Discussion Forum > Shades of the Diceman
Simon:
<< for a while I actually numbered all of the unactioned items in the AF notebook, and then randomized those numbers in Excel. >>
That sounds as if it's exactly the same as the method I started with, which was to keep a numbered list in Evernote and then use the randomizer to pick a number. That does give a truly random choice in that every task has exactly the same chance of being picked.
I'm actually trying out the method you originally described because I'm interested in the effect of the rules, which gradually increase the probability of a task's being picked as it gets older.
<< You can also try this if you want a higher probability of picking some tasks. Let's say you have four tasks -- A, B, C and D, and you assign them a score between 1 and 5 based on importance, with 5 being the highest. >>
Another way of achieving the same effect would be to enter a top priority task five times on the list, a slightly less priority task four times and so on.
But one does have to remember that when one increases the chances of one task on a list you are also reducing the chances of all the others.
<< for a while I actually numbered all of the unactioned items in the AF notebook, and then randomized those numbers in Excel. >>
That sounds as if it's exactly the same as the method I started with, which was to keep a numbered list in Evernote and then use the randomizer to pick a number. That does give a truly random choice in that every task has exactly the same chance of being picked.
I'm actually trying out the method you originally described because I'm interested in the effect of the rules, which gradually increase the probability of a task's being picked as it gets older.
<< You can also try this if you want a higher probability of picking some tasks. Let's say you have four tasks -- A, B, C and D, and you assign them a score between 1 and 5 based on importance, with 5 being the highest. >>
Another way of achieving the same effect would be to enter a top priority task five times on the list, a slightly less priority task four times and so on.
But one does have to remember that when one increases the chances of one task on a list you are also reducing the chances of all the others.
January 11, 2014 at 0:43 |
Mark Forster
Mark Forster
Yes, please ignore that button.
January 11, 2014 at 6:22 |
Alan Baljeu
Alan Baljeu
I am pleased that, finally, I find others appreciating the use of randomness to choose tasks.
I have felt very alone.
My way of getting around total randomness...i.e. swaying the choices to those that have higher "importance"....is to create a new list each morning, writing down everything that is pressing on my mind. I know you can create a never-ending list, but I list the important things and then stop when I find myself getting too creative (usually after about twenty or so items). Some items can be large, some small. No matter...whatever is foremost in my mind.
I have felt very alone.
My way of getting around total randomness...i.e. swaying the choices to those that have higher "importance"....is to create a new list each morning, writing down everything that is pressing on my mind. I know you can create a never-ending list, but I list the important things and then stop when I find myself getting too creative (usually after about twenty or so items). Some items can be large, some small. No matter...whatever is foremost in my mind.
January 11, 2014 at 9:01 |
Fiona
Fiona
Thanks, Fiona. That sounds a sensible way of doing it. How has it worked for you?
January 11, 2014 at 12:00 |
Mark Forster
Mark Forster
In the Life Balance software you have a slider-wheel to determine importance of life-areas. Like a cake diagram you give out percentages to each life-area. The program factors that in, when it calculates the list.
January 11, 2014 at 15:56 |
Christopher
Christopher
Actually, Mark, using a new list every day has worked very well - particularly at my workplace. On days that I have very pressing tasks, the list ends up quite small..sometimes only six items. The urgency I feel stops me from adding extraneous tasks. When I have no must-do tasks, the list becomes longer. That is the time when the nice-to-do tasks get a look in.
January 11, 2014 at 17:56 |
Fiona
Fiona
Mark wrote:
<< I don't understand how that works. The system would certainly favour completing tasks once they've been started, but I don't see why those would necessarily be more important than the ones which haven't been started. >>
Thanks Mark. I agree with you that the main feature of my proposal is to help drive tasks to completion once they've been started. The selection for importance is weaker, but I do think it's present. Here is how I was thinking that would work:
When the dice select the item, you have a choice: take action, or delete. If you take action, work for a while, then re-enter without completion, this fact (that you thought it worthwhile enough to take action rather than deleting it) gives the task some importance.
This is based on your principle that action itself is the best way to identify importance. Or maybe you would phrase it differently, or maybe you don't think the "act or delete" mechanism is selective enough to filter for importance.
In any case, I like Fiona's method better. :-)
<< I don't understand how that works. The system would certainly favour completing tasks once they've been started, but I don't see why those would necessarily be more important than the ones which haven't been started. >>
Thanks Mark. I agree with you that the main feature of my proposal is to help drive tasks to completion once they've been started. The selection for importance is weaker, but I do think it's present. Here is how I was thinking that would work:
When the dice select the item, you have a choice: take action, or delete. If you take action, work for a while, then re-enter without completion, this fact (that you thought it worthwhile enough to take action rather than deleting it) gives the task some importance.
