06-26-2020, 12:56 PM
(This post was last modified: 06-26-2020, 04:27 PM by Fordhammer.)
In the normal mid-season lull of media and action, combined with my abundance of time on my hands for the following month, I decided I wanted to look into a different aspects of this league’s history. Everything from me in the past has revolved around the draft, the history of past 1OA picks in both the DSFL and NSFL drafts, as well as analysis into my own class during our DSFL season, being last season. As you have seen in the title however, I instead looked into the fantasy football archives since its something you don’t see much coverage on its history. When the start of the season hits, fantasy drafting takes hold of the locker rooms for about a week and then there is mummering’s here and there about waivers. What isn’t mentioned, at least in my experience compared to other historical aspects of the league, is the past seasons of playing fantasy football and the history behind it.
When deciding what to actually look into, there was a few things that sprang to mind and in all honestly there will probably be another fantasy football history article done by me during this season on some of these aspects or many some sort of inspiration I get from someone or somewhere. Unfortunately, however, whilst all the group pages are achieved for many and many of seasons, there is 25 pages achieved away, only the S22 Master sheet is available to access. I did message people about this, and shamefully it seems the history and records weren’t really kept or are not available to access. As a result, the data I obtain for this and future articles will hopefully become the base line for future pieces by myself or others to start a collection of the fantasy history since it is an aspect of the league that I would say, a overwhelming majority of active users participate in. This got me to thinking what was brought up in conversations about fantasy that would be interesting to look into with the available information. What I first noticed and thought about was how people were asking questions not only about who to draft, rather how to draft being in a certain spot. This got me to thinking whether there is any major differences between the position in how successful they were in last season. This, brings us to the bulk of this media piece.
With the S22 master sheet being available, I could go through and see the results within in each group, seeing who came first, second, third etc. etc. Combining this with the achieved group pages, I was able to determine what pick each user had and compile a table, recording this information to see if there was any intriguing data points through analysis. Seen below in the first table, is the raw data used throughout the analysis. The group page was used to see who had what pick in that group, and their result according to the master sheet was noted into the table. This was done for each of the 31 groups.
![[Image: w56tMvU.png]](https://i.imgur.com/w56tMvU.png)
Similar to a golf system, users who came first are represented by a 1, second by a 2 and so forth. This means that the lower the total value, the better they performed. For example, if a pick achieved first place in each group, it would have a total value of 31, compared to 186 if the pick always came last.
If all groups were to be even and average out, the expected total of each pick should be 108.5. Across 31 groups and users of all skill and experience, it is fair to say there is a large enough sample space to achieve results that hold some value. For clarification, I asked BWII how groups and draft order were worked out, and he told me that everyone who signs up is entered into a randomiser and the top 6 names are taken, the order remains the pick order, and that continues for all the groups. This entails all our samples are random and non-biased.
As for the data, I found the results intriguing. Users who had the first or fourth pick performed the worse. Users with the sixth pick also performed slightly under the statistically expected result. Fifth pick performed right on average and second pick performed slightly better, at a rate similar to how the sixth pick performed under average. The biggest point in this data is the substantially large difference between the best performing pick, pick three, from the statistically expected result. With such a large difference from this value and compared to the results from the other picks, I ran the numbers through an outlier test. As seen in the table below, whilst not far off, it is not considered an outlier statistically as, for this test, the outlier mark was 1.89. However, it is certainly much further from the average, 108.5, than any other of the figures.
![[Image: iW3DCv8.png]](https://i.imgur.com/iW3DCv8.png)
To go further into the analysis, I decided to break down how many times the users in each pick were able to achieve each result, that is, how many times did each pick come first, second, third…sixth. This will show us a more concise version of the initial table, one that is easier to unpack and give a little more understanding as to the numbers above as we can see how exactly one pick is out performing the other
![[Image: iPHWJbd.png]](https://i.imgur.com/iPHWJbd.png)
If each pick was to achieve every result evenly in an idealistic world, each value in the table would be 5.17, so mainly 5’s with one 6. What we do see however, is that values for most of the picks do not match this. Pick one and six are both the closest to achieving the expected average, but both have multiple placements being achieved 4 times and sixth place and fourth place being achieved 7 times respectively. The most notable data extrapolated through this table is some of the imbalance between achieving a top three result, and thusly earning TPE, or a bottom three result, with some picks achieving places only once or as many as 9 times, almost a third of all results in that group.
When looking into the TPE earned per pick, first place awards 5 TPE, second place 3 TPE and third place gives 1 TPE. These numbers are therefore weighted, achieving more first and seconds with a lack of three can still see that pick giving out more TPE than a pick that has a balanced amount of first, seconds and third.
