🐷 Ethan Oliver Ralph / TheRalphRetort / Rad Roberts / Jcaesar187 / Rage Pig / "Killstream" / "Tequila Sunrise" - 5'1'' fat alcoholic, owner of a gunt, convicted felon and revenge pornographer, property of the ugly failed tranny pornstar Lucas Roberts. Has quadruple titties.

Status
Not open for further replies.
anuscabbage12 said:
what would also be interesting is doing lemons/total stream length, as the more recent streams he has tried to stretch them out to squeeze as many lemons as possible out of his dwindling audience. The stream today was over 4 hours. It may show a much clearer downward trend
Click to expand...
Yeah that's going to be harder to get for past streams. I can start tracking that from now on.
 
  • Like
  • Feels
Reactions: Vanquished Phoenix, REGENDarySumanai, Ulamog did nothing wrong and 4 others
Sam Losco said:
Here is a handy chart of the amount of lemons he's gotten since I've been sniping him. These aren't exact numbers but rounded, down I assume, to the nearest tenth as Dlive only shows two decimal places, so that's why they all end in 0. The 35k was the flat Earther stream. There was one stream early on I didn't finish so that one isn't included. His 3 day suspension is the end of Sept, so between the 15910 and 18820 points.
I wish I was smart enough to figure out how to easily export the earnings data from the site. I tried messing around with their API but it's undocumented so I really didn't get anywhere beyond the few examples they have.

View attachment 1665603
Click to expand...
Neat. Since you've been sniping him is interesting, but can you also give a longer timescale version? At a glance the most interesting thing about this graph is inconsistency. He's getting wildly different amounts night to night. That trendline is a vague mathematical construction which isn't intuitively visible. It's also kinda interesting that it's very consistently hit, miss, hit, miss, rather than getting streaks and runs. The day after every good stream is a bad stream almost every time. When it's not like that, it's the same basic concept except on a 2 day or 3 day scale instead of day to day. He's very consistently peaking and troughing. Weird.
I wish I knew anything about statistics and knew what any of that meant.
 
  • Like
Reactions: Vanquished Phoenix
anuscabbage12 said:
it might be too time consuming, but doesnt the chillstream channel have archives of the past streams? also your own?
Click to expand...
Unreliable. Chillstream has had tech issues and possible editing. I have had past tech issues cutting a stream short.

So what I did was go and get the lengths of the show from what he uplaods on libsyn. He tends to only play the first outro song for those so they are several minutes shorter than what is on Dlive. But whatever, close enough, he doesn't usually get lemons at that point anyway. It at least allows me to get duration for all the streams I have lemon counts for.
Screenshot from 2020-10-15 23-53-32.png

Erischan said:
Neat. Since you've been sniping him is interesting, but can you also give a longer timescale version? At a glance the most interesting thing about this graph is inconsistency. He's getting wildly different amounts night to night. That trendline is a vague mathematical construction which isn't intuitively visible. It's also kinda interesting that it's very consistently hit, miss, hit, miss, rather than getting streaks and runs. The day after every good stream is a bad stream almost every time. When it's not like that, it's the same basic concept except on a 2 day or 3 day scale instead of day to day. He's very consistently peaking and troughing. Weird.
I wish I knew anything about statistics and knew what any of that meant.
Click to expand...
I can't unless you know of a way for me to get lemon counts from streams in the past. Dlive doesn't have that data, Chillstream doesn't show lemons. I could try and dig through PO's restreams but that may be inconsistent and unreliable and I wouldn't stand by that data.

Edit - Since text is overlapping in the chart and not easily viewable, here is the link to the sheet I'm doing it in: https://docs.google.com/spreadsheets/d/1s7pJW1SkIid78uT366VeAVsJD0zg-4sHnMxcm60Zew8/edit?usp=sharing
 
  • Informative
  • Winner
Reactions: Vanquished Phoenix, Iroha Isshiki, Freedom Fries and 5 others
Sam Losco said:
Unreliable. Chillstream has had tech issues and possible editing. I have had past tech issues cutting a stream short.

