I haven't been sacked (yet) from my new job and my business interests appear to be thriving, so I am actively investigating the Jaguar XK8 market; expecting to spend between £10-20K.
I am regularly visiting e-bay and Autotrader to get a feel for the market. Almost all cars are sold as in 'excellent' condition with FJSH and nice photos to back the claims up. Most adverts admit to minor shopping trolley dents and worn drivers seat etc: fairy nuff.
What surprises me is the variation in age and mileage versus price.
Examples include:
17350 miles, 6 years old, £16995
21000 miles, 9 years old, £14445
42950 miles, 6 years old £28,995
86000 miles, 9 years old, £17995
148000 miles, 8 years old, £11850
I decided to spend a hang-over recovery morning typing in all the XK8s for sale on Autotrader (250 so far, more to come) into a data-base and see if I could find a relationship between Price, Mileage and Age.
In the end I came up with the following formula:
Price = ((log((1 / (M x A)) / 1000000)) + 0.5051) / 0.000058
Where:
Price = Advertised price in pounds
M = Mileage
A = Age in years
Would anyone like to comment on whether I should get a girlfriend?
|
>Would anyone like to comment on whether I should get a girlfriend?
You'll find that the Jaguar works out much, much cheaper.
Kevin...
|
good looking car especially in that lovely blue but the love affair stops in the drive,
suggest find a company that will hire you one for the weekend (google this) and see how the affair pans out
good luck
|
>>Would anyone like to comment on whether I should get a girlfriend?
Slinky Jag curves may help in the "getting" bit but the dodgy Excel habit may put the mockers on it being a long term thing.
|
>>Slinky Jag curves may help in the "getting" bit ...
Hmmm... AckShirley I'm a bit worried about that side of things. Shorely if a bird is attracted to you cos of yer flash motor then she might expect you to spend money on her? That's my theory, anyway.
Plus the fact that they probly wouldn't get to see your car until the next morning anyway; by which time they've already made up their mind.
>>...but the dodgy Excel habit may put the mockers on it being a long term thing.
Dunno 'bout that. Maybe some birds will be happy that I'm sitting there playing with me spreadsheets while they cook the dinner. At least they know I'm not looking at porn or out getting up to other mischief.
Anyway never mind this bird talk (1 throw-away phrase out of my whole post); what about some mathematical challenges to my theory. No-one's even asked me the size of my standard deviation. Ah well, maybe it's not that interesting. Sob.
|
Get a woman with a brain. It is far more satisfying...more so than you might think. If she has the aformentioned she won't give two hoots what you drive..(within reason) Although an Ariel square 4 and a Watsonian sidecar may be pushing your luck!!
VBR............MD.
|
|
|
I struggled to get any sensible numbers from the formula - perhaps I'm doing something wrong, but I enetered the formula exacty as posted. I got
Age Miles Price
0 0 #DIV/0!
1 12000 -165070.37
2 24000 -175450.71
3 36000 -181522.82
4 48000 -185831.06
5 60000 -189172.78
6 72000 -191903.17
7 84000 -194211.68
8 96000 -196211.4
9 108000 -197975.28
10 120000 -199553.12
11 132000 -200980.46
12 144000 -202283.51
13 156000 -203482.21
14 168000 -204592.02
15 180000 -205625.24
16 192000 -206591.75
Am I doing something wrong?
Perhaps, try This
P=R+((RRP-R)*exp(-c*(M/AveM)*A))
With
P - sale price in £
R - residual price that the car never drops below - 5K for XK8s?
RRP - new price
C - a suitable constant to set the decay - try 0.25 for a kick off
M - milage
AveM - average mileage at this age
A - age of the car in years
Using this, I get
Age Miles Price
0 0 66000
1 12000 51843.25
2 24000 40817.96
3 36000 32231.46
4 48000 25544.28
5 60000 20336.31
6 72000 16280.33
7 84000 13121.53
8 96000 10661.46
9 108000 8745.55
10 120000 7253.44
11 132000 6091.38
12 144000 5186.37
13 156000 4481.55
14 168000 3932.63
15 180000 3505.14
16 192000 3172.2
It may be that the milage needs another weighting, so the exponential reads
exp(-c*(M/AveM)-D*A))
but, I don't have the input data to work with to justify this.
I hope this is helpful.
Number_Cruncher
|
It's easier to set up an artificial neural network to provide results from a limited set of data.
It also allows the input of things like colours, optional extras etc that can't be easily handled by a formula.
This is left as an exercise for the reader!
--
I read often, only post occasionally
Edited by spikeyhead {p} on 09/10/2007 at 01:15
|
"It's easier to set up an artificial neural network to provide results from a limited set of data."
Speaking as an ex-AI researcher, I have to dispute that suggestion. It may be "quicker", it may be "more accurate", but I don't hold with "easier". Oh, no, indeed.
V
|
Unfortunately, I've just spent an hour and a half on a train thinking about this. I'm already married, though, so it's a change from spending the journey fantasising about Kirsten Dunst and a barrel of baked beans.
The basic problem you have here is that you're modelling asking prices. They bear no relationship whatsoever to anything to do with common sense. All they reflect is the individual's opinion of their vehicle, with a good dollop of wishful thinking and need for cash mixed in.
No, you need to model selling prices. AKA go to an auction and buy someone's month-old copy of the CAP black book. Better still, go to an auction and buy a Jag. Even better, get Joe Dowd (posts passim) to buy one for you.
V
Edited by Vin {P} on 09/10/2007 at 09:44
|
If you're interesting in hiring an XK8 to get a feel for one, I'm pretty sure Hertz and Europcar (among others) have so-called 'prestige' ranges that include them.
I looked into the possibility of hiring an XJ6 for a visit to East Midlands earlier this year - one of the hire fleets had everything up to Bentleys and Porsches, but only from certain towns/airports (not EMA). The four-figure deposit also raised an eyebrow, so check the small print on the collision damage wavier/insurance form carefully!
|
|
|
|
|