![[Image: KVUj8Xt.png]](https://i.imgur.com/KVUj8Xt.png)
If evenly distributed, again in an idealistic, perfectly even world, it would be 15.5. More interesting in this table is the continuation on what is found from the table before, breaking down the positions achieved by each pick. What we see is that whilst overall, fourth pick is the tied worst performer, it has the second most TPE winners, due to the fact that its losing positions, fourth, fifth a fifth, were heavily weighted by the amount of fifths and sixths and only one fourth. This reveals that for the most part, the picks are a lot more even then the previous tables and results may show. This extend however, until the third pick, where 71% of users who had this pick, achieved TPE compared to a rate of 42% for pick one or pick five. Pick three, therefore, continues to perform much higher than any of the others picks in the S22 of fantasy football.
![[Image: BA6uQVA.png]](https://i.imgur.com/BA6uQVA.png)
This table further completes and compliments the analysis above, showing further the supposed benefits and likelihood of success being pick 3. Whilst fourth pick had the second-most TPE earning results, it is only beating the last placed first pick users by 1 TPE. Yet again, in an idealistic world, each pick would have earned 46.5 TPE. We see fifth pick, despite having the equal lowest amount of TPE earning positions, is almost exactly on this mark and the only one.
~~~Conclusion~~~
So what does all these tables and statistics mean? The TD:LR of it all is that whilst random, it seems as though being pick three has the best chance of succeeding in fantasy. Why? That is a very good question that I don’t have an answer for. Maybe someone more experienced in fantasy could explain it. Unfortunately, with a lack of master sheets for previous season, I cannot see if this is a trend that is commonly occurred throughout fantasy history or is just a one-time occurrence. What can be done to balance it out? Well, nothing. The Users in each group are randomized. The draft order, therefore, is randomized. There is absolutely no bias in this system. What also is important is that some players don’t finish drafting, users even had drafted only one player and were promptly named “F. Armstrong” instead of their username on the master sheet accordingly. What also happens due to the randomising, is the creation of “easy groups” where a league vet is placed with a bunch of rookies or “groups of death” where either 6 experienced players are grouped together or alternatively, a rookie is placed with 4-5 league vets. Without fully going through each user’s fantasy history, impossible without knowing their results in all the seasons pass, the effect this has on this current analysis is unknown. The randomising takes this out, but only looking at one year makes it unknown whether it just so happened that the better fantasy players got pick three more often in this year.
Hopefully, after the S23 season is over, during its offseason there will be more data to pour over and to come back to this analysis and results to see how the two seasons compare. Also hope that this spreads some light on less glamourous and under-appreciated history of the league and that someone reading this either has or might know someone with previous season records that could be made available to me, if not everyone, to reminisce over or, for me personally, to track any trends to see if there really is one pick that is superior to others. Lastly, if you are first pick this season, make sure to reference this article if you don’t come in the top 3 and how the odds were against you. Even further, reference this article if the person in your group who has the third pick gets too big for their boots and you want to prove it wasn’t skill but rather luck in the pick they received that gave them their success. Because after all, who doesn’t want to look good and superior?
When deciding what to actually look into, there was a few things that sprang to mind and in all honestly there will probably be another fantasy football history article done by me during this season on some of these aspects or many some sort of inspiration I get from someone or somewhere. Unfortunately, however, whilst all the group pages are achieved for many and many of seasons, there is 25 pages achieved away, only the S22 Master sheet is available to access. I did message people about this, and shamefully it seems the history and records weren’t really kept or are not available to access. As a result, the data I obtain for this and future articles will hopefully become the base line for future pieces by myself or others to start a collection of the fantasy history since it is an aspect of the league that I would say, a overwhelming majority of active users participate in. This got me to thinking what was brought up in conversations about fantasy that would be interesting to look into with the available information. What I first noticed and thought about was how people were asking questions not only about who to draft, rather how to draft being in a certain spot. This got me to thinking whether there is any major differences between the position in how successful they were in last season. This, brings us to the bulk of this media piece.
With the S22 master sheet being available, I could go through and see the results within in each group, seeing who came first, second, third etc. etc. Combining this with the achieved group pages, I was able to determine what pick each user had and compile a table, recording this information to see if there was any intriguing data points through analysis. Seen below in the first table, is the raw data used throughout the analysis. The group page was used to see who had what pick in that group, and their result according to the master sheet was noted into the table. This was done for each of the 31 groups.
Similar to a golf system, users who came first are represented by a 1, second by a 2 and so forth. This means that the lower the total value, the better they performed. For example, if a pick achieved first place in each group, it would have a total value of 31, compared to 186 if the pick always came last.
If all groups were to be even and average out, the expected total of each pick should be 108.5. Across 31 groups and users of all skill and experience, it is fair to say there is a large enough sample space to achieve results that hold some value. For clarification, I asked BWII how groups and draft order were worked out, and he told me that everyone who signs up is entered into a randomiser and the top 6 names are taken, the order remains the pick order, and that continues for all the groups. This entails all our samples are random and non-biased.
As for the data, I found the results intriguing. Users who had the first or fourth pick performed the worse. Users with the sixth pick also performed slightly under the statistically expected result. Fifth pick performed right on average and second pick performed slightly better, at a rate similar to how the sixth pick performed under average. The biggest point in this data is the substantially large difference between the best performing pick, pick three, from the statistically expected result. With such a large difference from this value and compared to the results from the other picks, I ran the numbers through an outlier test. As seen in the table below, whilst not far off, it is not considered an outlier statistically as, for this test, the outlier mark was 1.89. However, it is certainly much further from the average, 108.5, than any other of the figures.
