly ginormous by English standards, in Seattle, described as the biggest ever in the region! – see the pic on
http://www.weatheraction.com/docs/WANews12No33.pdf
And note us NOT saying ‘there would be hail’ there is NOT us saying ‘there would be none’. These are NOT forecasts of all possibilities, read the title: “USA Key Developments & Extremes forecast July 2012″. Detail at this stage is lower priority and skill, which is why W.E chooses to “assess” it.

No, you didn’t get the thunder regions. The numbers of storms in the midwest and the Great Lakes/Northeast was approximately zero. Your claim that you nailed it is a joke.
No, you didn’t get the hail right. You claimed success for hail in Oregon, but you didn’t forecast hail in Oregon.
You claim that you NOT saying there would be hail is not the same as there saying there would NOT be hail, because you are dealing with “extremes” … that makes no sense. If you forecast hail in Illinois and it happens in Oregon, you don’t get to claim that as a success. It particularly makes no sense if you are forecasting “thunder”, that is not extreme in many parts of the US in July, it is totally common and expected.

I am pleased by our NW success here because while we have had tremendous success with eg NE blizzards in eg 2010-2011 these NW thunderstorms are more unusual than many other parts of USA and we have had difficulties in the past with getting to grips with the West / NW. So, subscribers who have asked for more there will be pleased; not that I expect W.E cares a damn about advancing forecasts. [What is WUWT for I wonder?].

Success? Piers, you are claiming success because it rained in Seattle on July 15th. Do you know how common that is? Take a look at the historical record. On average, it rains one day in four in Seattle on that date.

For those who are interested: We did decidedly under-forecast this event in the PNW, true, that is because I decided to tread gingerly there in view of the hurdles we have had to overcome to get here. Nevertheless so far, so good, or as they say in Tesco ‘Every Little helps’
(iii) We have been expecting more large hail around the world as opposed to small hail in line with our climate forecast of moving towards a new Little Ice Age and this greater prevalence of large/ giant/ enormous / big / whatever hail around the world has been well confirmed.

Piers, this kind of claim does your cause no good at all. Why? Because while it certainly may be true, without any kind of citation or backup, it’s just an anecdote. I know of no records to confirm your claim, although there may be some. But if there are, it’s your job to cite them in support of what you say.

7. W.E. a) Do you deny that this forecast period has skill – ie it is better than a random forecast?

Not my job. It is your job, since you claim it is better than a random forecast, to show that it is.

b) Do you deny that all objective assessments of our forecasts (taking a long enough period) in UK, Europe. USA, specific world events… show we have significant skill?

I know of only one “objective assessment” of your forecasts, which was of this forecast, although there may be more. In that forecast, you predicted 7 to 9 typhoons in the central/western Pacific in July 2008. In fact, your own citation said there were only two typhoons in that area during that time. You claimed 6 or 7 of those typhoon predictions as a resounding success, and whoever was doing the “objective assessment” was asleep at the switch and backed you up on the claim and said yes, it was a success.
In the same forecast you also said there was a 50% chance of typhoons around 28-30 July in the same region, viz:

Around 28-30 July
(action but NOT named storms)
Pacific active Tropical depressions likely but only 50% risk of developing into Typhoons forming (if they do) in west central North Pacific and headed NW/N towards Japan but probably veering away to NE

You claimed success because there were no typhoons … but with a 50% forecast you could have claimed success if there was a typhoon.

c) Are you going to examine the whole season Tropical storm forecast 2011 which I sent Anthony last year, and users described as a game changer.

I’d be happy to, but I’ve never seen it.

8. Now to the bigger picture starting from ‘The top’ Please look at:
http://www.weatheraction.com/docs/WANews12No33.pdf
Our SLAT (Solar Lunar Action Technique) – yes “Nostradamus!” to you too but citizens are you going to get serious on this blog or join the dustbin of history? For What it’s worth my opinion of the WUWT set-up here has dropped, on a scale of ten, from 9 to 1 in the last few weeks). Yes, Our SLAT is about predicting events on the sun, their intermediate Solar wind / ionospheric / geomagnetic effects and then jet stream and frontal activity and pressure pattern changes and finally weather in regions. Look at the pdf:
1. Our prediction of Earth-facing (significant) solar activity around (just preceding) the R4 period 13-14th or so – CONFIRMED (no apology for the caps) by X flare / AR 1520.

Let’s start with the fact that you are claiming success for a date (July 12th) for which the solar factor was “NSF”, which I assume means “No Significant Factor”. This is totally typical of you, you are more than willing to claim success for something you didn’t predict.
How many periods of enhanced solar activity have you predicted this year, and how many of them were accompanied by giant solar flares? So far, just in July we’ve seen:
R1: July 1
R5: July 3-4
R3: July 6-7
R4: July 13-14.5
Coming up, we have:
R5: July 17-19
R3: July 20-21
R4: July 27-28
R5: July 30-31
So if solar storms are more likely when the solar factor is above say 3, that’s 15 days out of the month. But of course, you claim success even if there is no solar factor if it’s within a day of your forecast … so that ups the count to 24 days out of the month on which you would claim success if there’s a solar flare … and you think I should be impressed by that???