This is based on your principle that action itself is the best way to identify importance. Or maybe you would phrase it differently, or maybe you don't think the "act or delete" mechanism is selective enough to filter for importance.
In any case, I like Fiona's method better. :-)
January 12, 2014 at 0:01 |
Seraphim
Seraphim
The original AutoFocus (AF1) could be modified as follows -- if you cycle through a page and nothing stands out, rather than just dismissing the entire page, use a random number to choose one task on the page. Complete that task and move on to the next page.
This would ensure that even the hardest tasks on the earlier pages would eventually be addressed. It would also let you deal with your list as you see fit, with the randomizer only taking over when you are stuck.
If there were 8 open items on a page where nothing stands out, generate a random number between 1 and 8. If it returns 3, do the third item.
This would ensure that even the hardest tasks on the earlier pages would eventually be addressed. It would also let you deal with your list as you see fit, with the randomizer only taking over when you are stuck.
If there were 8 open items on a page where nothing stands out, generate a random number between 1 and 8. If it returns 3, do the third item.
January 12, 2014 at 22:29 |
Simon
Simon
Since I last wrote I've been trying out several different ways of using the randomizer - including one which was similar to Seraphim's suggestion.
So far I have found the best method to be the one I suggested earlier:
- start with two tasks
- add another task each time you finish working on a task
- leave tasks on the list unless you have completely finished with them for the day
- start again from scratch each day
- always set the randomizer to the number of tasks actually on the list
- best to use an automatically numbered list (e.g. Word, Evernote) then you don't have to keep counting tasks.
The advantages of doing it this way are:
1. You can ensure that the high-priority tasks get preference by entering them first.
2. The number of tasks on the list will bear a reasonable relationship to the amount of time you have available to do them.
Disadvantages:
I haven't found any real disadvantages yet, though if you want to make sure you don't miss anything you will need to have a "master list" as well.
General Comment:
It's best to work off quite large tasks and use project lists where necessary.
So far I have found the best method to be the one I suggested earlier:
- start with two tasks
- add another task each time you finish working on a task
- leave tasks on the list unless you have completely finished with them for the day
- start again from scratch each day
- always set the randomizer to the number of tasks actually on the list
- best to use an automatically numbered list (e.g. Word, Evernote) then you don't have to keep counting tasks.
The advantages of doing it this way are:
1. You can ensure that the high-priority tasks get preference by entering them first.
2. The number of tasks on the list will bear a reasonable relationship to the amount of time you have available to do them.
Disadvantages:
I haven't found any real disadvantages yet, though if you want to make sure you don't miss anything you will need to have a "master list" as well.
General Comment:
It's best to work off quite large tasks and use project lists where necessary.
January 14, 2014 at 10:52 |
Mark Forster
Mark Forster
Simon:
<< if you cycle through a page and nothing stands out, rather than just dismissing the entire page, use a random number to choose one task on the page. >>
One problem with this is that there is often only one task on the page anyway.
<< if you cycle through a page and nothing stands out, rather than just dismissing the entire page, use a random number to choose one task on the page. >>
One problem with this is that there is often only one task on the page anyway.
January 14, 2014 at 10:54 |
Mark Forster
Mark Forster
Mark wrote:
<< 1. You can ensure that the high-priority tasks get preference by entering them first. >>
If the randomizer selects each task with equal probability, I don't understand how it gives preference to the ones entered first. Can you explain that again?
<< 1. You can ensure that the high-priority tasks get preference by entering them first. >>
If the randomizer selects each task with equal probability, I don't understand how it gives preference to the ones entered first. Can you explain that again?
January 14, 2014 at 15:46 |
Seraphim
Seraphim
This was a fun thread to read. Mark got comments how many times in a row? Looks like the programmer put a lot of work into the options. (Will not wander over and procrastinate even more. There's free wifi in the
I'll often use a random number on my existing list, to break out of a stall. I use the second-hand on my watch. Depending on the list length, I might subtract 20 or 30. Don't count lines with completed tasks. If I reach the end of the list, I go back to the top of the list; older tasks get a bit more attention, which is useful if I leave uncompleted tasks on the list (marked "worked on today") rather than moving them to the bottom.
Less set-up time than typing in a list. Easier to shift back to your regular system when the stall is broken. Can do without entering computer room.