To go further into the analysis, I decided to break down how many times the users in each pick were able to achieve each result, that is, how many times did each pick come first, second, third…sixth. This will show us a more concise version of the initial table, one that is easier to unpack and give a little more understanding as to the numbers above as we can see how exactly one pick is out performing the other
If each pick was to achieve every result evenly in an idealistic world, each value in the table would be 5.17, so mainly 5’s with one 6. What we do see however, is that values for most of the picks do not match this. Pick one and six are both the closest to achieving the expected average, but both have multiple placements being achieved 4 times and sixth place and fourth place being achieved 7 times respectively. The most notable data extrapolated through this table is some of the imbalance between achieving a top three result, and thusly earning TPE, or a bottom three result, with some picks achieving places only once or as many as 9 times, almost a third of all results in that group.
When looking into the TPE earned per pick, first place awards 5 TPE, second place 3 TPE and third place gives 1 TPE. These numbers are therefore weighted, achieving more first and seconds with a lack of three can still see that pick giving out more TPE than a pick that has a balanced amount of first, seconds and third.
If evenly distributed, again in an idealistic, perfectly even world, it would be 15.5. More interesting in this table is the continuation on what is found from the table before, breaking down the positions achieved by each pick. What we see is that whilst overall, fourth pick is the tied worst performer, it has the second most TPE winners, due to the fact that its losing positions, fourth, fifth a fifth, were heavily weighted by the amount of fifths and sixths and only one fourth. This reveals that for the most part, the picks are a lot more even then the previous tables and results may show. This extend however, until the third pick, where 71% of users who had this pick, achieved TPE compared to a rate of 42% for pick one or pick five. Pick three, therefore, continues to perform much higher than any of the others picks in the S22 of fantasy football.
This table further completes and compliments the analysis above, showing further the supposed benefits and likelihood of success being pick 3. Whilst fourth pick had the second-most TPE earning results, it is only beating the last placed first pick users by 1 TPE. Yet again, in an idealistic world, each pick would have earned 46.5 TPE. We see fifth pick, despite having the equal lowest amount of TPE earning positions, is almost exactly on this mark and the only one.
~~~Conclusion~~~
So what does all these tables and statistics mean? The TD:LR of it all is that whilst random, it seems as though being pick three has the best chance of succeeding in fantasy. Why? That is a very good question that I don’t have an answer for. Maybe someone more experienced in fantasy could explain it. Unfortunately, with a lack of master sheets for previous season, I cannot see if this is a trend that is commonly occurred throughout fantasy history or is just a one-time occurrence. What can be done to balance it out? Well, nothing. The Users in each group are randomized. The draft order, therefore, is randomized. There is absolutely no bias in this system. What also is important is that some players don’t finish drafting, users even had drafted only one player and were promptly named “F. Armstrong” instead of their username on the master sheet accordingly. What also happens due to the randomising, is the creation of “easy groups” where a league vet is placed with a bunch of rookies or “groups of death” where either 6 experienced players are grouped together or alternatively, a rookie is placed with 4-5 league vets. Without fully going through each user’s fantasy history, impossible without knowing their results in all the seasons pass, the effect this has on this current analysis is unknown. The randomising takes this out, but only looking at one year makes it unknown whether it just so happened that the better fantasy players got pick three more often in this year.
Hopefully, after the S23 season is over, during its offseason there will be more data to pour over and to come back to this analysis and results to see how the two seasons compare. Also hope that this spreads some light on less glamourous and under-appreciated history of the league and that someone reading this either has or might know someone with previous season records that could be made available to me, if not everyone, to reminisce over or, for me personally, to track any trends to see if there really is one pick that is superior to others. Lastly, if you are first pick this season, make sure to reference this article if you don’t come in the top 3 and how the odds were against you. Even further, reference this article if the person in your group who has the third pick gets too big for their boots and you want to prove it wasn’t skill but rather luck in the pick they received that gave them their success. Because after all, who doesn’t want to look good and superior?
Code:
1767 words by my count![[Image: image0.png]](https://cdn.discordapp.com/attachments/372602034198806530/707467132568338462/image0.png)
![[Image: Ahf3LtW.png]](https://cdn.discordapp.com/attachments/379532993070628864/703443515136278608/Ahf3LtW.png)
![[Image: 019p.png]](https://i.postimg.cc/J7XFN96k/019p.png)
![[Image: IOg9i8X.jpeg]](https://i.imgur.com/IOg9i8X.jpeg)
![[Image: 58742_s.png]](https://signavatar.com/58742_s.png)
![[Image: 58742_v.png]](https://signavatar.com/58742_v.png)
![[Image: 5fKn9dY.png]](https://i.imgur.com/5fKn9dY.png)
