fitday's credibility

[Guest guest]

Let me try to defend fitday (and the USDA database). I won't say it's

bug-free but " destroy any credibility " is way overboard.

1) The 4-4-9 computation that are you are using is a simplification of

several things. I'm no expert at this stuff but I can give some

examples. For example, the definition of " fat " is a complex thing and

the fat content used in the 4-4-9 computation is different than the fat

amount in the USDA (and thus fitday) tables. Another example is fiber.

Fitday doesn't distinguish between soluable and insoluable fiber but

the distinction is different when calculating calories. Fortunately,

fitday just uses the calories that come right out of the USDA database.

So as far as calories are concerned, fitday gets the right answers.

2) I don't know what fitday is doing with the vitamins and minerals but

I would be careful concluding that it's doing the wrong thing. Some

nutrients cannot simply be multiplied by 10 just because the food

quantity was. Also, the initial table in your example is based on a

2000c diet while the generated report is based on your age and sex. So

you shouldn't compare them. In any case, fitday doesn't even take into

account your weight - but I don't think there are general studies that

do, so it's not clear what IS the right computation! I do agree that

it would be helpful if fitday (or whatever software package you're

using) explained the calculations it was doing and why.

Bottom line - no matter what logging method you use, I would take its

vitamin/nutrition analysis with a very large grain of salt. Any

results are apt to be crude (and even more so when you factor in

variability in food samples) and you should keep that in mind.

However, calorie computation is a relatively simple thing and I have no

problem with the way fitday does it.

Don

On Saturday, May 8, 2004, at 11:55 AM, Rodney wrote:

> Hi folks:

>

> I have previously pointed out errors in the data sources we use to

> try to keep track of our nutrient intakes. This latest one looks

> like by far the worst I have tripped over so far. It looks so awful

> I now wonder if it may be dangerous to believe any of the results we

> get from these analyses.

>

> If you use Fitday.com to analyze your nutrient intake the following

> exercise should completely destroy any credibility you might have had

> for the output it gives.

>

> Try this for the food I did it for this morning (but similar types of

> errors have appeared in every one of the half dozen foods I have

> tested so far):

>

> Log on and hit the 'foods' button to enter a new food [it is

> important that it be the first food you enter for the day, as you

> will see later].

>

> Enter 'rice bran' in the box and select the 'rice bran' item from the

> listing you get. Adjust the amount of the food to a large amount

> (when using the data for analyzing a food I often adjust to 1700

> calories for reasons previously explained).

>

> Make a written note of the nutrient contents detailed in the table

> for future reference.

>

> Now click 'add to foods eaten'. Make sure it is the only food so far

> listed for the day. It should show the amount to be 1700 calories

> (if that was the amount you chose). If there are other foods in the

> box you have to delete them if you wish to continue with this

> exercise. Hit the 'save changes' button.

>

> Now hit the 'reports' button. Then hit 'nutrition today'. Now

> compare the data it shows you for micronutients contained in this

> 1700 calories worth of food with the data it previously gave you for

> the exact same amount of the exact same food. They are almost all

> completely different. Ignoring the zeros only two of the sixteen are

> the same. One, vitamin E, it is 347% greater. Iron is 80% greater.

> Some are less.

>

> Next take the protein and carbs (minus fiber) and multiply by 4 and

> add the fat x 9 and see how many calories you get. You get

> 1912 ......... 12.5% more than the 1700 it had previously said you

> had. Try the same calculation for what bran ........... the number

> of calories is 14% LESS. Try it for oat bran .......... the number

> it gives is 39.8% more. These differences are way too big to be

> explained by rounding errors or whatever.

>

> Sorry folks. Anyone who thinks the fitday results mean anything

> better think again. And since these things are all interlinked via

> the USDA database it may be unwise to believe that the other similar

> software is any better.

>

> I had thought I had a deficiency of Ca and Zn. Who knows. Given

> this fiasco, I could easily believe I have an excess.

>

> I could email fitday about it. But they claim to have more than

> three-quarters of a million people using their system and they did

> not reply to a previous email of mine on similar issues. So I don't

> think I will waste my time doing that. Nor waste any more of my time

> using their system - nor anyone else's system if I cannot test it

> extensively before having to pay for it.

>

> Have a nice day.

>

> Rodney.

>

>

>

>

>

[Guest guest]

Hi Don:

Thanks for that. But let me respond by asking for you to explain to

me how it could possibly make sense for fitday, in the example I

noted in my post - in considerable detail giving the food product and

the amount I quoted - to provide two numbers, for the vitamin E

content of EXACTLY the same quantity of EXACTLY the same food (rice

bran) in response to a SINGLE enquiry from a SINGLE individual, at

the SAME time of day on the SAME date .............. one of which

is more than 300% greater than the other? (As I noted, apart from

those that indicated zero, only two of the (sixteen) micronutrient

numbers were the same, or for that matter even close. The folic acid

and Zn data were the ones that were the same.)

