Is 70ms 14 or 12% faster than 80ms ?

classic Classic list List threaded Threaded
14 messages Options
Reply | Threaded
Open this post in threaded view
|

Is 70ms 14 or 12% faster than 80ms ?

Albert Chan
Is 70ms 14 or 12% faster than 80ms ?

Reading many stackoverflow response, I am still confused.
Most answers claimed (1 - 70/80), or 12.5% faster

If the math is right, it is impossible to be 100% faster.
10 times faster (1000%) is meaningless (time is negative)

I normally avoid using "percentage faster", and prefer ratio:
Speed at 80/70 = 1.14X

--

https://stackoverflow.com/questions/31506554/is-70-ms-14-or-12-faster-than-80-ms?noredirect=1&lq=1

https://stackoverflow.com/questions/8127862/how-do-you-calculate-how-much-faster-time-x-is-from-time-y-in-terms-of

https://math.stackexchange.com/questions/716767/how-to-calculate-the-percentage-of-increase-decrease-with-negative-numbers/716770#716770



Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Oliver Kroth


Am 15.05.2018 um 13:19 schrieb Albert Chan:

> Is 70ms 14 or 12% faster than 80ms ?
>
> Reading many stackoverflow response, I am still confused.
> Most answers claimed (1 - 70/80), or 12.5% faster
>
> If the math is right, it is impossible to be 100% faster.
> 10 times faster (1000%) is meaningless (time is negative)
>
> I normally avoid using "percentage faster", and prefer ratio:
> Speed at 80/70 = 1.14X
>
>
it takes 12.5% less time
that makes it 14.28...% faster (1 / (1-0.125))

--
Oliver

Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Alysson Cunha
In reply to this post by Albert Chan
The term "fast" refers to "speed". To express speed, you need to specify amount and time.

If a process takes 70ms to complete, then it's speed is 14,29 times per second (1 / 0,07)

If a process takes 80ms to complete, then it's speed is 12,5 times per second.

so, the speed of 70ms is (14,29/12,5) times faster than 80ms's speed = 14,289 % faster....

On Tue, May 15, 2018 at 8:19 AM, Albert Chan <[hidden email]> wrote:
Is 70ms 14 or 12% faster than 80ms ?

Reading many stackoverflow response, I am still confused.
Most answers claimed (1 - 70/80), or 12.5% faster

If the math is right, it is impossible to be 100% faster.
10 times faster (1000%) is meaningless (time is negative)

I normally avoid using "percentage faster", and prefer ratio:
Speed at 80/70 = 1.14X

--

https://stackoverflow.com/questions/31506554/is-70-ms-14-or-12-faster-than-80-ms?noredirect=1&lq=1

https://stackoverflow.com/questions/8127862/how-do-you-calculate-how-much-faster-time-x-is-from-time-y-in-terms-of

https://math.stackexchange.com/questions/716767/how-to-calculate-the-percentage-of-increase-decrease-with-negative-numbers/716770#716770






--
Alysson Cunha / AlyssonRPG
http://www.rrpg.com.br - Jogue o tradicional RPG de mesa online
http://www.alyssoncunha.com.br - Me conheça melhor
Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Viacheslav Usov
In reply to this post by Albert Chan
On Tue, May 15, 2018 at 1:19 PM, Albert Chan <[hidden email]> wrote:

> Is 70ms 14 or 12% faster than 80ms ?

Times: t vs T, t = (1 - s)T, meaning t is s x 100% shorter than T.

Speeds: v = 1/t vs V = 1/T, v = (1/(1 - s)) V = (1 + f)V, meaning v is f x 100% faster than V.

Obviously 1/(1 - s) = (1 + f), so f = s / (1 - s), and s = f / (1 + f).

With t = 70 and T = 80, s = 0.125, f = 0.143.

> If the math is right, it is impossible to be 100% faster.

Given f = 1 and T = 80, s = 0.5, t = 40.

Cheers,
V.
Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Albert Chan
In reply to this post by Oliver Kroth
>> Is 70ms 14 or 12% faster than 80ms ?
>>
>> Reading many stackoverflow response, I am still confused.
>> Most answers claimed (1 - 70/80), or 12.5% faster
>>
>> --Albert
>
> it takes 12.5% less time
> that makes it 14.28...% faster (1 / (1-0.125))
>
> --
> Oliver

Thanks.  Stack-overflow is wrong !

