*Doc's S47 ISFL Fantasy Season Post-Mortem - Printable Version +- International Simulation Football League (https://forums.sim-football.com) +-- Forum: Community (https://forums.sim-football.com/forumdisplay.php?fid=5) +--- Forum: Media (https://forums.sim-football.com/forumdisplay.php?fid=37) +---- Forum: Graded Articles (https://forums.sim-football.com/forumdisplay.php?fid=38) +---- Thread: *Doc's S47 ISFL Fantasy Season Post-Mortem (/showthread.php?tid=51002)

*Doc's S47 ISFL Fantasy Season Post-Mortem - DocLars - 05-01-2024 When I joined the ISFL last offseason, I wasn’t sure what to think of all the different things that were offered. A casino? Fantasy football about fantasy football? My initial thought was that these were things that could be fun for some, but not something I would get actively involved in. Fast forward to week 1 and I’m sitting in my office chair slobbering, shouting at the screen to throw the ball to the dots I drafted. Now that the season is over and I’ve won my group, I’ve accepted that ISFL fantasy football is the ultimate game of skill. As I continue to dive into this new hobby, I wanted to take some time to recap how the S47 fantasy season went. Who was the most valuable player? Who appeared on the most championship teams? Who was the best late round steal or waiver pickup this season? All that and more below! Fantasy Difference Makers Starting off, I wanted to identify who the most valuable player was across the ISFL fantasy football landscape. To accomplish this, I attempted to calculate each player’s value to see how much a difference they made for their fantasy team. I tried to avoid making this too complicated, so I sorted each position by total score and then took the value of the rank that would be the worst player rostered in an ideal league. For example, since there is a max of 6 quarterbacks rostered in a group, we can calculate the replacement value by using the score of the 6th ranked quarterback. From there, we can take the difference of a player’s points scored vs the replacement value and get their final value. Here is what the numbers looked like for the top 20 players:  With these numbers in mind, we can confidently say that Reginald Shrubbery is your S47 ISFL Fantasy MVP. No player provided more value, and he’d be the easy #1 pick if you were redrafting S47 today. But you could have figured that out in about 10 seconds just by looking for the highest scoring RB on the fantasy sheet, so let’s use this data to dive a little bit deeper. Sleeper Picks While the Fantasy MVP is certainly Shrubbery, he had an average draft position of 2nd overall, meaning you expected him to score you a ton of points every week. For our next exercise, we are going to try to find this year’s sleepers by comparing average draft position to actual value provided. Again, I’m not going to spend time developing a fancy formula with a lot of variables because I think we can use a simple solution to get us close enough. To determine “expected value”, I first redrafted a group based solely on player value to determine what the ideal draft would look like. Basically, this is what would happen if every team made the perfect pick whenever they were on the clock. From there, I took the value generated by the player taken at that draft position and set it as the base “expected value” for that pick. For example, Mister Hogmally would have been drafted 7th overall in the ideal draft and he produced a value (calculated earlier on in this article) of 112. If you had the 7th pick, that would be the expected value. Likewise, the 70th selection in the draft would have an expected value of 0 so any player drafted in that spot would not be expected to provide any value. Using this season’s ADP data, I then sorted and ranked them to determine where they would be drafted in an average draft. Now that we have their expected draft position and their ideal draft position, we can calculate the difference in actual value vs expected value. Going back to the same example above, the player with the 7th highest ADP (in this case Ace Anderson) would be expected to provide a value of 112 because that’s what the actual 7th best pick provided. I’ll note that using average ADP instead of ADP rank would probably give this metric a little more validity, but I already spent enough of my time on this. If they provided more value than expected, it was a great pick. Less and you could have done better. You get the idea, so let’s look at the top 10 value picks of the season: There were quite a few “league winning” picks that you could find late in the draft, but the title of S47 League Winner goes to: Mister Hogmally! Even though his average ADP had him being drafted at the end of the 10th round (he wasn’t even drafted in 13 of 25 groups) he finished as the 7th most valuable player in the entire league. If you managed to snag him, there is a good chance you won your group. And just for fun, here were this season’s worst values: Anatomy of a Group Winner We’ve already determined which players made the biggest difference for their teams, but were those players also the ones that appeared on the most group winning teams? Was there any noticeable strategy that group winners had that set them apart? Was it just dumb luck? Let’s dive in. First, I looked at each group winner’s final roster to see the most common players seen on these teams. There were 7 players who finished on over 30% of championship teams. The highest owned players by position were as follows: QB: Thor (28%) RB: Skywalker (52%) WR: Robinson (32%) OL: Verwae, Mabry (28%) K: Leonidas I (28%) DL: Armstrong (32%) LB: Maclean, Redd (32%) CB: LeBayers (28%) As far as I could tell, the worst player to appear on any championship roster at season's end was Kairo Knight, who was ranked #131 in value. Waiver moves Next, let’s explore if there is a correlation between waiver activity and group finish. Going into this, I assumed the group winners would be the most likely to be actively making moves throughout the season. This was mostly true, though weirdly users that finished 3rd were nearly as active as group winners were. Last place finishers were by far the least active on waivers, with 18 out of 25 not making a single move. On top of that, there was not a single last place finisher that used all of their available waiver moves. 