Lies and dammed lies!

ercolano

Member
Messages
14
Reaction score
0
There are lies, dammed lies, and there are statistics. We all hate statistics, we all know how data can be presented in a myriad of forms in order to sustain, or discredit, whatever argument as needs dictate. But here in the Dpreview forums there is nothing we like more than disputable numbers; we can talk about them until the cows come home, or maybe don't come home. Whatever. Here goes:

The images posted in Dpreview challenges are annotated with the camera model and lens used to take the shot, it is potentially a great mine of information. Even more so analyzing the winners. And straight away we enter into a point of contention; are we analyzing which equipment is most likely to get a winning shot or are we simply observing what equipment is most likely to be used by the sort of photographer who is most likely to get a winning shot?

But let's not allow such pedantry to dampen out enthusiasm, let's move straight on to the details. Our sample data is taken from the first, second and third placed entries in 32 randomly selected contests, plus the winner of a 33rd contest to round things up to a hundred. Contests that were limited to specific hardware or were orientated around post processing techniques such as HDR were not included.

And the winner is...Canon. Or is it? Well, first let's look at the raw data: 41% of the entries were taken with a Canon camera, 30% Nikon, 14% Olympus and 10% Pentax, whilst Sony/Minolta rolled in with 4%. Oh yes, there was one image taken with a Lumix camera as well.

The trouble with the raw data is that it takes no account of market share. Just suppose that 99% of cameras sold were Canon, if only 60% of winning entries were taken with Canon cameras that would be a pretty poor show.

So in order to dig down into this data we need some figures for market share, which of course depend an awful lot on when and where you do survey, and what cameras you include. As 89% of the entries were taken on DSLR cameras, we could google around for DSLR market share figures from 2007 onwards and do a very rough average: Canon and Nikon about 40% each, Sony 8%, Olympus and Pentax 4% each. Before hitting the reply button to contest such crude data, it is worth bearing in mind that modest variations in those figures do not actually make much difference on the analysis we are about to make.

The thing we see here is that Canon cameras success in the Dpreview challenges is more or less the same as it's market share. A similar, but slightly lower result can be found for Nikon. But the real winner in this context is Olympus; 14% of winning entries achieved with just 4% of market share. Or put another way, an Olympus photographer is nearly 4 times as likely to win a challenge as a Canon photographer. Pentax get's the runner up place; 10% of winning entries from it's own 4% share, or 2.5 times more likely to get a top three place than a Canon.

Sony does not fair so well, but to be fair to them we must remember that they are only just recovering from the dry spell that followed Sony's takeover of Minolta.

Now, let's dig deeper and look at the actual models rather than brands. More than a third of the Canon entries were taken using full frame cameras, specifically the 5D as there was only one entry taken on a 1D. Given that the 5D's sales make it a specialty camera, that is quite a significant proportion. Canon compact cameras account for just 14% of Canon shots, but that is still more than the number of Canon entries taken on the Digital Rebel line of cameras, just a tad below 10% or 4% overall. Given the Digital Rebels huge market share over recent years we might see that as a pretty poor result. The most common type of Canon camera, at 44%, was of course the midrange xxD series (10D to 50D).

Trying to break down the other bands, particularly the lesser ones, we run into problems with lack of data, but the overall trend is similar (except for full frame of course). Also with Nikon the best selling low end models (D40/50/60) are thin on the ground. The most common Olympus camera is the midrange E5xx series. Likewise the Pentax images are dominated by midrange *istD's,K10D and K20D cameras.

Perhaps the most significant thing about the results is that it is difficult to draw any hard conclusions. If we were to try and make any sense of it we would perhaps conclude that you have a reasonable chance of winning with any brand of camera, and perhaps more significantly there is no reason to regard minor brands such as Pentax or Olympus as being inferior, far from it! The relative lackluster performance of the entry level models, despite their enormous sales, could perhaps in part be explained but the large number of people who buy these cameras to use as a fast point and shoot camera, the sort of people who generally do not get involved in photo contests. On the other hand, there is no shortage of entry level DSLR users on Dpreview, so maybe the data is telling us something: if you are serious about your photography then spend that little extra to get a camera in the midrange band.

Now let's spare a thought for the compact cameras which were used for 11% of entries. That's not a high number, but not insignificant either, especially when we remember that more Canon compacts made the winners pedestal than Digital Rebels . What is significant is that they were all low pixel density models without exaggerated zoom lenses; the sort of compact camera that is unfortunately becoming increasingly rare. Maybe this sector will be taken over by the new generation of compact interchangeable lens cameras, but so far it is not happening; there was one entry for such a camera, the Olympus Pen. Of course it is early days yet...
 
Moving on from brands and cameras, let's have a look at the lenses; after all, many would argue that the lens is more important than the camera.

Twenty two percent of these winning entries were taken using prime lenses. Non of these were wide angle primes. A few were long telephotos, but the bulk of the prime lenses were in the 50mm to 100mm range; 13% were described as macro lenses but the use was by no means limited to close focus shots. This type of lens ranging from very cheap 50mm f1.8 lenses to 100mm macro lenses represent an excellent tool for the keen photographer on a budget as the intrinsic simplicity of the lenses means they can yield a quality comparable to top professional lenses at modest prices, and clearly there is no shortage of people taking advantage of this.

Most people were using lenses of the same brand as the camera but third party lenses were not excluded notching up a 13% share of which 6 were Sigma, 4 Tokina and 3 Tamron. The Sigma lenses were of various types whilst the Tokina lenses were all zooms that opened up wider than 18mm.