-- %faster should be relative to speed

speed = 1/70= (100% + %faster)(1/80)
%faster = 80/70 - 100% = 14.3%

-- %slower should also be relative to speed
-- Example: 70ms job now take 80ms

speed = 1/80 = (100% - %slower)(1/70)
%slower = 100% - 70/80 = 12.5%

--> %slower will NEVER reach 100%



Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Albert Chan
> speed = 1/70= (100% + %faster)(1/80)
> %faster = 80/70 - 100% = 14.3%

Just want to add:  
% faster should only be use for small percentage
Bruce Dawson recommend around 20%

https://randomascii.wordpress.com/2018/02/04/what-we-talk-about-when-we-talk-about-performance/

We don't say 100% faster, we say twice as fast
-- the number is bigger, good for marketing
-- using speed ratio is unambiguous (faster if ratio > 1)
-- work well even if code run slower, say 0.96X
-- trivial to calculate

speed vs oldest version = 1.143 * 0.96 = 1.097X





Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Albert Chan
If what everyone here say is true, "Lua Programming Gems" is wrong
(unless it is trying to *redefine* the meaning of % faster)

Quote from chapter 2, Lua Performance tips :
https://www.lua.org/gems/sample.pdf

When I say that a program is "X% times faster" than another it means
that it runs in % less time. (A program 100% faster would take no time
to run).  When I say that a program is "X% times slower" than another
I mean that the other is % faster.  (A program 50% slower means that
it takes twice the time).

--

I ALWAYS use speed ratio, say 1.143X, to avoid this ambiguity

Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Dirk Laurie-2
2018-05-15 23:20 GMT+02:00 Albert Chan <[hidden email]>:

> If what everyone here say is true, "Lua Programming Gems" is wrong
> (unless it is trying to *redefine* the meaning of % faster)
>
> Quote from chapter 2, Lua Performance tips :
> https://www.lua.org/gems/sample.pdf
>
> When I say that a program is "X% times faster" than another it means
> that it runs in % less time. (A program 100% faster would take no time
> to run).  When I say that a program is "X% times slower" than another
> I mean that the other is % faster.  (A program 50% slower means that
> it takes twice the time).

That tells you precisely how the the terms "faster" and "slower" are
used _in that article_, in order to avoid misunderstanding by readers
whose usage may be different.

It is exactly the same kind of remark as:

> I ALWAYS use speed ratio, say 1.143X, to avoid this ambiguity

except that it does not confidently claim "ALWAYS".

Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Tim Hill


On May 15, 2018, at 11:12 PM, Dirk Laurie <[hidden email]> wrote:

When I say that a program is "X% times faster" than another it means
that it runs in % less time. (A program 100% faster would take no time
to run).  When I say that a program is "X% times slower" than another
I mean that the other is % faster.  (A program 50% slower means that
it takes twice the time).

That tells you precisely how the the terms "faster" and "slower" are
used _in that article_, in order to avoid misunderstanding by readers
whose usage may be different.

… except he *defines* the phrase “X% times faster” but then *uses* the phrase “100% faster” and not “100% times faster”.

I’m not entirely sure “X% times faster” is a good choice. What is “10% times faster” supposed to mean? It certainly *doesnt* mean “10% faster”, nor does it mean “1/10th times faster”.

—Tim

Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Dirk Laurie-2
2018-05-16 9:34 GMT+02:00 Tim Hill <[hidden email]>:

>
>
> On May 15, 2018, at 11:12 PM, Dirk Laurie <[hidden email]> wrote:
>
> When I say that a program is "X% times faster" than another it means
> that it runs in % less time. (A program 100% faster would take no time
> to run).  When I say that a program is "X% times slower" than another
> I mean that the other is % faster.  (A program 50% slower means that
> it takes twice the time).
>
>
> That tells you precisely how the the terms "faster" and "slower" are
> used _in that article_, in order to avoid misunderstanding by readers
> whose usage may be different.
>
>
> … except he *defines* the phrase “X% times faster” but then *uses* the
> phrase “100% faster” and not “100% times faster”.
>
> I’m not entirely sure “X% times faster” is a good choice. What is “10% times
> faster” supposed to mean? It certainly *doesnt* mean “10% faster”, nor does
> it mean “1/10th times faster”.

Aha. Humpty Dumpty (the famous quote on what words mean)  is not
enough for this discussion. We need the White Knight too (the nice
distinctions on what the song is, what it is called, what its title is,
and what the title of the song is called).