My list includes a wide variety of energy and interest levels.
An easier variation on Simon's method of giving some lines more priority is to decide which ones get more attention (number of stars?), then tap (count) twice or more as you reach those lines. So 1-read book, 2-cook supper, 3-cook supper, 4-read forums. Easier than copying the list each time priorities change.
Sometimes, being forced to do something (by the random number) actually forces you to make a decision. If the choices really are equal (food court line), I'd waffle and be unable to choose. Then I'd flip a coin, and realize I really didn't like what it chose. In that case, it was a system to force me to make a decision, not a system to ensure things got equal attention.
Seraphim, if I read it correctly, if you start a task but don't finish it, you don't erase it, but you do write it again at the end. That way it's both number 3 and number 27 -- twice the chance of being selected. If you do another chunk, it's numbers 3, 27 and 31. When it's finally done, you get to stroke out several lines at once. I did something like that for several months. The same task might get entered (and not done) on several week plans. Stroking it off multiple pages was a big reward!
Done one pass of the forum. On to the MIT.
I'll often use a random number on my existing list, to break out of a stall. I use the second-hand on my watch. Depending on the list length, I might subtract 20 or 30. Don't count lines with completed tasks. If I reach the end of the list, I go back to the top of the list; older tasks get a bit more attention, which is useful if I leave uncompleted tasks on the list (marked "worked on today") rather than moving them to the bottom.
Less set-up time than typing in a list. Easier to shift back to your regular system when the stall is broken. Can do without entering computer room.
My list includes a wide variety of energy and interest levels.
An easier variation on Simon's method of giving some lines more priority is to decide which ones get more attention (number of stars?), then tap (count) twice or more as you reach those lines. So 1-read book, 2-cook supper, 3-cook supper, 4-read forums. Easier than copying the list each time priorities change.
Sometimes, being forced to do something (by the random number) actually forces you to make a decision. If the choices really are equal (food court line), I'd waffle and be unable to choose. Then I'd flip a coin, and realize I really didn't like what it chose. In that case, it was a system to force me to make a decision, not a system to ensure things got equal attention.
Seraphim, if I read it correctly, if you start a task but don't finish it, you don't erase it, but you do write it again at the end. That way it's both number 3 and number 27 -- twice the chance of being selected. If you do another chunk, it's numbers 3, 27 and 31. When it's finally done, you get to stroke out several lines at once. I did something like that for several months. The same task might get entered (and not done) on several week plans. Stroking it off multiple pages was a big reward!
Done one pass of the forum. On to the MIT.
January 14, 2014 at 19:27 |
Cricket
Cricket
Seraphim:
<< If the randomizer selects each task with equal probability, I don't understand how it gives preference to the ones entered first. Can you explain that again? >>
Easy.
Say the first two tasks you enter are Email and Paper In-tray. You set the randomizer to 2. So each task has a 50% chance of being selected.
Since you need to work some more on whichever task was selected you leave it on the list. Now you add a third task, say Tax Return, and set the randomizer to 3. Each task now has a 33.33% chance of being selected. Then you add a fourth task, say Invoicing.
Your list now is:
1. Email
2. Paper
3. Tax Return
4. Invoicing
Set the randomizer to 4 and each task has a 25% chance of being selected.
Note that the first two tasks Email and Paper have had four opportunities to be selected (at 50%, 33.33%, and 25% odds), then Tax Return has had two opportunities at 33% and 25% respectively, while Invoicing has only had one opportunity at 25% odds.
And so on.
<< If the randomizer selects each task with equal probability, I don't understand how it gives preference to the ones entered first. Can you explain that again? >>
Easy.
Say the first two tasks you enter are Email and Paper In-tray. You set the randomizer to 2. So each task has a 50% chance of being selected.
Since you need to work some more on whichever task was selected you leave it on the list. Now you add a third task, say Tax Return, and set the randomizer to 3. Each task now has a 33.33% chance of being selected. Then you add a fourth task, say Invoicing.
Your list now is:
1. Email
2. Paper
3. Tax Return
4. Invoicing
Set the randomizer to 4 and each task has a 25% chance of being selected.
Note that the first two tasks Email and Paper have had four opportunities to be selected (at 50%, 33.33%, and 25% odds), then Tax Return has had two opportunities at 33% and 25% respectively, while Invoicing has only had one opportunity at 25% odds.
And so on.
January 14, 2014 at 20:33 |
Mark Forster
Mark Forster
Cricket:
<< Mark got comments how many times in a row? >>
I think it was four, but they weren't in a row - there were intervening tasks.
