02-13-2021, 03:34 AM
(This post was last modified: 02-15-2021, 06:02 AM by Asked Madden.)
Well, after a lot of procrastinating and being sidetracked with a lot of new responsibilities and positions within the league, I have finally got back around to spreading some love to the fantasy football addicts in this league. If you can remember, what i did was go through and gather data from all the groups about how people in each pick position performs and do what is some relatively simple analysis on it to see if there is a pick position that is better or worse than the rest, in case you need an excuse for why you did not get TPE or your league friend in the same group beat you.
Whilst we are currently in S27, so far this project of mine has only been done for S22 and S24, meaning there are two seasons, S25 in this post, as well as the S26 version that will be out shortly. Unfortunately the S23 sheet doesn’t exist as it was lost/removed with the change of fantasy heads I believe. With the S26 version, for the ease of the few people who actually care about what is in this article, I will be making a spreadsheet with all the data and analysis available to everyone for all the seasons. why not?
So, the first part in all this is trolling through all the archived group pages on the forum, going to their serpentine draft order, then utilising the S25 fantasy score sheet, working out who was drafting in each spot and their final result. For S25, there were a total of 36 groups, however I decided to exclude group 35 due to them only having 5 people in the group and making the numbers less round and easy to deal with. All this data is compiled into the table below, and everything else done in terms of analysis all stems from this data and extrapolating it into different forms and variations to test performance in more meaningful or cared about ways. A reminder, since if you came first in fantasy that is equivalent to 1 point, and last place is 6 points, this table works like a game of golf, lowest total wins.
![[Image: 2BzqBFm.png]](https://i.imgur.com/2BzqBFm.png)
Just observing the total numbers, it appears that this season was all relatively close to one another, much more than previous seasons. Since there is 35 groups, and each group consists of 21 points (6 +5 +4 +3 +2 +1 = quick maths), there is a total of 735 points in this data pool, which dividing by six gives us what the average and the outcome in a perfectly balanced fantasy football would be, a total of 122.5 for each group.
Second and Sixth pick both are just above the average, whilst Pick three is still by very close to average for what is fifth place. First pick comes in at the back of the back, but remains only 8.5 above the equal outcome number.. This means per group, they performed 0.243 above average, a very insignificant amount. Below average we have Pick four, which is equidistant from the average as fifth place Pick three, whilst the best performing pick by raw numbers, Pick 5, also remains very close to the average and the range of other scores.
In a new table to the series, to bump word count as well as another way to compare seasons, the below is a simple method to see how spread out pick positions were from their average in each season. The table is pretty self-explanatory, with expected outcome being the average (for this season it is the 122.5 number) you expect to see each group have a perfectly balanced result and the range being the Highest minus the Lowest. What we would want to see is as small of a spread as possible, meaning that pick position had a small effect on the eventual outcome of that season. These compare tables will be used for some of the other measures I typically used since they are a great way to provide summaries and contrasts to the past seasons.
![[Image: GyBvNVS.png]](https://i.imgur.com/GyBvNVS.png)
This table shows that it is true that Season 25 was the closest for pick position, not only having the smallest range and by far the smallest distance from the best performing pick to average, but tying with the highest picks difference to average. Season 24 is really far off, especially at the high end, and may show that the users picking at that position were simply underperforming/worse fantasy players? It is hard to tell with only 3 seasons of data but the more and more that I do this, the clearer and more comprehensive of a picture this paints.
Now, whilst the raw data about how each position performed in each group is the back-bone of this series, what I assume most people really care and see as the most important breakdown of these articles, is how pick position resulted in TPE. Of course, the top 3 positions in a group get TPE. So, if a position is able to consistently achieve such results, whether it be 1, 3 or 5 TPE earnt, you will at least be getting something for playing instead of just the $1 million dollars for completing the draft. Some people would go to the high risk for a high reward position, whilst others would prefer the low risk but low reward option in life, and this helps reflect on how the pick positions fall on that scale. With 35 groups, the expected average outcome would be 5.833*.