Moreover, didn't Lewis Carroll in his alter ego as C.L.Dodgson, that
eristic algebraist and logician, say something about precisely the
topic of "times faster" and "times as fast as"?

I'll have to re-read Curiousa Mathematica :-)

Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Albert Chan
In reply to this post by Dirk Laurie-2
On May 16, 2018, at 2:12 AM, Dirk Laurie <[hidden email]> wrote:

2018-05-15 23:20 GMT+02:00 Albert Chan <[hidden email]>:
If what everyone here say is true, "Lua Programming Gems" is wrong
(unless it is trying to *redefine* the meaning of % faster)


That tells you precisely how the the terms "faster" and "slower" are
used _in that article_, in order to avoid misunderstanding by readers
whose usage may be different.


The article remove %faster ambiguity, but still is misleading.
Readers will not know it can only be used _in that article_.

I thought it were just a *reminder* of terms, not a re-definition.
A better reminder is to say X% faster meant X% faster in speed.

Another case of "faster" re-definition:

https://www.livescience.com/27920-quantum-action-faster-than-light.html

Spooky! Quantum Action Is 10,000 Times Faster Than Light

How fast do quantum interactions happen? 
Faster than light, 10,000 times faster

It is exactly the same kind of remark as:

I ALWAYS use speed ratio, say 1.143X, to avoid this ambiguity

except that it does not confidently claim "ALWAYS".

But I DO !

Speed ratio is much more precise than "% faster"
(Besides, I were un-sure what "% faster" meant ...)



Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Albert Chan
In reply to this post by Tim Hill

On May 15, 2018, at 11:12 PM, Dirk Laurie <[hidden email]> wrote:

When I say that a program is "X% times faster" than another it means
that it runs in % less time. (A program 100% faster would take no time
to run).  When I say that a program is "X% times slower" than another
I mean that the other is % faster.  (A program 50% slower means that
it takes twice the time).

That tells you precisely how the the terms "faster" and "slower" are
used _in that article_, in order to avoid misunderstanding by readers
whose usage may be different.

… except he *defines* the phrase “X% times faster” but then *uses* the phrase “100% faster” and not “100% times faster”.

I’m not entirely sure “X% times faster” is a good choice. What is “10% times faster” supposed to mean? It certainly *doesnt* mean “10% faster”, nor does it mean “1/10th times faster”.

—Tim


Good catch !
Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Tim Hill
In reply to this post by Dirk Laurie-2

On May 16, 2018, at 3:39 AM, Dirk Laurie <[hidden email]> wrote:


… except he *defines* the phrase “X% times faster” but then *uses* the
phrase “100% faster” and not “100% times faster”.

I’m not entirely sure “X% times faster” is a good choice. What is “10% times
faster” supposed to mean? It certainly *doesnt* mean “10% faster”, nor does
it mean “1/10th times faster”.

Aha. Humpty Dumpty (the famous quote on what words mean)  is not
enough for this discussion. We need the White Knight too (the nice
distinctions on what the song is, what it is called, what its title is,
and what the title of the song is called).

Moreover, didn't Lewis Carroll in his alter ego as C.L.Dodgson, that
eristic algebraist and logician, say something about precisely the
topic of "times faster" and "times as fast as"?


Indeed, as Carroll was pointing out, English is a very ambiguous language, hence the need for precision and clarity in definitions. I’m not sure the “X% times faster” definition would qualify for either in this case.

“X% times faster” is awfully close to the mis-use of that horrible phrase “rate of speed”, which is erroneously used (particularly by law enforcement in the US for some reason) as a more refined way to say “speed” when in fact it means acceleration (“speed” being “rate of change of position” and “rate of speed” thus either being tautology or “rate of rate of change of position”).

—Tim


Reply | Threaded
Open this post in threaded view
|

Re: Is 70ms 14 or 12% faster than 80ms ?

Albert Chan
> “X% times faster” is awfully close to the mis-use of that horrible phrase “rate of speed”, which is erroneously used (particularly by law enforcement in the US for some reason) as a more refined way to say “speed” when in fact it means acceleration (“speed” being “rate of change of position” and “rate of speed” thus either being tautology or “rate of rate of change of position”).
>
> —Tim

Someone did a poll on % faster/slower questions
45% got it right (37 votes)
37% got it wrong (31 votes)
18% just gives up (15 votes)

https://arstechnica.com/civis/viewtopic.php?f=23&t=1110666