<< Seraphim, if I read it correctly, if you start a task but don't finish it, you don't erase it, but you do write it again at the end. >>
No, you just leave it on the list where it was. You don't re-enter it.
<< That way it's both number 3 and number 27 -- twice the chance of being selected. If you do another chunk, it's numbers 3, 27 and 31. >>
That's not it. You don't need any of that. Each throw gives all the tasks equal preference. The prioritizing of the first tasks on the list comes from the number of throws they are present for and the higher chances of being selected when there are only a few tasks on the list. See my previous post to Seraphim.
<< Mark got comments how many times in a row? >>
I think it was four, but they weren't in a row - there were intervening tasks.
<< Seraphim, if I read it correctly, if you start a task but don't finish it, you don't erase it, but you do write it again at the end. >>
No, you just leave it on the list where it was. You don't re-enter it.
<< That way it's both number 3 and number 27 -- twice the chance of being selected. If you do another chunk, it's numbers 3, 27 and 31. >>
That's not it. You don't need any of that. Each throw gives all the tasks equal preference. The prioritizing of the first tasks on the list comes from the number of throws they are present for and the higher chances of being selected when there are only a few tasks on the list. See my previous post to Seraphim.
January 14, 2014 at 20:49 |
Mark Forster
Mark Forster
I'm still fascinated by the random-number generator and am currently experimenting with a variation of Simon's method.
I'm using a notebook in which I cross out out tasks which I've finished working on for the time being. Unfinished and recurrent tasks are then re-entered at the end of the list. The list is treated as one long list. Pages with no tasks remaining are removed.
I set my randomizer to a maximum equal to the number of lines on a page. My notebook has 32 lines on a page, so the randomizer is set to produce an integer between 1 and 32 inclusive.
This is just a convenient number which produces reasonable results, but you can use a lower or higher number if you wish.
I use the randomizer to produce a number and move forward that number of spaces (onto the next page if necessary). I INCLUDE lines which have been crossed out in the count - this is important.
If the task I land on is crossed out I move to the next uncrossed-out task (on the next page if necessary).
When I reach the end of the list, I circle back to the beginning of the list, ignoring empty lines on the last page.
This method has the effect of increasing the likelihood of the older tasks on the list getting done. Note that if you don't include lines with crossed-out tasks in the count, then every task would have an exactly equal chance and there would be no preference for older tasks.
I'm using a notebook in which I cross out out tasks which I've finished working on for the time being. Unfinished and recurrent tasks are then re-entered at the end of the list. The list is treated as one long list. Pages with no tasks remaining are removed.
I set my randomizer to a maximum equal to the number of lines on a page. My notebook has 32 lines on a page, so the randomizer is set to produce an integer between 1 and 32 inclusive.
This is just a convenient number which produces reasonable results, but you can use a lower or higher number if you wish.
I use the randomizer to produce a number and move forward that number of spaces (onto the next page if necessary). I INCLUDE lines which have been crossed out in the count - this is important.
If the task I land on is crossed out I move to the next uncrossed-out task (on the next page if necessary).
When I reach the end of the list, I circle back to the beginning of the list, ignoring empty lines on the last page.
This method has the effect of increasing the likelihood of the older tasks on the list getting done. Note that if you don't include lines with crossed-out tasks in the count, then every task would have an exactly equal chance and there would be no preference for older tasks.
January 15, 2014 at 18:39 |
Mark Forster
Mark Forster
I'd imagine the bias toward selecting older tasks is equivalent to a bias toward *starting* tasks rather than *completing* tasks. How does it work out in actual practice?
January 16, 2014 at 6:22 |
Seraphim
Seraphim
"Note that the first two tasks Email and Paper have had four opportunities to be selected (at 50%, 33.33%, and 25% odds), then Tax Return has had two opportunities at 33% and 25% respectively, while Invoicing has only had one opportunity at 25% odds."
While that's true, it's a common fallacy to believe that this means there is bias to selecting them. The fact that they are still present means they weren't selected in previous random selections, therefore they cannot be counted towards any advantage. All that matters is their chance of being selected on the next selection, which is equal to the chance of anything else present being selected.
The only way to create an advantage for selected tasks is to bias the selector towards them. One method might be each time the selector is used the remaining tasks are represented by an additional copy in the list (which could be a virtual copy by simply writing the next available line number next to them so if the selector picks 3 or 33 that is the same task. On the next selection 3, 33 or 37 would select it and so on).