![[Image: lkaYPgk.png]](https://i.imgur.com/lkaYPgk.png)
On first look, this seems pretty even all round and despite not being perfectly even, each pick is quite balanced, especially shown with their chance of coming second, even coming first has no high outliers. It is the lower result positions that have some numbers way off base, First pick has 10 fifth place finishes, almost a third of all results, and the only combination of pick position and finish result to happen in the double digits., whilst on the opposite end of the spectrum, third pick only managed third place 2 times, the lowest combination. Also of note is second pick finishing in sixth place 9 times, more than any other pick.
Turning this data into a somewhat better table/analysis for determining the strength or weakness or a pick position, since the top 3 earn TPE and is the goal for most users to be in one of those spots, putting this data into the top 3 and bottom 3 can provide a figure about chance of getting a TPE position, as well as including the amount of TPE earnt as a pick position may have more top 3 finishes, but if they are all third place, another pick position may end up earning more TPE in a situation of a higher risk, higher reward.
![[Image: aF5nmZl.png]](https://i.imgur.com/aF5nmZl.png)
Once again, if all groups were robotic and even, the average expected outcome would be 17.5 top 3 finishes. Third pick had the lowest amount of top 3 finishes, yet finishes almost fourth in TPE earnt, only one behind sixth pick despite having 5 less TPE finishes. First pick being quite spread out amongst their top 3 finishes, puts them in a solid last for TPE earnt, despite being just below the average amount of top 3 finishes you should expect from a pick position. On the other side of that, Both Fourth and Fifth pick had 20 top 3 finishes, as a percentage that is 57% of people from that position that earnt TPE. Fourth pick earns more TPE due to having the extra second place finish over fifth place. Sixth pick and second pick are the interesting results here, as the former has almost as many top 3 finishes as the tie top picks, but has significantly less TPE, meaning they are consistent but not high performing. The latter, second pic, is only 1 TPE off being equal for 2nd in TPE earnt, but sits in the middle of the pack in terms of top 3 finishes, demonstrating that their placements were higher, the same amount of first and second finishes as the top spot fourth pick, just a little less consistent.
Since there will be a S26 edition to this soon enough, I am going to save some of the extra season by season analysis and talk for that edition, since i would just be restating everything with extra data anyways. So until then, enjoy this and hopefully fantasy football can keep getting some media on it (R.I.P Art) because we all enjoy it so much, especially with the fantasy bot (Thanks Infinite for making a new one) and adds not only some more realistic elements to our league, but an extra layer of depth to the league that make us more invested because we choose names from a piece of (internet) paper
Whilst we are currently in S27, so far this project of mine has only been done for S22 and S24, meaning there are two seasons, S25 in this post, as well as the S26 version that will be out shortly. Unfortunately the S23 sheet doesn’t exist as it was lost/removed with the change of fantasy heads I believe. With the S26 version, for the ease of the few people who actually care about what is in this article, I will be making a spreadsheet with all the data and analysis available to everyone for all the seasons. why not?
So, the first part in all this is trolling through all the archived group pages on the forum, going to their serpentine draft order, then utilising the S25 fantasy score sheet, working out who was drafting in each spot and their final result. For S25, there were a total of 36 groups, however I decided to exclude group 35 due to them only having 5 people in the group and making the numbers less round and easy to deal with. All this data is compiled into the table below, and everything else done in terms of analysis all stems from this data and extrapolating it into different forms and variations to test performance in more meaningful or cared about ways. A reminder, since if you came first in fantasy that is equivalent to 1 point, and last place is 6 points, this table works like a game of golf, lowest total wins.
Just observing the total numbers, it appears that this season was all relatively close to one another, much more than previous seasons. Since there is 35 groups, and each group consists of 21 points (6 +5 +4 +3 +2 +1 = quick maths), there is a total of 735 points in this data pool, which dividing by six gives us what the average and the outcome in a perfectly balanced fantasy football would be, a total of 122.5 for each group.
Second and Sixth pick both are just above the average, whilst Pick three is still by very close to average for what is fifth place. First pick comes in at the back of the back, but remains only 8.5 above the equal outcome number.. This means per group, they performed 0.243 above average, a very insignificant amount. Below average we have Pick four, which is equidistant from the average as fifth place Pick three, whilst the best performing pick by raw numbers, Pick 5, also remains very close to the average and the range of other scores.
