Game of Stones Season 3 Mid-Season Blog 2 - Stats Special!
Hello all! Figured I'd pop in to do some typing during this international break. Keep the fingers warm for April, etc.
This week we have a super cool topic on our hands: stats! Yep welcome to the Gameweek 30 stats special!
Basic Stats
Table time! Starting off with some basic average, max/min, etc. stats from each season.
| Season | Average | High (Single Gameweek) | Low (Single Gameweek) | Average (Season) | High (Season) | Low (Season) |
|---|---|---|---|---|---|---|
| 1 | 32.92 | 85 | 4 | 1251 | 1621 | 1024 |
| 2 | 37.34 | 95 | 4 | 1418 | 1912 | 1108 |
| 3 | 34.39 | 84 | 4 | 997 | 1177 | 672 |
| 3a | -- | -- | -- | 1307 | 1537 | 879 |
For row 3a, I multiplied the weekly average by 38, added 360 (9*40) to the season high, and added 207 (9*23) to the season low. This gives us a good approximation of what the final numbers for Season 3 will look like, since MatthewR and Matt are both around that 1177 mark and 40 points per week and Andrew is currently averaging 23 points per week.
Okay, well what does this mean? Well to start, it's hard to compare Season 2 with Season 3, since Season 2 only had 12 managers. So, some notes:
- This season is overperforming Season 1 by just over a point per week
- Very similar highs and lows on a single gameweek basis in Seasons 1 and 3 (and 2 isn't far off)
- Much lower predicted low for Season 3 (sorry Andrew)
This doesn't really tell us that much, so let's dive in a little deeper and look at it at a manager level. Starting with average points per gameweek:
| Season | 20-25 | 25-30 | 30-34 | 34-37 | 37-40 | 40-50 | 50+ |
|---|---|---|---|---|---|---|---|
| 1 | 0 | 4 | 7 | 2 | 1 | 2 | 0 |
| 2 | 0 | 1 | 1 | 5 | 3 | 1 | 1 |
| 2a | 0 | 2 | 5 | 3 | 1 | 1 | 0 |
| 3 | 1 | 2 | 3 | 4 | 4 | 2 | 0 |
For row 2a, I cut 3 points off the numbers from Season 2 to mimic if the season was played with 16 managers instead of 12 and the average was more in line with Season 3. This is a really interesting table to me, because it shows us a few things:
1. The managers in the league are getting better over time
Even comparing OG Season 2 vs Season 3, we see as many managers averaging 34+ in a league with 4 more people. Let alone the adjusted Season 2b, with 10 managers averaging 34+ vs just 5. In a purely numerical analysis, it looks to me like the 3 managers under 30 points per week are influencing the average more than initially thought.
And in fact, the median is almost a full point higher, at 35.16.
2. There is more parity at the top
Looking at Season 1, there's just 1 manager sitting in the 37-40 range behind the two leaders at 40+. And then behind them sits just another pair in the 34-37 range. In Season 2, there's a manager at 50+, trailed by a single manager in the 40s. Behind them, only 3 more in the high 30s. This season, our leaders in the 40 range are chased by 4 managers in the high 30s.
The "trailing pack" has grown from 1 manager, to 3 managers (well, this is debatable), to now being 4 managers deep. Maybe one day we'll get to the point where we have a proper title race. (copium)
3. Stragglers are hit hard when there are more managers
Season 2 sees a huge spike in points per week for managers sitting near the bottom. Just 2 managers were picking up less than 34 points per week. Season 1 had 11 managers under than number. Season 3 has 6!
Maybe I'm overthinking this a bit, but my intuition says that managers who see their team do well, will want to play the game more. Make more waivers, set their lineups, read the updates. If managers can blame 'oh I had a bad matchup but my team did well,' losses hurt way less.
Which means: this stat looks to me like a huge indicator that going from 16 -> 12 managers is the right call. Averaging less than 30 means that you're getting 1, maybe 2 returns in a week, which just isn't as much fun. There needs to be a balance between difficulty picking up players and earning points.
The PF vs Points Conundrum
For this section, I'm going to try to tackle the age old question of if Points For actually corresponds to Points. To do so, I made a chart! It's pretty ugly but shows there is a pretty consistent curve we see season to season.
Note that Season 3 is using calculated PF and P based on current manager's average points per week in each category.
All Data
Sorted by Season (uh, backwards)
And since these curves all look pretty similar, we can extrapolate a Points For per Point value for each of these seasons:
| Season | Points For per Point |
|---|---|
| 1 | 22.8 |
| 2 | 25.86 |
| 3 | 25.60 |
| All | 24.92 |
I'm not really sure if this is a good measure of anything - considering it has Season 3 Andrew at 55 Points For per Point - but it does reinforce the notion that one Point is worth about the same amount every season. My theory here is that this may be a more conclusive measure of luck. It isn't about how many PA you have against you in any given week, but how consistently your Points For returns Points. General rule of thumb is "lower value = luckier."