By the same reasoning I should do the same lottery numbers every week so they increasingly have more opportunities to be selected and win me some money. The fact that so many people do exactly this shows how common the fallacy is.
While that's true, it's a common fallacy to believe that this means there is bias to selecting them. The fact that they are still present means they weren't selected in previous random selections, therefore they cannot be counted towards any advantage. All that matters is their chance of being selected on the next selection, which is equal to the chance of anything else present being selected.
The only way to create an advantage for selected tasks is to bias the selector towards them. One method might be each time the selector is used the remaining tasks are represented by an additional copy in the list (which could be a virtual copy by simply writing the next available line number next to them so if the selector picks 3 or 33 that is the same task. On the next selection 3, 33 or 37 would select it and so on).
By the same reasoning I should do the same lottery numbers every week so they increasingly have more opportunities to be selected and win me some money. The fact that so many people do exactly this shows how common the fallacy is.
January 16, 2014 at 13:55 |
Chris
Chris
Chris, I also pondered whether this is a fallacy, and I don't think it is. It isn't the same situation as the lottery. In any one trial, every task has an equal chance of being selected - older tasks do not have a higher probability - you are right about that. But if you look at the results over 10 trials (for example), and make a list of the 10 tasks that got selected, you would find that tasks that were on the list for more of those trials appeared in the selected results more often.
January 16, 2014 at 15:09 |
Seraphim
Seraphim
See sample spreadsheet here: http://www.dropbox.com/s/77rvirdv7ltk2yu/task%20prob.xlsx
Row 18 uses the formula =RANDBETWEEN(1,COUNTA(K2:K16)) to select a task from the list in rows 2:16. When it selects the task, I then replace the formula with the chosen random number, and then decide whether to take the task off the list (completed) or let it ride (recurring or unfinished). I think it illustrates Mark's point that a task that remains on the list has a nonzero chance of being selected again. It actually worked out this way in this example, where "check email" was selected 3 times in 9 trials, at which point I deleted it because I figured I'd be done clearing out my email by then. :-)
Row 18 uses the formula =RANDBETWEEN(1,COUNTA(K2:K16)) to select a task from the list in rows 2:16. When it selects the task, I then replace the formula with the chosen random number, and then decide whether to take the task off the list (completed) or let it ride (recurring or unfinished). I think it illustrates Mark's point that a task that remains on the list has a nonzero chance of being selected again. It actually worked out this way in this example, where "check email" was selected 3 times in 9 trials, at which point I deleted it because I figured I'd be done clearing out my email by then. :-)
January 16, 2014 at 15:25 |
Seraphim
Seraphim
Chris said: "While that's true, it's a common fallacy to believe that this means there is bias to selecting them."
It's only a fallacy if the initial selections are considered to have already been made. Before any selections are made, it's perfectly valid, and that was Mark's point. You put what you want to do soon on the list first, when there are only two items, because at that point there is a strong likelihood that they will be done soon, as opposed to the likelihood that you will do something soon that you haven't added yet.
To illustrate this, the chance of getting two heads in a row when tossing a penny is 1/4 if I haven't tossed either penny yet, but it's 1/2 if I've already tossed one head, and 0 if I've already tossed one tail.
It's only a fallacy if the initial selections are considered to have already been made. Before any selections are made, it's perfectly valid, and that was Mark's point. You put what you want to do soon on the list first, when there are only two items, because at that point there is a strong likelihood that they will be done soon, as opposed to the likelihood that you will do something soon that you haven't added yet.
To illustrate this, the chance of getting two heads in a row when tossing a penny is 1/4 if I haven't tossed either penny yet, but it's 1/2 if I've already tossed one head, and 0 if I've already tossed one tail.
January 16, 2014 at 20:22 |
Austin
Austin
One only needs to look at Mark's percentages to see that he didn't make the fallacy referenced here. If he were making the fallacy, then he would have used a larger number than 25% to describe the chance of getting an item on the list when there are four items, since that item was on there longer. But he used 25% -- an equal probability for older tasks as for any other task.
January 16, 2014 at 20:27 |
Austin
Austin
The system Mark described, where if you land on a crossed-off line you do the next one gives more emphasis to the ones after crossed-off lines.
<<I use the randomizer to produce a number and move forward that number of spaces (onto the next page if necessary). I INCLUDE lines which have been crossed out in the count - this is important.>>
The task immediately after a crossed-off task has twice the chance of being done.
1
2
3
4
5x
6
7
8
9
10
Role between 1 and 10. If you get a 5 or a 6, you do #6.