So what I did was go and get the lengths of the show from what he uplaods on libsyn. He tends to only play the first outro song for those so they are several minutes shorter than what is on Dlive. But whatever, close enough, he doesn't usually get lemons at that point anyway. It at least allows me to get duration for all the streams I have lemon counts for.
Click to expand...
Yeah there is obviously a correlation between the two. If he had only gone 2-3 hours today, he would have made less than 10k. It did not have an effect on the flat earth episode. but it seems to control for the most of the other spikes other than 10/02, which I guess is the richard spencer vs styx show.

Maybe it was a new audience that day, for the 35k spike (flat earthers fans)? or just the retardation of the topic? Or people argued and got heated?

The effect may be small, but doing it by total time, probably would give a slight more downward trend. It seem that takening out outliers like the flat earth/spencer vs styx/election debate ie polticial events/ones offs, he makes a alot. But his regular shows and audience/earning follow a much more distinct downward trend
 
Sam Losco said:
Unreliable. Chillstream has had tech issues and possible editing. I have had past tech issues cutting a stream short.

So what I did was go and get the lengths of the show from what he uplaods on libsyn. He tends to only play the first outro song for those so they are several minutes shorter than what is on Dlive. But whatever, close enough, he doesn't usually get lemons at that point anyway. It at least allows me to get duration for all the streams I have lemon counts for.
View attachment 1665664

I can't unless you know of a way for me to get lemon counts from streams in the past. Dlive doesn't have that data, Chillstream doesn't show lemons. I could try and dig through PO's restreams but that may be inconsistent and unreliable and I wouldn't stand by that data.

Edit - Since text is overlapping in the chart and not easily viewable, here is the link to the sheet I'm doing it in: https://docs.google.com/spreadsheets/d/1s7pJW1SkIid78uT366VeAVsJD0zg-4sHnMxcm60Zew8/edit?usp=sharing
Click to expand...
You could put this graph in a statistics textbook with the title "An easily recognizable correlation." Mystery of the peaks and troughs has been solved.
 
  • Like
  • Agree
Reactions: Iroha Isshiki and BOLDYSPICY!
anuscabbage12 said:
Yeah there is obviously a correlation between the two. If he had only gone 2-3 hours today, he would have made less than 10k. It did not have an effect on the flat earth episode. but it seems to control for the most of the other spikes other than 10/02, which I guess is the richard spencer vs styx show.

Maybe it was a new audience that day, for the 35k spike (flat earthers fans)? or just the retardation of the topic? Or people argued and got heated?

The effect may be small, but doing it by total time, probably would give a slight more downward trend
Click to expand...
Yeah it was the retardation of the guest had people donating a lot to fuck with him.

There is a correlation, but an inconsistant one for sure. The noteable exceptions are explained by content. The aforementioned 35k for the flat earther stream. The 18k the day he returned to Dlive from his suspension is another exception that can be explain by content/circumstances. The first two streams, the only other two where he seemed to earn more than the stream length would indicate, the first one was 9/11 with Ryan Dawson, and the next day Richard Spencer. So again, likely explained by content.
 
  • Like
Reactions: Vanquished Phoenix
Sam Losco said:
The noteable exceptions are explained by content.
Click to expand...
Yeah it maybe that his regular shows are just going to trash, which may be more borne out by the data and tracking the total time (he attempts to extend the show to get lemons) and he'll have to rely on his random content shows of 9/11 events, big guests/ibs scraps and other political events like debates and election coverage
 
Erischan said:
Neat. Since you've been sniping him is interesting, but can you also give a longer timescale version? At a glance the most interesting thing about this graph is inconsistency. He's getting wildly different amounts night to night. That trendline is a vague mathematical construction which isn't intuitively visible. It's also kinda interesting that it's very consistently hit, miss, hit, miss, rather than getting streaks and runs. The day after every good stream is a bad stream almost every time. When it's not like that, it's the same basic concept except on a 2 day or 3 day scale instead of day to day. He's very consistently peaking and troughing. Weird.
I wish I knew anything about statistics and knew what any of that meant.
Click to expand...

