FORUMula1.com - F1 Forum

[CookinFlat6]

I think many are aware of GA's history of making questionable predictions while at the BBC. Now at Autosport his latest gem is this prediction, 'scientifically' worked out from the second test

1st MER
2nd McL
3rd FI
4th FER
5th WIL
6th SAU
7th LOT
8th STR
9th RBR
10th CAT
11th MAR

Now I am not sure whether to laugh or scream, but some of these just seem ridiculous. We will soon find out after the third test if he has done it again. What do we think about this?

Poll to run till after final test

[LiamCatterson]

What's wrong with EJ predicting? He's 98% of the time accurate

[spankyham]

If you're a Mercedes fan you will be hoping his predictions don't turn out like last year

Sent from my GT-I9500 using Tapatalk

[CookinFlat6]

Very true, the only fly in the ointment is GA placing Merc first

[Hammer278]

Him predicting Merc to be anything other than 1st will mean his brain has turned to scrambled eggs to it's a ray of light for him to get that right.

But Force India 3rd....Ferrari 4th.....and of course his dear old McLaren in 2nd. Thankfully they're not in first place this time.

[Jabberwocky]

The last 3... really. No way will RBR be that low. Plus Marussia and Caterham, I am sure the Ferrari Engine will move Marussia ahead

Sent using NCC-1701

[Hammer278]

Lol I didn't even bother to look so far down to check out RBR...wow!

[stonemonkey]

Is this prediction for the championship or just the opening race?

[CookinFlat6]

Is this prediction for the championship or just the opening race?

Gary Anderson predicts F1's pecking order
AUTOSPORT's technical expert GARY ANDERSON takes a look behind the lap times from last week's Bahrain test and gives his verdict on the teams so far

It's too early to start drawing any definitive conclusions on the relative performance of Formula 1's 2014 cars. But by taking the fastest lap achieved by each car during last week's four-day Bahrain test, it's possible to get a few clues as to who looks hot and who doesn't.

As usual, I've converted the figures into a percentage deficit to the fastest time (Nico Rosberg's final-day lap of 1m33.283s). This allows you to make comparisons not just on one circuit, but also over varying lap lengths and with historical data.

Its one of his scientific specials

[stonemonkey]

Hmmm, not sure how you can judge potential performance of teams while they sort out the issues we've been seeing.

[RyRy]

He's just used a ML algorithm to used past statistics with the test stats to produce those predictions.

We all know how things can change so these type of algorithms just don't work half the time.

Sent from my Nexus 4

[stonemonkey]

Sounds like a SISO algorithm.

[CookinFlat6]

or a GA in/GA out algo, pretty much same as a SISO

[stonemonkey]

Don't get me wrong, I think he has a fair bit of knowledge but this is as ridiculous as a Caterhams nose.

[Nin-Chin]

Who's the douchebag who predicted that he was correct because you are sooooo wrong (apart from the Merc bit obviously).