1
2
3x
4
5x
6x
7
8
9x
10
Now 4 and 9 have two chances each, and 7 has three. #1 has nothing in front of it to be scratched off, so it will never get that boost.
Some easy ways to solve this:
As Mark suggested, roll a larger die, so crossing-off #10 gives #1 a boost.
Skip the crossed-off lines.
If you land on a crossed-off line, go to the one before it. That way 2 and 8 have twice the probability, and 4 has three times. Pretty soon, 2 and 4 will be crossed-off, and #1 will get five times the probability.
++++
Going back to rolling for page, then line number, on a list that's normally done with AF1. Lines on pages with fewer active lines will get chosen more often. That might be a good thing. Those lines have been already been put off many times.
<<I use the randomizer to produce a number and move forward that number of spaces (onto the next page if necessary). I INCLUDE lines which have been crossed out in the count - this is important.>>
The task immediately after a crossed-off task has twice the chance of being done.
1
2
3
4
5x
6
7
8
9
10
Role between 1 and 10. If you get a 5 or a 6, you do #6.
1
2
3x
4
5x
6x
7
8
9x
10
Now 4 and 9 have two chances each, and 7 has three. #1 has nothing in front of it to be scratched off, so it will never get that boost.
Some easy ways to solve this:
As Mark suggested, roll a larger die, so crossing-off #10 gives #1 a boost.
Skip the crossed-off lines.
If you land on a crossed-off line, go to the one before it. That way 2 and 8 have twice the probability, and 4 has three times. Pretty soon, 2 and 4 will be crossed-off, and #1 will get five times the probability.
++++
Going back to rolling for page, then line number, on a list that's normally done with AF1. Lines on pages with fewer active lines will get chosen more often. That might be a good thing. Those lines have been already been put off many times.
January 16, 2014 at 20:40 |
Cricket
Cricket
Seraphim:
<< I'd imagine the bias toward selecting older tasks is equivalent to a bias toward *starting* tasks rather than *completing* tasks. How does it work out in actual practice? >>
No not really. By the end of the day the two starting task on my sample list, Email and Paper, will have been present in every throw throughout the day. That means they stand a good chance of being selected several times, even though the odds lengthen as the day goes on against their being selected on any particular throw.
<< I'd imagine the bias toward selecting older tasks is equivalent to a bias toward *starting* tasks rather than *completing* tasks. How does it work out in actual practice? >>
No not really. By the end of the day the two starting task on my sample list, Email and Paper, will have been present in every throw throughout the day. That means they stand a good chance of being selected several times, even though the odds lengthen as the day goes on against their being selected on any particular throw.
January 16, 2014 at 23:16 |
Mark Forster
Mark Forster
Chris:
<< While that's true, it's a common fallacy to believe that this means there is bias to selecting them. The fact that they are still present means they weren't selected in previous random selections, therefore they cannot be counted towards any advantage. All that matters is their chance of being selected on the next selection, which is equal to the chance of anything else present being selected.>>
I said that the tasks would remain on the list for as long as work was needed on them, which in the case of Email and probably Paper would be all day. If there were 50 throws during the day the advantage they would get from being present for 50 throws as compared to another task which was only present for 10 throws would be considerable, even if the odds were the same for every throw - which they are not.
<< While that's true, it's a common fallacy to believe that this means there is bias to selecting them. The fact that they are still present means they weren't selected in previous random selections, therefore they cannot be counted towards any advantage. All that matters is their chance of being selected on the next selection, which is equal to the chance of anything else present being selected.>>
I said that the tasks would remain on the list for as long as work was needed on them, which in the case of Email and probably Paper would be all day. If there were 50 throws during the day the advantage they would get from being present for 50 throws as compared to another task which was only present for 10 throws would be considerable, even if the odds were the same for every throw - which they are not.
January 16, 2014 at 23:22 |
Mark Forster
Mark Forster
The system which I described yesterday (January 15, 2014 at 18:39) is working extremely well.
I've made one amendment. It relates to when the randomizer takes one to a crossed out task, and you move to the next active task. I call that a "slide". My amendment is that if a slide takes one to the end of the page, then you continue the slide from the top of the SAME page.
The reason I made this amendment was that I realized that the first task on the first page was at a disadvantage under my original rules because it could not be "slid" to from another page.
One amendment I'm thinking of making is to mark tasks which need to be done quickly or several times a day - I'd use a dot but anything will do. Then use the randomizer to select a task in the usual way but if there are any marked-up tasks in the gap between the selected task and the previously actioned task, then these tasks should be done before the selected one. Marking up tasks in this way should be done extremely sparingly.