If we look at some of the more interesting points on the first graph, we can pick out some observations:
- Season 2 Matthew (me!) had the highest number of PF/P of the league winners at 23.32 - Season 1 Darryan had 18.06 and Season 3 MatthewR had 15.42
- Season 1 Andrew (18.54) and Season 1 Scott (17.89) had noticeably low values, meaning they picked up way more points than average at their PF level (65 and 64 points, respectively)
- Other managers in the 1150-1200 PF range included S1 Kevin (48), S1 Ricky (47), S1 Emily (44), S1 Joshua (41), S2 Jesse (48), S3 Coop (51), and S3 Soph (37)
- Season 3 Andrew at 55.06, which I think we can safely chalk up as an outlier
I think this is where a conclusion should go, so I think I can safely say that having more Points For does correlate to having more Points. Jokes aside, we can see some pretty major groupings, even between seasons, so I'll be really curious to see how this sort of data progresses as the seasons go by. Maybe eventually I'll get a stats degree and can put an r^2 value or something here to sound smart.
Records
For this section I wanted to give a quick update on some of our rolling records that I keep track of for this season. Stuff like highest score in a gameweek, biggest win, etc.
Highest Single Gameweek (Season 3)
| GW | Score | Manager | Opp | Opp Score | DGW | |
|---|---|---|---|---|---|---|
| 1 | 24 | 84 | Rebecca | Coop | 39 | X |
| 2 | 22 | 71 | Zach | Luke | 37 | |
| 3 | 27 | 70 | Zach | Emily | 18 | |
| 4 | 26 | 67 | Matt | Jesse | 14 | |
| 5 | 3 | 63 | Matt | Gavin | 13 |
This puts Rebecca's 84 pointer at 5th on the all-time list:
Highest Single Gameweek (All-Time)
| Season | GW | Score | Manager | Opp | Opp Score | DGW | |
|---|---|---|---|---|---|---|---|
| 1 | 2 | 37 | 95 | Matthew | MatthewR | 59 | X |
| 2 | 2 | 33 | 87 | Matthew | Darryan | 21 | |
| 3 | 1 | 29 | 85 | Amanda | Darryan | 85 | X |
| 5 | 2 | 36 | 84 | Matthew | Andrew | 26 | |
| 5 | 3 | 24 | 84 | Rebecca | Coop | 39 | X |
Okay, not gonna make these tables for all of them, but some other fun ones include:
- The "Biggest Victory" table includes 3 50+ point victories, led by Matt's 67 pointer in the 1st table above
- Both of the Zach vs Luke matchups making it onto the "Highest Combined Gameweek" list at 2 and 3, with 110 and 108 combined points. Spot #1 is taken by Rebecca vs Coop from above, of course
- "Highest Score in a Loss" is tied between Gavin and Matt, who both lost with 51 pointers
- "Lowest Score in a Win" is Emily's 16-15 win against Jesse
- Scott features on this list 3 times, as does Emily (both with 1W2L)
And we move onto some of the accolade categories! Including:
- Weekly High Scores, which tracks who scored the highest in any given week
- Matt and MatthewR have the highest with 4 each
- Kevin, Rebecca, Soph, and Zach each have 3
- Weekly Low Scores, which tracks the opposite
- Andrew has 11
- Emily has 7
- Manager of the Months
- MatthewR leads with 3
- Kevin has 2
- Emily, Luke, Matt, and myself each have 1
- Longest Winning Streak
- Kevin had an 11 winstreak from GW17-27
- MatthewR had two separate long winstreaks: 9 and 8
- Longest Losing Streak
- Andrew had a 13 loss streak from GW14-26
- Dylan had a 9 loss streak from GW17-25
Game Changers
Hopefully, y'all have enjoyed the little write-ups I've been doing here (see mid-season blog 1 if you want to read about it in way more detail), but they also come with stats!
So, I've mostly been tracking things like most appearances, total points while on the list, averages while on the list, etc. I'd love suggestions if anyone has any more ideas, but here are some of the stats!