In a new table to the series, to bump word count as well as another way to compare seasons, the below is a simple method to see how spread out pick positions were from their average in each season. The table is pretty self-explanatory, with expected outcome being the average (for this season it is the 122.5 number) you expect to see each group have a perfectly balanced result and the range being the Highest minus the Lowest. What we would want to see is as small of a spread as possible, meaning that pick position had a small effect on the eventual outcome of that season. These compare tables will be used for some of the other measures I typically used since they are a great way to provide summaries and contrasts to the past seasons.
This table shows that it is true that Season 25 was the closest for pick position, not only having the smallest range and by far the smallest distance from the best performing pick to average, but tying with the highest picks difference to average. Season 24 is really far off, especially at the high end, and may show that the users picking at that position were simply underperforming/worse fantasy players? It is hard to tell with only 3 seasons of data but the more and more that I do this, the clearer and more comprehensive of a picture this paints.
Now, whilst the raw data about how each position performed in each group is the back-bone of this series, what I assume most people really care and see as the most important breakdown of these articles, is how pick position resulted in TPE. Of course, the top 3 positions in a group get TPE. So, if a position is able to consistently achieve such results, whether it be 1, 3 or 5 TPE earnt, you will at least be getting something for playing instead of just the $1 million dollars for completing the draft. Some people would go to the high risk for a high reward position, whilst others would prefer the low risk but low reward option in life, and this helps reflect on how the pick positions fall on that scale. With 35 groups, the expected average outcome would be 5.833*.
On first look, this seems pretty even all round and despite not being perfectly even, each pick is quite balanced, especially shown with their chance of coming second, even coming first has no high outliers. It is the lower result positions that have some numbers way off base, First pick has 10 fifth place finishes, almost a third of all results, and the only combination of pick position and finish result to happen in the double digits., whilst on the opposite end of the spectrum, third pick only managed third place 2 times, the lowest combination. Also of note is second pick finishing in sixth place 9 times, more than any other pick.
Turning this data into a somewhat better table/analysis for determining the strength or weakness or a pick position, since the top 3 earn TPE and is the goal for most users to be in one of those spots, putting this data into the top 3 and bottom 3 can provide a figure about chance of getting a TPE position, as well as including the amount of TPE earnt as a pick position may have more top 3 finishes, but if they are all third place, another pick position may end up earning more TPE in a situation of a higher risk, higher reward.
Once again, if all groups were robotic and even, the average expected outcome would be 17.5 top 3 finishes. Third pick had the lowest amount of top 3 finishes, yet finishes almost fourth in TPE earnt, only one behind sixth pick despite having 5 less TPE finishes. First pick being quite spread out amongst their top 3 finishes, puts them in a solid last for TPE earnt, despite being just below the average amount of top 3 finishes you should expect from a pick position. On the other side of that, Both Fourth and Fifth pick had 20 top 3 finishes, as a percentage that is 57% of people from that position that earnt TPE. Fourth pick earns more TPE due to having the extra second place finish over fifth place. Sixth pick and second pick are the interesting results here, as the former has almost as many top 3 finishes as the tie top picks, but has significantly less TPE, meaning they are consistent but not high performing. The latter, second pic, is only 1 TPE off being equal for 2nd in TPE earnt, but sits in the middle of the pack in terms of top 3 finishes, demonstrating that their placements were higher, the same amount of first and second finishes as the top spot fourth pick, just a little less consistent.
Since there will be a S26 edition to this soon enough, I am going to save some of the extra season by season analysis and talk for that edition, since i would just be restating everything with extra data anyways. So until then, enjoy this and hopefully fantasy football can keep getting some media on it (R.I.P Art) because we all enjoy it so much, especially with the fantasy bot (Thanks Infinite for making a new one) and adds not only some more realistic elements to our league, but an extra layer of depth to the league that make us more invested because we choose names from a piece of (internet) paper
Code:
1,483 words + simple tables/stats/analysis![[Image: image0.png]](https://cdn.discordapp.com/attachments/372602034198806530/707467132568338462/image0.png)
![[Image: Ahf3LtW.png]](https://cdn.discordapp.com/attachments/379532993070628864/703443515136278608/Ahf3LtW.png)