2. Proton flux hit (eg) and various other parameters – massive aurorae and upping of geomagnetic activity (also about the coming R5 17-19th – WATCH IT!) – CONFIRMED

Confirmed? What was confirmed? I don’t understand this forecast.

3. Three thunder regions in USA – CONFIRMED

Nonsense. There was less thunderstorm, rain, and hail in the Great Lakes Northeast and north central US than there was earlier in the month, almost no action at all.

4. APPROX pressure pattern USA – CONFIRMED

How does one CONFIRM an “approximate” pattern? Where is the data you used to CONFIRM it?

5. Various detail USA – NOT CONFIRMED (and some parts confusing, USA is large).

Same problem.

6. Widespread thunder FLOODS UK – CONFIRMED (with ‘brilliant’ timing) and I note in comments on WeatherAction a doubter has become a ‘backer’.

The UK? We’re talking about your predictions for the US, we don’t even have access to your predictions for the UK.

7. Approx pressure in Britain & Ireland – WELL CONFIRMED

I don’t have your UK forecasts, how would I know?

8. Met Office having to upgrade a yellow to an amber warning (a change we predicted) – CONFIRMED (although this is not Long Range forecasting it is ‘end-gaming’ like our correcting the hurricane tracks of standard models one day ahead – see our 2011 ATS work, enquiries to piers@weatheraction.com)
9. Jet Stream south of Britain – CONFIRMED
10. Thunderstorms etc N/W Europe – CONFIRMED
11. Tornado events N/W Europe – CONFIRMED although the very significant tornado wave in Poland was east of what we expected, however there may have also been events in Germany.
12. General circulation Europe – ESSENTIALLY CONFIRMED but the penetration of thunderstorms got into Russia ending (much of) their heat (for now) early.
13. Up ticks of frontal activity compared to standard Met in New Zealand – these are consistently monitored but we issue no specific forecasts of circulation. – CONFIRMED
14. Very extreme events compared with normal (not specific forecasts) in various parts of world – Enormous floods in West Russia and South Japan for example and unusually cold blasts in South Africa. – CONFIRMED

Piers, you can make all of the claims you want about things that we have never seen. You may be right, but there’s no way for me to tell, so why are you wasting electrons on the subject?

General point. The solar and frontal action in this R4 was probably a bit above what we expected, in some respects – towards an R5.
Piers Corbyn, WeatherAction

Let me explain to you the general problem I have with your forecasts, Piers.
1. Lack of specificity as to the event. If you forecast “thunder” in big red letters, I expect to see lots of thunderstorms. For you, if one man in the area reports hearing thunder, you claim success. If you claim “tornado swarms” and there is one tornado in the area, you claim success. If you forecast “giant hail”, even you admit that you don’t know what “giant hail” is, but any hail seems to bring forth a claim of “CONFIRMED”.
2. Lack of specificity as to the timing. You break things into 3-4 day periods, and then you claim success if something happens outside of those periods. You say that’s OK because you are forecasting extreme events, viz:

And note us NOT saying ‘there would be hail’ there is NOT us saying ‘there would be none’. These are NOT forecasts of all possibilities, read the title: “USA Key Developments & Extremes forecast July 2012″.

With three day periods, if you include a day on either side, basically there’s nothing left. If you claim extreme events as a success in periods where you have note forecast extreme events, sure, you can get a CONFIRMED for the extreme events period … but you seem to forget that that very CONFIRMED means that your lack of a forecast for the extreme event when it did happen is a failure …
3. Lack of specificity as to the location. You predicted hail in the north central and Great Lakes/northeast regions. I can’t find a single hailstorm in the indicated regions, although there may be one … meanwhile, you are claiming success for a hailstorm in Oregon.
Look, Piers, let me be crystal clear about my opinion of your forecasting ability—I DON’T KNOW. You may very well be onto something, you may well be ahead of the pack. The problem for me is that your vague forecasts and your fanciful claims of CONFIRMED when it is not confirmed in the slightest and the general lack of citations for your claims and the lack of historical records of your forecasts makes it impossible to say if you are right or not.
For example, do your R1 to R6 or however high it goes solar activity forecasts correspond to solar flares? I haven’t a clue, and you haven’t given us enough information to find out. Since we have good records on solar flares, and you must have good records of your back forecasts, it seems like that would be a no-brainer for you to check out. But I can’t check it out, I don’t have access to your past forecasts … and in that situation, the fact that you haven’t produced a chart to say “Look at how well the solar activity lines up with my forecasts” means that we have absolutely no way to judge the validity of your method.
And that, in turn, makes your claim of success when a giant flare happens in a time when you’ve predicted no significant solar activity, but by gosh it’s close to an R4 period, less than impressive.
In a nutshell, when even you don’t know what you mean by “giant hail”, saying “There was even some large/giant (??) hail I don’t know how giant, certainly ginormous by English standards, in Seattle”, how on earth are we supposed to assess the success of a forecast of “giant hail”? In the US, the weather service doesn’t even call hail “large” until it is over 2″, and they don’t have a category called “giant hail”, so how can we determine if there was “giant hail” or not?
(Although to be fair, in this case it was easy, the two highlighted areas around the Great Lakes where you forecast “thunder, hail, and tornados” and “thunder, giant hail, and tornados” in red, and for which you are claiming CONFIRMED, had no hail at all, no tornados at all, and little thunder … confirmed? NOT.)
w.