Incidentally, I just repeated the request and got the same answer I

had previously.

If you want to blame this on the USDA database then you must be of

the opinion that it gives two quite different numbers for the

nutrient content of same quantity of the same food. Do you believe

this is the case? If that is so it does not make logical sense and I

have not seen evidence of it.

[i do not have a lot of confidence in the USDA database either,

incidentally. As I have posted here previously it apparently

believes that potatoes float - indeed do so quite high in the water.

Their data appear to say they believe 118 mls of potato weighs 78

grams (SG = 0.66). They did not reply to my email when I pointed

this out to them. And I have checked. The potatoes here do not

float (SG actually is about 1.06, calculated by weighing and

measuring their displacement). This is, of course, another reason to

distrust output from fitday, although for sure this would not be

Fitday's error.]

Did you take the trouble, before posting, to check for yourself the

data I specified? If you did please explain how you could think that

such discrepancies are defensible.

If you didn't, may I suggest you try it for yourself and come back

and tell us what you found and whether you still consider it

acceptable - or, more to the point, useful for our purposes of trying

to ensure we cover the micronutrient requirements while considerably

reducing total calories?

Perhaps there is an explanation. We will all (me included) be a lot

better informed if you can provide us with it.

Thanks.

Rodney.

PS: And while you are at it, perhaps you could also explain how the

approximate nature of the 4-4-9 factors could account for a 39.8%

discrepancy in total calories. Instead of 4-4-9 they would have to

be 5.6-5.6-12.6. By chance, do you work for the government? ;;

^ )))

>

> > Hi folks:

> >

> > I have previously pointed out errors in the data sources we use to

> > try to keep track of our nutrient intakes. This latest one looks

> > like by far the worst I have tripped over so far. It looks so

awful

> > I now wonder if it may be dangerous to believe any of the results

we

> > get from these analyses.

> >

> > If you use Fitday.com to analyze your nutrient intake the

following

> > exercise should completely destroy any credibility you might have

had

> > for the output it gives.

> >

> > Try this for the food I did it for this morning (but similar

types of

> > errors have appeared in every one of the half dozen foods I have

> > tested so far):

> >

> > Log on and hit the 'foods' button to enter a new food [it is

> > important that it be the first food you enter for the day, as you

> > will see later].

> >

> > Enter 'rice bran' in the box and select the 'rice bran' item from

the

> > listing you get. Adjust the amount of the food to a large amount

> > (when using the data for analyzing a food I often adjust to 1700

> > calories for reasons previously explained).

> >

> > Make a written note of the nutrient contents detailed in the table

> > for future reference.

> >

> > Now click 'add to foods eaten'. Make sure it is the only food so

far

> > listed for the day. It should show the amount to be 1700 calories

> > (if that was the amount you chose). If there are other foods in

the

> > box you have to delete them if you wish to continue with this

> > exercise. Hit the 'save changes' button.

> >

> > Now hit the 'reports' button. Then hit 'nutrition today'. Now

> > compare the data it shows you for micronutients contained in this

> > 1700 calories worth of food with the data it previously gave you

for

> > the exact same amount of the exact same food. They are almost all

> > completely different. Ignoring the zeros only two of the sixteen

are

> > the same. One, vitamin E, it is 347% greater. Iron is 80%

greater.

> > Some are less.

> >

> > Next take the protein and carbs (minus fiber) and multiply by 4

and

> > add the fat x 9 and see how many calories you get. You get

> > 1912 ......... 12.5% more than the 1700 it had previously said

you

> > had. Try the same calculation for what bran ........... the

number

> > of calories is 14% LESS. Try it for oat bran .......... the

number

> > it gives is 39.8% more. These differences are way too big to be

> > explained by rounding errors or whatever.

> >

> > Sorry folks. Anyone who thinks the fitday results mean anything

> > better think again. And since these things are all interlinked

via

> > the USDA database it may be unwise to believe that the other

similar

> > software is any better.

> >

> > I had thought I had a deficiency of Ca and Zn. Who knows. Given

> > this fiasco, I could easily believe I have an excess.

> >

> > I could email fitday about it. But they claim to have more than

> > three-quarters of a million people using their system and they did

> > not reply to a previous email of mine on similar issues. So I

don't

> > think I will waste my time doing that. Nor waste any more of my

time

> > using their system - nor anyone else's system if I cannot test it

> > extensively before having to pay for it.

> >

> > Have a nice day.

> >

> > Rodney.

> >

> >

> >

> >

> >