*Draft Pick Analysis - 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: *Draft Pick Analysis (/showthread.php?tid=16556)

*Draft Pick Analysis - Modern_Duke - 11-19-2019 The Premise: I wanted to find out whether or not the draft usually goes the way it should. Are players picked in the right order? How often do players fall through the cracks? How deep are the cracks (lol)? Or, essentially, do GMs know what they are doing? The Question: How should that be measured? I'm not going through and giving a draft grade to every draft pick ever. (It's 1.5x, you say?...No, I won't do it. That's ridiculous. Good day sir.) So what am I try to find out? How do you determine if a draft pick was successful? There's no catch-all stat like WAR for football. Looking at stats doesn't work since there are so many different positions. The Options: The way I see it, there are two options that are easily (well, this wasn't exactly easy, but still) tracked and apply to every position. One is TPE. TPE should correlate well with on the field performance, plus it obviously measures how active players are. The Issue (Part I): I just said TPE should correlate well with on-field performance, but if you're familiar with the league, you know that "should" is kind of the key word there. Also, TPE is fluid. Though I suppose TPE at a player's peak would do well. Plus, the TPE Tracker handily keeps track of Peak TPE. But, not everybody is in the TPE Tracker (there is a Retired Players page but it doesn't look like every player ever is on there). The Options (con't): The other option I thought of is Pro Bowl appearances. It applies to every position, and it obviously directly relates to on-field performance. It would have been hard to compile everything, but luckily I remembered JuOSo did just that last month (http://nsfl.jcink.net/index.php?showtopic=16205&hl=%22pro+bowl%22). The Issue (Part II): Pro Bowl appearances can be fluky. Pro Bowls don't matter. A player who made 1 career pro bowl in Season 1 and then went inactive is deemed to be a successful pick? The Choice: Peak TPE is probably the way to go. I didn't use Peak TPE though. Because: 1. It seemed like more work and 2. I had already started putting this all together with the pro bowl thing when I thought of it. The Method: Yeah, I literally went through that list of every pro bowler ever and every draft ever and noted when each pro bowler was selected (I found some shortcuts to improve the process, but like I said before, this wasn't that easy). If you listen closely, you will likely hear a faint, fair-off voice yelling "NEEERRRRD" right now. The Problems: -Typos (on the draft pages, on the list of pro bowlers, my own, etc.) made some players hard to find. Speaking of... -Hard to find players - of the 300+ pro bowlers, I couldn't determine where about 20 of them were drafted. They just don't seem to exist on the draft posts. Who are these mystery men? From what I could tell, I think in the early seasons some people who were probably late sign ups may have just gotten picked up on waivers and were never actually drafted? Frankly, I'm not being paid well enough to get to the bottom of it. -JuOSo's list only went through Season 17. I actually didn't even look if Season 18 pro bowl roster was announced yet, but either way I'm not changing anything now. Sorry, Season 18 breakout players. The Results: Ok, so what I did was run the numbers on each draft slot, and find out how many pro bowlers were selected there. So, there have been 17 players selected 1OA (I know, it's actually 18. Shut up, I just covered that like 2 seconds ago). 15 of those players made pro bowls, or 88.2%. This goes all the way down to pick #122, which is the lowest draft pick to result in a pro bowl appearance that I found. So #122 resulted in 1 pro bowler out 17 drafts is 5.9% (oh no I just realized a big flaw in this. Not every draft has 122+ prospects. Draft pick #122 is probably like 50% or 33% pro bowlers.....Ok that would actually be a relatively simple fix, but I don't feel like doing it. Deal with it.) The Premise, Again: Right, so let's remember why I was doing this in the first place (other than money), to see if draft picks are made smartly, or something like that. So, in theory, early picks should have a very high chance of becoming a pro bowler, and low picks shouldn't. If