I'm writing a book about Bayesian statistics, decision theory, information theory, and connected topics. It's coming along slowly, but I thought I'd share a bit of a preview. This is an early draft of the first 11 pages, including the title, a table of contents, and a few pages of writing. Much will change I'm sure. Any purchases will be considered encouragement for me to actually finish.
I'm writing a book about Bayesian statistics, decision theory, information theory, and connected topics. It's coming along slowly, but I thought I'd share a bit of a preview. This is an early draft of the first 11 pages, including the title, a table of contents, and a few pages of writing. Much will change I'm sure. Any purchases will be considered encouragement for me to actually finish.
I'm writing a book about Bayesian statistics, decision theory, information theory, and connected topics. It's coming along slowly, but I thought I'd share a bit of a preview. This is an early draft of the first 11 pages, including the title, a table of contents, and a few pages of writing. Much will change I'm sure. Any purchases will be considered encouragement for me to actually finish.
We took this video of our daughter playing with a block when she was 6 months old, and were amused to see that it looks like she briefly performs a magic trick with it.
DISCLAIMER: This information is provided for entertainment purposes only.
NOTE: Tips have been very irregular lately and my forecasting assumptions assume they are likely to be regular, so the forecasts should be taken with a *large* grain of salt, especially for channels with short histories. However, they still might be of interest, aong with the historical LBC numbers.
I used [TipPredict](https://github.com/eggplantbren/TipPredict) to forecast LBC tips received over the next month for some publishers on LBRY. The forecast period is always one month from the day this file is updated.
This is a CSV file containing a low, a medium, and a high prediction for each of the creators on the list. I made quite basic "status quo" assumptions, which assume that the next month is quite likely to look similar to previous months apart from typical statistical fluctuations. If people find this service useful, I can make the assumptions much more realistic and sophisticated, including things like trends, spikes when you publish, and the fact that what's observed for other channels has some bearing on what you might predict for a particular channel.
So far the list of channels I've analysed is quite short (but it's longer than what you see in the thumbnail). If you'd like to be added to the list or removed from it the next time I update this file, contact me on Keybase or the LBRY Discord.
Formally, conditional on my assumptions, the probability of the next month's tips being between the low and high forecast is 90%, and it's 50/50 whether it's above or below the medium forecast. Some creators have more uncertain forecasts than others. This is due to two main effects: (i) a shorter tip history, so there's less data to extrapolate from, and/or (ii) greater variation in the size of tips received in the past, which propagates through to greater uncertainty in the size of future tips received.
DISCLAIMER: This information is provided for entertainment purposes only.
NOTE: Since 2019-04-20, when I switched to using chainquery to fetch the data, the results now include withdrawn tips.
I used [TipPredict](https://github.com/eggplantbren/TipPredict) to forecast LBC tips received over the next month for some publishers on LBRY. The forecast period is always one month from the day this file is updated.
This is a CSV file containing a low, a medium, and a high prediction for each of the creators on the list. I made quite basic "status quo" assumptions, which assume that the next month is quite likely to look similar to previous months apart from typical statistical fluctuations. If people find this service useful, I can make the assumptions much more realistic and sophisticated, including things like trends, spikes when you publish, and the fact that what's observed for other channels has some bearing on what you might predict for a particular channel.
So far the list of channels I've analysed is quite short (but it's longer than what you see in the thumbnail). If you'd like to be added to the list or removed from it the next time I update this file, contact me on Keybase or the LBRY Discord.
Formally, conditional on my assumptions, the probability of the next month's tips being between the low and high forecast is 90%, and it's 50/50 whether it's above or below the medium forecast. Some creators have more uncertain forecasts than others. This is due to two main effects: (i) a shorter tip history, so there's less data to extrapolate from, and/or (ii) greater variation in the size of tips received in the past, which propagates through to greater uncertainty in the size of future tips received.
DISCLAIMER: This information is provided for entertainment purposes only. For instance, prior to 2019-04-10, there was a very dumb bug which made the numbers all wrong. Hey, at least *that* bug is gone now.
I used TipPredict (https://github.com/eggplantbren/TipPredict) to forecast LBC tips received over the next month for some publishers on LBRY. This is a CSV file containing a low, a medium, and a high prediction for each of the creators on the list. The assumptions I used were quite basic "status quo" type assumptions, which assume that the next month is quite likely to look similar to previous months. If people find this service useful, I can make the assumptions much more realistic and sophisticated, including things like trends, and spikes when you publish.
So far the list of channels I've analysed is quite short (but it's longer than what you see in the thumbnail). If you'd like to be added to the list or removed from it the next time I update this file, contact me on Keybase or the LBRY Discord.
Formally, conditional on my assumptions, the probability of the next month's tips being between the low and high forecast is 90%, and it's 50/50 whether it's above or below the medium forecast. Some creators have more uncertain forecasts than others. This is due to two main effects: (i) a shorter tip history, so there's less data to extrapolate from, and/or (ii) greater variation in the size of tips received in the past, which propagates through to greater uncertainty in the size of future tips received.