I'm amazed at how well this system is working (even without the possible amendment). The randomness takes away all the procrastination element - which is much bigger than I'd realized. And the random numbers never cease to surprise me - they don't behave in the way I would instinctively expect at all.
I've made one amendment. It relates to when the randomizer takes one to a crossed out task, and you move to the next active task. I call that a "slide". My amendment is that if a slide takes one to the end of the page, then you continue the slide from the top of the SAME page.
The reason I made this amendment was that I realized that the first task on the first page was at a disadvantage under my original rules because it could not be "slid" to from another page.
One amendment I'm thinking of making is to mark tasks which need to be done quickly or several times a day - I'd use a dot but anything will do. Then use the randomizer to select a task in the usual way but if there are any marked-up tasks in the gap between the selected task and the previously actioned task, then these tasks should be done before the selected one. Marking up tasks in this way should be done extremely sparingly.
I'm amazed at how well this system is working (even without the possible amendment). The randomness takes away all the procrastination element - which is much bigger than I'd realized. And the random numbers never cease to surprise me - they don't behave in the way I would instinctively expect at all.
January 16, 2014 at 23:39 |
Mark Forster
Mark Forster
Cricket:
<< Going back to rolling for page, then line number, on a list that's normally done with AF1. Lines on pages with fewer active lines will get chosen more often. That might be a good thing. Those lines have been already been put off many times. >>
Lines haven't been "put off" in the sense we'd put something off. They just haven't been selected. The randomizer is completely indifferent to what tasks it selects.
Just to illustrate how it works, my first page (started yesterday) now has four tasks left on it. These tasks are on lines 1, 2, 5 and 9 of a 32-line page.
There are 24 lines on the page (1, 10-32) which if they were landed on would result in line 1 being selected. There is only one line on the page which would result in line 2 being selected, but if line 1 were to be selected then on the next throw there would be 25 lines on the page (1, 2 10-32) which would result in line 2 being selected.
It's quite feasible that all four tasks would be selected before the randomizer moves off the page.
Afternote:
I just missed clearing all four tasks off page 1. I threw a 27 from the last page which took me to line 22 on the first page, from where I slid to line 1. Then 5 came up and I slid to line 9. I then threw a 2 and slid all around the page to line 2. Only two numbers (31 and 32) would have prevented me from hitting line 5, the only unactioned line left. Of course I threw 31 which took me off the page.
Next time I hit that page there is a 100% change of dealing with that one remaining task.
Tally after 2 days on the system:
Page 1 - 1 task remaining
Page 2 - 0
Page 3 - 8
Page 4 - 19
Page 5 - 22
Page 6 - 26
Page 7 - 14 (of 16)
<< Going back to rolling for page, then line number, on a list that's normally done with AF1. Lines on pages with fewer active lines will get chosen more often. That might be a good thing. Those lines have been already been put off many times. >>
Lines haven't been "put off" in the sense we'd put something off. They just haven't been selected. The randomizer is completely indifferent to what tasks it selects.
Just to illustrate how it works, my first page (started yesterday) now has four tasks left on it. These tasks are on lines 1, 2, 5 and 9 of a 32-line page.
There are 24 lines on the page (1, 10-32) which if they were landed on would result in line 1 being selected. There is only one line on the page which would result in line 2 being selected, but if line 1 were to be selected then on the next throw there would be 25 lines on the page (1, 2 10-32) which would result in line 2 being selected.
It's quite feasible that all four tasks would be selected before the randomizer moves off the page.
Afternote:
I just missed clearing all four tasks off page 1. I threw a 27 from the last page which took me to line 22 on the first page, from where I slid to line 1. Then 5 came up and I slid to line 9. I then threw a 2 and slid all around the page to line 2. Only two numbers (31 and 32) would have prevented me from hitting line 5, the only unactioned line left. Of course I threw 31 which took me off the page.
Next time I hit that page there is a 100% change of dealing with that one remaining task.
Tally after 2 days on the system:
Page 1 - 1 task remaining
Page 2 - 0
Page 3 - 8
Page 4 - 19
Page 5 - 22
Page 6 - 26
Page 7 - 14 (of 16)
January 16, 2014 at 23:57 |
Mark Forster
Mark Forster
Hi Mark
Totaling your tasks for 2 days is a staggering 130 tasks! Even if I low balled the computation at 10 minutes per task, that's over 10 hours per day! Each task is randomly chosen. What if you have a few projects that require more than 10 minutes of concentration at a time. If you chose a few tasks that you spent an hour or two working on, that would bring your day totals to well over 10 hours a day. This can't be your everyday working hours! I can rack up 12-13 hours but it's usually only a few projects taking the bulk of my time and effort. I don't think I've EVER completed 130 tasks in two days. I don't complete that many in an entire week! LOL! In fact, I only work 10-12 hours when I'm overloaded by outside circumstances. I never voluntarily pile that much time and effort on my back unless it's ONE passion project that's swept me into a trance. What's your secret? Or are you front-loading your week to free up the weekend?