you were to graph the likelihood of success (oh don't worry - bet your ass I got graphs coming), it should look like a steep slope that flattens out or some kind of cool waterslide. If instead, GM's don't know what they are doing, it will look more like, I don't know, a bumpy staircase? Um, anyway here are the results for the top 10 picks and then the whole results -- in graph form! #1: 88.2% #2: 76.5% #3: 82.4% #4: 76.5% #5: 64.7% #6: 76.5% #7: 64.7% #8: 64.7% #9: 52.9% #10: 52.9% The Graph: The Notes: -That looks about right. -I mentioned above 15 of 17 1OA picks have been pro bowlers. The two that missed are Season 8's Rickey Ramero, and Season 17's Ashley Owens. Here another flaw in the method should be clear; younger players haven't had as many chances to make pro bowls, obviously. Ramero played for a while and from what I saw put together some decent seasons. I guess he just had a breakout season. Owens, meanwhile, has tremendous things in store for an absolutely loaded Yeti team. -Picks #18 and #19 is a mini sweet spot, with 13 pro bowlers. That's more than picks #13 and #14. -Third rounders (#21-30) don't seem to be much better than fouth rounders (#31-40) - 27 vs. 25 -Highest spot with multiple pro bowlers is #93, in Seasons 1 and 2. Which brings me to my next point The Refined Results: A lot of this data reflects early season selections. This does make sense. Like, it was easier to make a pro bowl for S1 players, considering there was no literally no competition for it. Of the 300 or so pro bowl players, 142 are from seasons 1 to 3. (Season 12 had the next most pro bowlers, with 22). In fact, every pro bowl player selected past #54 came in the first two drafts. Basically, what I'm getting at is, do the results change if excluding the early seasons? Also an unintentional result of this question is, have GM's gotten better at drafting over time? The Graph Part 2: The Notes Part 2: -The red line represents just the 10 most recent drafts S8-S17. -This looks exactly the same, doesn't it? -Well actually, check out the first couple of picks. Looking at these 10 drafts, 1OA has 2 misses (discussed above), but 2OA is only 6/10, and then 3OA is actually 9/10 (the only missing one so far is S16's Lazslo Fourty-Two, who has tremendous things in store for an absolutely loaded Yeti team.) 6OA is 8/10. -But yeah, other than that it's pretty much the same The Deep Dive: Ok, so that covers 'success rate'. But does that tell us everything? Like I alluded to earlier, a player who made 1 career pro bowl and then went inactive is deemed to be equally as successful a pick as an 8-time pro bowler? Luckily, I also have the numbers for how many pro bowls each pick went to. Let's take a look at average number of pro bowls per selection: The Graph Part 3: The Notes Part 3: -That looks about right too. -First overall pick makes, on average, 3.35 career pro bowls. No pressure, Bjorn. -Actually second overall makes more pro bowls, with 3.53 on average -Third highest is actually #6 overall, 2.82 -Other random blips (well, not random, there's perfectly good reasons) include #18 overall with on average 1.47 pro bowls and #23 with 1.06, the highest pick with an average over 1 The Fatigue: At this point I went even deeper into the numbers and kind of got lost in what I was even trying to find anymore. So, yeah. That's about all that's worth sharing. The Verdict: Do GM's know what they are doing? -Generally, sure. What did we learn from this? -Hardly anything. Was doing this worth it? -*checks* Only 1400 words?! Probably not. *Draft Pick Analysis - run_CMC - 11-19-2019 This article is such a pleasant read and frankly I just love it lol I'll get that pro bowl soon! (one hopes) *Draft Pick Analysis - Dylandeluxe - 11-19-2019 The commitment to this is astounding *Draft Pick Analysis - JuOSu - 11-20-2019 Love articles like this, happy you used my data too haha. Fantastic work! *Draft Pick Analysis - majesiu - 11-20-2019 Great article. #3rdOAGang *Draft Pick Analysis - infinitempg - 11-20-2019 that graph is such a satisfying exponential decay. great article! also if you wanna use peak TPE you can probably get the JSON file from @Eco's tracker! *Draft Pick Analysis - manicmav36 - 11-20-2019 Awesome work! I would assume that #18 is abnormally high in part due to Owen Taylor, but who caused #23 to be so high?