2nd, 4th, and 5th place finishers all averaged about 1 waiver move. Intuitively, the top scoring groups also generally had the most waiver moves. The top 3 groups averaged nearly 11 waiver moves per group compared to the overall group average of 6.84. Did Draft Position Matter? Since the large majority of points come from players that were drafted, let’s now look into draft positions of group winners to determine if there was a position more likely for winners to be drafting out of this season. From the looks of it, you could find success from any of the first 4 picks, but it was rare for a user with the 5th or 6th pick to end up winning their group. On the opposite side of things, the 6th pick was the most common draft position for group losers. It’s hard to draw any concrete conclusions from the data, but it does seem like you’d want to pick as early as possible if this season’s results are any indication of what will happen in the future. Looking Forward So we’ve recapped the top players and looked at what a normal group winner does, but how can we use this information to make informed decisions in next season’s draft? First, and in my opinion most importantly, you should be thinking about what positions to draft when. Predicting who will lead their position is important, but picking the number 1 CB is a lot less valuable than picking the number 5 RB. How do we know which positions should be attacked first? Personally, I’m looking for two things: Value Predictability Of course, getting the most value out of the pick is the end goal and the only thing that matters. We want to draft the player that is going to give us the most points compared to the other players at their position. That being said, some positions are much more difficult to predict than others. For example, Patterson was the 2nd most valuable QB this season but (on average) was drafted behind 5 other quarterbacks and even went undrafted in 9 out of the 25 groups. While he was the 23rd most valuable player this season, worthy of a 4th round pick, it would have been difficult to predict his performance. If we can confidently predict where players at a position will finish, it makes it more advantageous to try to draft it earlier. With that in mind, let’s take a position by position look at last season’s data. QB Value: B Predictability: C Outside of the FLEX positions, the top QB provided the most value in S47. Not only did Thor provide the 19th most value overall, he was usually drafted 1st at the position, meaning his top finish was predictable. Once we get past Thor, however, things get messy. Patterson and McRobinson both provided a solid amount of value, but neither were drafted in the top 3 at their position on average. Additionally, the only other QBs to provide value after that were usually not even drafted. There was value to be gained at this position if you could grab Thor, but if he was off the board you were probably better off waiting. RB Value: A+ Predictability: A As we’ve touched on before, elite RBs are by far the biggest difference makers for ISFL fantasy success. On top of that, they are (mostly) predictable. 4 of the top 6 RBs were also in the top 6 by ADP, with Skywalker being drafted top 10 as well. Swaggert and Mediocre are the biggest outliers in this group, but for the most part we are pretty good at predicting this position. Picking a highly ranked running back with your first pick was the best strategy last season and will likely continue to be for the foreseeable future. WR/TE Value: A Predictability: C WR/TE is much deeper than RB, and it still does provide a substantial amount of value if you can snag a top 10 player. This is especially true if you identified one of the top 3 tight ends, as they occupied 3 of the top 4 spots on this list. However, this is much easier said than done. ADP for the top players at these positions was all over the place. Calhoun was the only player drafted in the top 5 at the position to actually finish there, and only half of the top 10 in value were drafted there. DL Value: D Predictability: D There is not much variance in defensive line scores, and the top player at the position was only drafted in 3 leagues. I would not waste my time trying to get an advantage here. LB Value: C+ Predictability: B- The top linebacker provided a solid amount of value this past season, and 3 of the top 4 finishers at the position were drafted in the top 4. It’s probably easier to predict than most other defensive positions if this past season was any indication. DB Value: F Predictability: F Unfortunately, DBs just don’t really make a difference with the current scoring system. The top player at the position only produced 19 more points than the 6th ranked DB and, similarly to the DL leader, was not drafted in over 50% of groups. Just going by the numbers, you should address this position last. OL Value: C Predictability: A It’s not a very exciting position, but there is actually a decent amount of value to be gained here and it is pretty predictable. Of the top 6, 5 of them were drafted there. Mabry was the exception, but that’s also because most drafts finished before the Baltimore trade happened. That being said, the top 5 were all fairly close so you