Data that varies with time is called a "time series". Time series generally have a few different components, the most important of which are the "trend" component and the "seasonal" component. "Seasonal" doesn't necessarily refer to literal seasons, though this is the origin of the terminology...it could just as easily be the day of the week or the day of the month, in this case reflecting things like the work schedules or pay periods of Ethan's most prolific paypigs. Through a process called "time series decomposition", one can heavily mitigate the distracting influence of seasonal variation to isolate the underlying trend that is actually changing over time. An example of what the various components might look like is this (the raw data being labeled "observed" and the error/noise term being labeled "residual"):

time_series_ex.png

Decomposition of time series - Wikipedia

en.wikipedia.org en.wikipedia.org

I don't think we have enough data to perform this process with any real fidelity, unfortunately, and by the time we do, the Retort will probably have been aborted, making doing so redundant anyway.

What would interest me more is whether or not there's an association between the show's topics and Ethan's profits (obviously this is true for very high profile guests like Styx or extremely controversial topics like the Flat Earth guy, but is it true for the general dreg shows we see most nights?) This could potentially be done by running the text of the recaps through an NLP model like bag-of-words and checking for any associations between the topics and Ethan's lemons but that's outside of my expertise.
 
  • Informative
Reactions: Vanquished Phoenix, IAmNotAlpharius, johnisdeer and 5 others
Haru Okumura said:
What would interest me more is whether or not there's an association between the show's topics and Ethan's profits (obviously this is true for very high profile guests like Styx or extremely controversial topics like the Flat Earth guy, but is it true for the general dreg shows we see most nights?) This could potentially be done by running the text of the recaps through an NLP model like bag-of-words and checking for any associations between the topics and Ethan's lemons but that's outside of my expertise.
Click to expand...
or just remove the outliers and track his regular show, get the trend line and take out the data points +/- 2sd for example
 
  • Thunk-Provoking
Reactions: IAmNotAlpharius
Sam Losco said:
Ok, I think I got this right. I only counted it as a show with a guest if they were named in the title, with one exception where Ashton Birdie shows up out of nowhere. All the others should be a planned guest appearance.
Click to expand...
wow. he has alot of guests lol. without events like presidential town halls, hes had 5 non guest shows in about a month. in the without guest obviously, total time is directly related to lemon count, excluding major events. with guests its more dependent on the guest, but the correlation does exist between time and lemons

we may actually be helping ralph get more money here. he should be the one doing this on his own time lol. Ralph just extend the length of your show and you'll make more money. do a 10 hour stream

the overall trend in the with guests is more a crescent, maybe indicating a bottoming out/fluctuation around a bottom mean. Where as the without guests its just straight down without the major events
 
Last edited:
  • Informative
Reactions: Vanquished Phoenix
Sam Losco said:
Ok I think I'm going to stick with this one. I switched to LibreOffice Calc and after having to figure out how to make the chart work right (I actually never did but found a work around), I have it all on one chart with no overlapping labels. As the right text says, blue stars are shows with guests.
View attachment 1665758
Click to expand...
Interesting results. The lemons seem to correlate more closely with how long the show actually runs for rather than whether or not a guest is present. With the exception of 09/25, where I assume a notable guest was present, and 09/15, where I assume an... unpopular... guest was present, the only datapoint that I find noteworthy is 10/02, where the lemon/duration trend is bucked quite substantially. I see a guest was on that show, so maybe they were drawing in more lemons, but it seems odd to me that Ralph wouldn't drag that show out if his lemons were surging like that. Did anything significant take place that night?

In any case I think graphs like this could benefit from adding another line for viewership numbers. The more data we track, the more irregularities we can spot. If we could add another one tracking Ralph's blood-alcohol level that would also be of interest but sadly we'd need the Gunt's cooperation for that one...
 
  • Agree
Reactions: A_Single_Egg
anuscabbage12 said:
what would also be interesting is doing lemons/total stream length, as the more recent streams he has tried to stretch them out to squeeze as many lemons as possible out of his dwindling audience. The stream today was over 4 hours. It may show a much clearer downward trend
Click to expand...
I noticed he even did this in the earlier years.
 
I honestly dont know what's harder to keep track of, the thread when it was blowing up or now. Ralph now seems like a pathetic hisk but even more boring because nothing of what hes actually doing is worth laughing at like the Gunt sex tape
 
  • Thunk-Provoking
Reactions: PhoBingas and Autistic
Status
Not open for further replies.
Back