A curiosity: the only shot taken in micro four thirds format was done with an Olympus Pen sporting a 20mm Lumix prime. Should we call that a second party lens?

Vacation zooms were almost inexistent, just one Nikon 18-200mm and a single Pentax 18-250mm; clearly people who win contests are prepared to switch lenses to achieve their goals.

Another inexistent type of lens (zero instances) was the ultra cheap external focus 18-55mm kitlens of the type that Nikon and Canon offer on their cheapest cameras. Considering they are perhaps the most widely diffused type of lens we could perhaps draw some conclusions about what can actually be achieved with such glass (and plastic!).

Now all these observations are just scratching the surface. If we had electronic access to the complete database then we could extract far more statistics and dig much deeper. None of which resolves the first conundrum however; are we looking at the equipment that winners use, or is this the equipment that makes a photographer a winner?
 
Forgotten? I fear not. Inappropriate perhaps.

Not looking at the camera when enjoying the picture is excellent advice.

But does that mean the photographer should ignore what equipment might be best to achieve a given goal?

To the beholder the equipment is irrelevant, to the photographer it isn't.
 
TROLL ALERT TROLL ALERT TROLL ALERT.... let's get im lads
OMG Merry Christmas anywho and I've been bashed enuff so back off
--
May the light be with you
Stop global whining
Stupid should hurt
 
So in order to dig down into this data we need some figures for market share, which of course depend an awful lot on when and where you do survey, and what cameras you include. As 89% of the entries were taken on DSLR cameras, we could google around for DSLR market share figures from 2007 onwards and do a very rough average: Canon and Nikon about 40% each, Sony 8%, Olympus and Pentax 4% each. Before hitting the reply button to contest such crude data, it is worth bearing in mind that modest variations in those figures do not actually make much difference on the analysis we are about to make.
But sales is a bad way of judging this particular case. You've ignored the possibility that contributors to the DPReview challenges don't have a representative sample of cameras. Your data already shows that DSLRs have won a grossly disproportionate share of challenges relative to their total market share. It seems safe to assume that's because the challenges participants don't follow the market as a whole in their camera choices. If they don't follow the market in camera type, what's to say that they follow it in brand choice, either?

A better approach would be to compare the distribution of brands among the top entries compared to the distribution among all submissions. That would eliminate any brand bias in the overall submission pool and get you closer to your goal.

You also need to consider the possibility that there's a brand bias in voting. We know that there are some dedicated brand fanboys around here; it's possible that some of them consider brand loyalty in their voting, not just aesthetic quality. That may throw a big monkey wrench into your calculations.
--

As with all creative work, the craft must be adequate for the demands of expression. I am disturbed when I find craft relegated to inferior consideration; I believe that the euphoric involvement with subject or self is not sufficient to justify the making and display of photographic images. --Ansel Adams
 
Forgotten? I fear not. Inappropriate perhaps.

Not looking at the camera when enjoying the picture is excellent advice.

But does that mean the photographer should ignore what equipment might be best to achieve a given goal?

To the beholder the equipment is irrelevant, to the photographer it isn't.
Bla bla bla...take your ego-boosting drivel somewhere else, it's tiresome and pathetic.
 
Yes, I would have liked to compare the data against the user base of Dpreview challenge entrants, both at a global level and at an individual contest level. But as I noted, I do not have access to DPreview's database and that would make the whole thing a bit laborious.

I do however have a strong suspicion that the inadequate method of comparing against market as opposed to user base is the real reason behind the relative success of the Olympus and Pentax brands. But that is just my opinion so it is just as worthless as the flawed analysis ;-)

Regarding fanboys, I believe equipment data is hidden during voting phases.
 
I usually do. But it's raining. I enjoyed totting up the numbers, at least as a one off exercise. I'm sorry if it caused you any offense.
 
On the other hand if you take a good shot with a not very good camera, in many cases the image would look that much better had it been taken with better equipment.

Is that difference enough to give a photographer a leading edge in challenges?

As I noted, a small proportion of the winning entries were taken on compact cameras, but they were all low density types. Out of 100 winning entries not one was taken on a high density ( > 20MP/cm2) compact, and yet they are by far the most widely diffused type of digital camera. Is that just by chance? Or is it that the good photographers who take the winning entries just happen to prefer to use better equipment?
 
On the other hand if you take a good shot with a not very good camera, in many cases the image would look that much better had it been taken with better equipment.

Is that difference enough to give a photographer a leading edge in challenges?

As I noted, a small proportion of the winning entries were taken on compact cameras, but they were all low density types. Out of 100 winning entries not one was taken on a high density ( > 20MP/cm2) compact, and yet they are by far the most widely diffused type of digital camera. Is that just by chance? Or is it that the good photographers who take the winning entries just happen to prefer to use better equipment?
I think the good photographers just win more often, and they tend to use better gear, because most good photographers are into it enough that they crave better gear.
 
..for DPR to promote different cameras that will drive traffic to amazon.com for purchasers. DPR is owned by amazon.
 
I can't believe all the negative comments and troll acusations regarding the OP! As far as I can tell, his crude statistics and speculations are pretty much harmless. What is the big problem?
 
You have entirely too much time on your hands, go grap your instamatic and take some pictures, Oh yes,
Merry Christmas
 

Keyboard shortcuts

Back
Top