Most Game Changer Apps
| Apps | Player | |
|---|---|---|
| 1 | 11 | Mohamed Salah |
| 2 | 9 | Cole Palmer |
| 3 | 6 | Matheus Cunha |
| 4 | 5 | Alexander Isak |
| 4 | 5 | Morgan Rogers |
Highest Game Changer Points
| Points | Player | |
|---|---|---|
| 1 | 167 | Mohamed Salah |
| 2 | 117 | Cole Palmer |
| 3 | 70 | Alexander Isak |
| 4 | 65 | Matheus Cunha |
| 5 | 59 | Erling Haaland |
Highest Game Changers Average (2+ apps)
| Points | Player | Apps | |
|---|---|---|---|
| 1 | 16.67 | Justin Kluivert | 3 |
| 2 | 15.18 | Mohamed Salah | 11 |
| 3 | 14.75 | Erling Haaland | 4 |
| 4 | 14.50 | Omar Marmoush | 2 |
| 5 | 14.33 | James Maddison | 3 |
I mean, it's mostly just the same players over and over again, but it's nice to see Morgan Rogers popping up with 5 appearances, Kluivert being so key for Zach, and somehow James Maddison is elite? (Nobody tell Roy Keane)
I also track how many points the Game Changers score total in a week (so, all 8 players), with the highest scoring weeks being:
- 112 - DGW 24
- 109 - GW 13
- 105 - GW 17 & GW 6
- 103 - GW7
- 101 - GW2
Another pseudo Game Changers stat I've been tracking is the highest scorer(s) of the gameweek. If you look at the gameweek picture in the Weekly Updates, you'll notice that no matter if the player makes the Game Changers list or not, if they are owned, they're added to the list. I don't actually ever mention it so if you're super studious and ignore the screenshots you'll miss this. So! Let's look at some interesting things from this list:
- Salah makes the list 9 times - surprise surprise - with his highest % of Rebecca's points being 64% in week 19 (16/25)
- This is also the highest % this season, with other 50+% players including Salah (GW3), Salah (GW17), Diallo (GW21), Cucurella (GW27)
- The highest non-Salah score was Palmer's 25 pointer in Gameweek 6
- We've seen the highest score of the week be a tie 7 times (with one of those times (GW4) being between 3 players!)
Hopefully I've been able to add some interesting points to this list rather that it only being about Mohamed Salah, because holy shit that guy has been picking up goals and assists at a ridiculous rate this season.
Miscellaneous
And finally, I always enjoy a miscellaneous stat dump, so here that is, broadly categorized:
Gameweek Scores
- Double Gameweek 24 was the highest scoring gameweek with a 38.69 average
- Gameweek 7 was the highest scoring non-double gameweek with 38.63 (GW17 had a 38.56 avg as well!)
- On the topic, Gameweeks 7 and 17 had 8 managers score 40+
- The highest score of the week has only been less than 50 3 times: Gameweeks 8-10
- The lowest score of the week has been 20+ 4 times: Gameweeks 7, 18, 19, and 22
Gameweek Breakdown
Monthly Scores
- The highest scoring month was February, with a 37.76 average over 4 gameweeks
- Manager of the Month has been won on PF 6 times out of a possible 8 (with 1 tie, included in the 6)
- Managers who achieved maximum points in a month yet lost on PF:
- MatthewR (3)
- Gavin (2)
- Jesse (2)
- Dylan
- Rebecca
Manager Scores
- 3 managers have scored in the 40s 10+ times: MatthewR (15), Luke (14), and Matt (10)
- The only other manager that has been in a 10-point category as many times as MatthewR is Darryan, in the 30s (15)
- Every manager has scored 50+ except for Jesse
- Matt and Zach are the only managers who have scored 50+ 6 times
- Zach is also tied for 3rd most weeks in the 20s (11)
- Only 3 managers have double digits in multiple 10-point categories
- Andrew: 10s (10), 20s (10)
- Coop: 20s (10), 30s (12)
- Dylan: 20s (11), 30s (11)
Manager Breakdown
Luck
- MatthewR is the luckiest manager with 854 points against this season (29.4 average per week)
- Andrew and Scott are the unluckiest managers, with 1109 and 1107 points against (~38 average per week)
- Gavin is the closest to playing against the average, with 1000 points against (PF average is 997)
- Luke is overperforming his luck by 9 places (13th in PA, 4th overall)
- Emily is underperforming her luck by 9 places (2nd in PA, 11th overall)
Luck Table
Expected Points
- MatthewR only leads the xP Table by 8 points over Luke, whereas he leads 17 points over Matt in the actual table
- Darryan is underperforming his expected position by 6 places (8th in xP, 14th overall)
- Jesse is overperforming his expected position by 5 places (15th in xP, 10th overall)
Expected Points Table
Conclusion
And I think that's just about all I feel like typing for this blog, so thanks for reading! I may or may not be doing another of these this season, considering it is almost April somehow, but if anyone has any requests for topics / updates on anything league-wise I'd be happy to provide (maybe in a shorter format).
In terms of book keeping, in April / May you can expect to receive a few communications (that will require action on your end), including:
- Interest form for next season
- This will also include a section for pulling in new managers for the second league, if you have any!
- End of Season Accolades Nominations
- Form for people to nominate Managers, Players, etc. for various awards, similar to last season
- End of Season Accolades Vote
- I will compile the nominations and send out a form to vote for the various awards
- Feedback Form
- As always there is the generic feedback form which you are welcome to use, but I'll also send out a more specified feedback form for this season to try to gauge what can be improved for next season!
- I'm a big fan of getting feedback and people's thoughts on things, good or bad, so definitely don't hesitate to raise things to me if there are any improvements to be made
- This form may be attached to the End of Season Accolades Vote, since I do think that is the one that will get the most traffic
That should cover it! A quick warning that I will be traveling for the last couple gameweeks of the season, so the Weekly Updates for those may be an interesting format. I'll be sure to send out something, though! Even if it is just images from my travels.
Thanks for reading and I hope you enjoyed the raucous Nations League matches today!