don’t necessarily have to be the first one to pull the trigger. I’d address it before DB or DL if possible, though. K Value: C Predictability: A There’s not a ton of value to be gained or lost at the kicker position, but we are mostly alright at predicting it. Similar to OL, snagging the top option at this position can make sense if you’ve filled the more valuable slots already. Draft Plan As you can see, with only 6 players being rostered at most positions, there isn’t always a ton of value to be gained by drafting the right one. Of course, there will be likely be an outlier or two that will jump up a value tier so feel free to try your hand at being Nostradamus but if you wanted to go into your draft with a plan of attack, I’d suggest the following tiered priority list: Tier 1: RB1 This should be pounded into your head at this point. With your first pick, you should be drafting your RB1. There really isn’t any other acceptable option. Tier 2: RB2, WR 1-3, FLEX At this point, you should be filling the rest of your skill positions with the best pick available. 27 of the top 36 players in value last season were either RB, WR, or TE so I wouldn’t recommend looking at other positions until you’ve filled these unless you have a really strong feeling about a player (Thor at QB sticks out as maybe the lone acceptable deviation here). Though 5 of the top 12 players were WR/TE, I’d still probably lean towards filling both RB spots first since it’s more predictable and you could potentially snag 2 of the top 6 if another manager makes a bad selection. Tier 3: QB, LB QB would make the most sense to attack after you’ve filled your FLEX spot, but it was somewhat hard to predict last season. The other position with a similar level of value added is LB, so I’d try to fill these next if you think one of the top options is still on the board. Tier 4: OL, K Both OL and K provided decent value and were easy to predict, making them a higher priority than positions like DL or DB. Tier 5: DL, DB In the current system, I can’t imagine a scenario where you’d want to pick either of these positions before the last 2 rounds (unless you are the last one in your group that needs to pick a different position). Both are fairly unpredictable and don’t provide much value even if you do guess right. Just pick the highest ranked guy when you finally have to fill this position and hope for the best. It probably won’t matter. Outliers If you are drafting based on our methodology above, I believe you’d be putting yourself in a good spot to take home some fantasy TPE at the end of the season. But let’s see if we can take this a step further and develop some tricks for identifying outliers. For the most part, player ADP correlates pretty closely to value added. However, there are some clear outliers on the chart. Players like Swaggert and Hogmally had league winning upside despite being drafted in the later rounds. I specifically looked at the following surprise players to try and see if I could discern any trends: Willie Swaggert Mister Hogmally Pitter Patter Purdy Mediocre NCADV RAINN This may not hold true every season, but there two team variables that seemed to correlate with producing fantasy outliers. No WR3 If my research was correct, there were only three ISFL teams to not roster a WR3 this season: Colorado, Arizona, and Sarasota. Two of the three ended up producing 4 of the 5 outliers in our sample (Mediocre, Paris Johnson, Hogmally, and RAINN). It might be worthwhile looking for skill players on teams without a WR3 that might be able to step up due to that absence. Bad Defense Logically, it makes sense that teams with bad defenses would have more reason/opportunity to air it out and accumulate stats. Berlin, Colorado, New York, and Honolulu all had bottom 4 defenses in both TPE and PA, with Berlin and Colorado being able to produce outliers this season. Guys like Pitter Patter and Willie Swaggert ended up being big producers despite not having high TPE. Once again, 4 of the 5 outliers in our sample had a bottom 4 defense. It doesn’t matter as much how efficient you are when you always have to score on offense to keep up. Overall, drafters were mostly pretty good at predicting who would be top scorers, but I’d recommend thinking about this criteria when you’re looking to fill that final FLEX spot. That’s All Folks For the few of you that stuck with me through all of this, I appreciate you! Thank you to @CROney3 for providing me with some previous research and documentation he used as it saved me a lot of time when calculating ADPs. Thank you to @TubaDeus for giving me access to all of the draft sheets so that I could pull the ADP data! And thank you to @LtHudz for backing off and letting me win group 9. Also, I think fantasy articles get 2x? Idk, the media graders get paid enough to figure that out for me. Until next time, the Doctor is out. RE: Doc's S47 ISFL Fantasy Season Post-Mortem - UptownCord - 05-01-2024 I for one fully support this Reginald Shrubbery propaganda. RE: Doc's S47 ISFL Fantasy Season Post-Mortem - CROney3 - 05-01-2024 Holy forkin shirtballs RE: Doc's S47 ISFL Fantasy Season Post-Mortem - Pvtpenne - 05-01-2024 Wow, great stuff RE: Doc's S47 ISFL Fantasy Season Post-Mortem - Baron1898 - 05-02-2024 you should be tremendously proud of this RE: Doc's S47 ISFL Fantasy Season Post-Mortem - JoeSteel - 05-02-2024 Where are the Safeties RE: Doc's S47 ISFL Fantasy Season Post-Mortem - Wiggli - 05-02-2024 Awesome work. I joined a little late and didn't get oriented early enough to join fantasy, which I think is a shame. It looks like a lot of fun and it would have kick started learning about the players more. A lot of great insights about performance in fantasy that should be useful to everyone in the future, appreciate the effort.