130 tasks in 2 days? Doesn't all that switching scramble up your focus and determination to complete particular jobs? Even before my brain damage I didn't have that amount of control over my brain to switch gears 60 times during a work session. My hat's off to you! What's your secret?
Totaling your tasks for 2 days is a staggering 130 tasks! Even if I low balled the computation at 10 minutes per task, that's over 10 hours per day! Each task is randomly chosen. What if you have a few projects that require more than 10 minutes of concentration at a time. If you chose a few tasks that you spent an hour or two working on, that would bring your day totals to well over 10 hours a day. This can't be your everyday working hours! I can rack up 12-13 hours but it's usually only a few projects taking the bulk of my time and effort. I don't think I've EVER completed 130 tasks in two days. I don't complete that many in an entire week! LOL! In fact, I only work 10-12 hours when I'm overloaded by outside circumstances. I never voluntarily pile that much time and effort on my back unless it's ONE passion project that's swept me into a trance. What's your secret? Or are you front-loading your week to free up the weekend?
130 tasks in 2 days? Doesn't all that switching scramble up your focus and determination to complete particular jobs? Even before my brain damage I didn't have that amount of control over my brain to switch gears 60 times during a work session. My hat's off to you! What's your secret?
January 17, 2014 at 15:15 |
learning as I go
learning as I go
learning:
I should think that's about right. 10 mins is an average. Some of the tasks take much longer than that, some no time at all (e.g. they just get crossed out because there's nothing to do). I use the system for everything - not just work.
I should think that's about right. 10 mins is an average. Some of the tasks take much longer than that, some no time at all (e.g. they just get crossed out because there's nothing to do). I use the system for everything - not just work.
January 18, 2014 at 2:07 |
Mark Forster
Mark Forster
This system continues to amaze me. I've got some really sticky tasks done today, including my tax return - for once it hasn't been left to the last minute!
Tally after 3 days on the system:
Page 1 - 0 tasks remaining
Page 2 - 0
Page 3 - 0
Page 4 - 6
Page 5 - 6
Page 6 - 17
Page 7 - 18
Page 9 - 27
Page 10 - 12 (of 12)
You can clearly see the effect of the emphasis on the older pages. After three days there are only 6 left of the 128 tasks I entered on day 1. The lack of resistance caused by the random selection has also made my work much faster. And the list has shrunk rather than expanded - there are only 87 tasks in it at the moment, compared with 90 at the end of Day 2.
The really amazing thing is that there is nothing that I feel behind on. There's not really even anything I feel I would like to have processed a bit faster. It's just got everything done, without fuss and without stress and, as Learning has pointed out above, the "everything" I've got done has been huge.
Tally after 3 days on the system:
Page 1 - 0 tasks remaining
Page 2 - 0
Page 3 - 0
Page 4 - 6
Page 5 - 6
Page 6 - 17
Page 7 - 18
Page 9 - 27
Page 10 - 12 (of 12)
You can clearly see the effect of the emphasis on the older pages. After three days there are only 6 left of the 128 tasks I entered on day 1. The lack of resistance caused by the random selection has also made my work much faster. And the list has shrunk rather than expanded - there are only 87 tasks in it at the moment, compared with 90 at the end of Day 2.
The really amazing thing is that there is nothing that I feel behind on. There's not really even anything I feel I would like to have processed a bit faster. It's just got everything done, without fuss and without stress and, as Learning has pointed out above, the "everything" I've got done has been huge.
January 18, 2014 at 2:19 |
Mark Forster
Mark Forster
This thread is getting too long to be manageable. I'm going to close it now. I'm opening a new thread on the same subject as I think there's more to say on the subject - which I at least am finding very interesting.
New thread at: http://markforster.squarespace.com/forum/post/2275650
New thread at: http://markforster.squarespace.com/forum/post/2275650
January 18, 2014 at 10:37 |
Mark Forster
Mark Forster
<< There's also a button to pick a different set if you don't like the current list. >>
It's not really random if you keep pressing the button until you get the answer you want!