Freezing Saddles Scoreboard backend "soft launch"

[Mikey]

@dbb 75087 wrote:

I am completely cool with leaving the logic as is. My obligatory rides were, according to my Garmin, 2.1 miles to eliminate any rounding issues.

The really funny part is that I’ve been entering my distances manually.

I guess I need to go back and re-enter it.

[Mikey]

hozn,

So if I manually entered 1.0 miles in strava, and it converted it to meters, then back to miles as 0.9999. . . then I get scored for <1.0 miles. Not cool. I have an easy fix though. I will enter 1.001 miles for that day in strava.

[hozn]

@Mikey 75091 wrote:

hozn,

So if I manually entered 1.0 miles in strava, and it converted it to meters, then back to miles as 0.9999. . . then I get scored for <1.0 miles. Not cool. I have an easy fix though. I will enter 1.001 miles for that day in strava.

Yes, that is what the math says since Strava rounds after converting the data (and we don’t).

Doing the conversions in Google illustrates what’s going on here:

1.0 miles in meters = 1609.34
1609.34 meters in miles = 0.999998

I agree; it doesn’t make a lot of sense. I’d say it’s a Strava bug, but we may want to adjust our scoring logic to round to 4 decimal places or something

And yes, entering 1.01 miles (or 1.0001) is the fix. I suppose this only affects people manually entering data, since probably those are the only people to actually log rides of exactly 1.00 miles.

[jopamora]

@Vicegrip 75086 wrote:

must have a triple on the box bike

More like a motor

[peterw_diy]

@hozn 75084 wrote:

This is hilarious! Maybe Grant can fix that one (using Strava’s auto-correct? If that’s even a thing anymore?) so it doesn’t skew the graph.

Aw, c’mon, hozn, it’s just a little more code to write. If Grant can export data from the Garmin, you just need to iterate over all the data points and query some reputable goedata source to get the surface elevation at each X,Y and fix each point’s elevation. Like, this: https://developers.google.com/maps/documentation/elevation/

Hmm, I wonder if anybody has bothered to write a cheat app that could take Garmin data and adjust all the points’ timestamps. Yes, of course someone has: http://www.digitalepo.com/ Nice name.

BTW, this is a nice geometry reminder. I tend to think of grade % as if it were an angular measure in radians, with 100% being vertical. But, no, a 100% grade is just 45 degrees / 50 radians. Strava shows Grant hitting grades as high as 513%, or 78 degrees. Cool.

[hozn]

@peterw_diy 75095 wrote:

Aw, c’mon, hozn, it’s just a little more code to write. If Grant can export data from the Garmin, you just need to iterate over all the data points and query some reputable goedata source to get the surface elevation at each X,Y and fix each point’s elevation. Like, this: https://developers.google.com/maps/documentation/elevation/

Yes, I believe that is exactly what Strava’s “Correct” elevation feature does. I think the granularity of the sources isn’t great — on roads it’s probably fine, but off-road riders get screwed out of elevation changes — or at least, I think that’s the case. Typically I think the Garmin data is better — or at least more consistent. But obviously it can go wrong

[Mikey]

@hozn 75092 wrote:

Yes, that is what the math says since Strava rounds after converting the data (and we don’t).

Doing the conversions in Google illustrates what’s going on here:

1.0 miles in meters = 1609.34
1609.34 meters in miles = 0.999998

I agree; it doesn’t make a lot of sense. I’d say it’s a Strava bug, but we may want to adjust our scoring logic to round to 4 decimal places or something

And yes, entering 1.01 miles (or 1.0001) is the fix. I suppose this only affects people manually entering data, since probably those are the only people to actually log rides of exactly 1.00 miles.

fixed

I was wondering why I never showed up with a sleaze ride. D’oh

[ronwalf]

@hozn 75092 wrote:

I agree; it doesn’t make a lot of sense. I’d say it’s a Strava bug, but we may want to adjust our scoring logic to round to 4 decimal places or something

An easy fix would be to add 1 meter to the strava distance before converting. That should fix any rounding errors, and shouldn’t change the point total otherwise.

[Anonymous]

@hozn 75092 wrote:

I suppose this only affects people manually entering data, since probably those are the only people to actually log rides of exactly 1.00 miles.

This is not necessarily true. My grocery store is one mile round trip from my house, give or take a zig-zag through the parking lot. During last year’s freezing saddles, one day I rode to the grocery store with the Strava app recording my ride, checked my mileage before turning it off and going inside, noted that Strava said I had ridden 1.0 miles, put away my groceries, etc. However, what Strava actually recorded on my ride was 0.99 miles* I did not get my 10 points until I rode somewhere else later that day. So even if you are using Strava’s own app, it is a good idea to make sure you ride at least 1.1 miles, just to be sure.

*well, I mean 0.99 miles is what was displayed on the ride info looking at the web site, not talking about the precision of the actual data that is stored, which I don’t know. the ride info is displayed on the web to 2 decimal places but only displayed on the app while recording to one decimal place, which was rounded to 1.0

[jrenaut]

The only real solution is to ride more.

[Rod Smith]

@dcv 75046 wrote:

Dasgeh, whats going on here?

E-assist. :rolleyes:

[Rod Smith]

@jrenaut 75116 wrote:

The only real solution is to ride more.

Which she did.

[Rod Smith]

@dbb 75087 wrote:

I am completely cool with leaving the logic as is. My obligatory rides were, according to my Garmin, 2.1 miles to eliminate any rounding issues.

That’s not gonna help your team in the sleaze ride competition. :confused:

[dasgeh]

@hozn 75092 wrote:

Yes, that is what the math says since Strava rounds after converting the data (and we don’t).

Doing the conversions in Google illustrates what’s going on here:

1.0 miles in meters = 1609.34
1609.34 meters in miles = 0.999998

I agree; it doesn’t make a lot of sense. I’d say it’s a Strava bug, but we may want to adjust our scoring logic to round to 4 decimal places or something

And yes, entering 1.01 miles (or 1.0001) is the fix. I suppose this only affects people manually entering data, since probably those are the only people to actually log rides of exactly 1.00 miles.

I actually did a sleaze ride yesterday of exactly 1 mile. And I think it’s not counting as a “day”

[hozn]

I am going to fix the code so that we round with the same precision as Strava when storing.

[Author]

Posts