A Fresh Look at FU Stats: A Review of 2019 - Year of the Grenades

Discussion in 'ORFFU' started by fresh, Sep 6, 2019.

  1. fresh

    fresh Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,055
    Likes Received:
    1,065
    Megalo! Megalo! Megalo!

    Grab a cup of tea or a glass of whiskey, a packet of Kit Kat and make sure you won’t get disturbed for at least the next half hour if you plan on reading all this in one go. This 6 part extravaganza (might be more depending on word limits in posts) covers Gisborne FC, our home and away season, finals, club records, league records and club H2H records.

    This a little rushed because I wanted to get this out before AFL Finals had begun so there will be typos and mistakes. There may even be mistakes in my formulas but hopefully I’ve ironed all these out. Luckily none of my work colleagues have realised I was doing this at work over the past couple of weeks, they’ve only noticed the work pile up on my desk and my todo list expand. Not that they’ve offered to help me out with any of that work, selfish c**ts.

    If anything needs fixing up or you’d like to see anything else, let me know and I’ll chuck it in. But without further ado, I’d like to present to you a review of 2019….
     
    • Like Like x 4
  2. fresh

    fresh Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,055
    Likes Received:
    1,065
    Gisborne FC

    Grenades’ Best 15 in 2019
    Defence
    : Laird (Ave 96.8, 9th best averaging defender), Thomas (82.5, 34th), Pittard (79.5, 47th), Hore (70.6, 89th)
    Emergency: Talia (69.5, 96th)
    Sum of defenders averages: 329.5 (82.4 per defender)

    Midfield: Neale (121.3, 2nd), Harmes (95.1, 39th), Ellis-Yolmen (93.4, 49th), Fiorini (87.9, 61st)
    Emergency: Seedsman (72.3, 118th)
    Sum of midfielder averages: 397.6 (99.4 per midfielder)

    Ruck: Witts (99.7, 5th)
    Emergency: Campbell (52.5, 43rd)

    Forward: Kelly (103.5, 6th), Billings (93.8, 14th), Lambert (88.0, 18th), Gresham (84.1, 27th)
    Emergency: Douglas (54.7, 165th)
    Sum of forwards averages: 369.3 (92.3 per forward)

    Interchange: Brayshaw (83.3, 136th highest overall average), Petrevski-Seton (81.6, 153rd)
    Emergency: Wilson (69.0, 308th)
    Sum of interchange averages: 164.9 (82.5 per interchange player)

    Sum of best 15 averages: 1348.1 (89.9 per player)

    After a couple of trades in the preseason, @Tomster had given up Powell-Pepper, Murray and picks for Laird, Ellis-Yolman and picks. Looking back now it’s an astute trade.

    Out went Powell-Pepper who looked good enough to be at least an 85+ averaging mid this year. However, he went backwards and his 2019 average of 74.2 would have meant he wouldn’t have made Gisborne’s best 15, but instead been named as the emergency midfielder. Sam Murray wasn’t going to be missed so could been seen as steak knives in that deal.

    In came much needed defensive and midfield reinforcements.

    Neale, Brayshaw and Harmes looked like they would be the best 3 mids for the Grenades this year. How things change. Kelly could slot in nicely but with forward eligibility he’s more valuable there. Fiorini improved to an 88 point average to claim a spot at M4. After losing Powell-Pepper, a more risky backup in Ellis-Yolman came in and made the best 15. Would he get games was the question. Seedsman was likely to play, Masten had started to decline and Guelfi and Douglas were unlikely to contribute much. If anything, Ellis-Yolman would hopefully provide some cover...but he managed a bit more than that. Although he only played in the first half of the year, his 93.4 average contributed strongly towards a team on the improve. Ellis-Yolman played from rounds 4 to 11, during which time the Grenades went 7-1. The margin for these wins and losses were at least 120+ so he wasn’t a difference maker. But, when he didn’t play in the second half of the year, Powell-Pepper suited up for the Buffalos. Luckily for the Gisborne coach, Powell-Pepper only went at a 68 average in the back half of the year so even playing Talia and his average of 69 meant the coach was in front trade value wise.

    Let’s not forget that young Dale Thomas was traded in mid way through the year and provided ample scoring power for the Grenades. While Ellis-Yolman was out in the second half of the year, Thomas stepped in and ran at a 93 average. He was only good for 9 games, but at a cost of MSD pick 36 and a 3rd round PSD it was worth the risk.

    Last year’s best defensive 4 for the Grenades were Wilson (80.2), Hamling (71.5), Jones (70.4) and Talia (69.7). Summing those averages gives you 291.9 or 72.3 per player. It looked as if Laird would have another 100+ averaging season and, had he maintained an average of 105, Gisborne’s best defensive 4 would’ve scored 329 per week or 82 per player (I’m including Thomas in this purely to make it easier on myself). Effectively Talia was meant to be replaced by Laird. While those numbers roughly worked out that way, it was partly thanks to Laird and partly thanks to Gisborne picking up a Hore (pun intended) in the draft that ensured the defenders had enough troops to get through the year. Taking just 6 defenders or forwards into the year and you’re cutting it fine imo. Hore was the lucky 7th defender.

    Without Laird, the best defensive 4 would’ve had a total average of 302.2, or 75.5 per defender. Those extra 30 points per week that Laird provided equates to about 500 points across 17 rounds. Had Gisborne not benefited from that little bit extra, they would’ve had the 3rd highest total points for rather than the highest. And guess what, the Grenades won 2 matches by less than 30 points this year...the Prelim and Grand Final.

    I’ve copied in Gisborne’s scoring history for the year. Yellow bars are the 3 losses they copped. Orange line is the median score across the entire year and the pink dotted line is the 3 round average.

    What a difference a week makes. A score of 1257 in round 9 followed by 1476 in round 10. The first half of the year was fairly consistent with a drop in scoring seen from round 12 to 17. Lucky Thomas joined the Grenades. Thankfully for the coach, scores picked up across the 3 finals.

    [​IMG]

    So I’d say that Gisborne’s premiership was a combination of good trading, good drafting and also a bit of luck. Instead of getting 22 games out of him, imagine if Gisborne’s only playing ruck went down. Or one or more of the mids or forwards being out for an extended period and having to rely on players averaging in the 50s such as Guelfi or Douglas. The back half of the year seems to be an obstacle course where your best players need to dodge being struck down by injury. Gisborne sidestepped these well.

    This should probably go into the H&A section but wanted to compare Gisborne to the rest of us. The dashed lines mark the runner up, solid line the premiership team.

    upload_2019-9-6_13-33-45.png

    So for the 5 years I have detailed data for, Gisborne continues the trend of premiers having a below average defensive line but a top 5 midfield line. Gisborne also became the first team to be ranked 1st in total points scored and win the FU premiership. This year was also the first year we saw the top 2 teams ranked for total points scored faced off in the final week. Gisborne improved by 10 places in the total points scored column from 11th to 1st. This seems to have been driven by the forward line going from 11th best to top scoring in 2019 and the midfield going from 15th best to 5th best. Depth (3rd best) was still a strong point for the team. One thing that’s evident is that you need at least average depth to have a shot at winning. Serengeti’s 8th placed finish in points scored from the interchange being worst amongst FU premiers.

    upload_2019-9-6_13-52-28.png
     
    Last edited: Sep 6, 2019
    • Like Like x 3
  3. fresh

    fresh Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,055
    Likes Received:
    1,065
    Home & Away Season

    Over the years we’ve seen on average about 3 teams drop out of the top 8 each year. After 2018, I was expecting Hughenden and Gisborne to break into the top 8 and thought Cow Bay was another year away. Time to get a new magic 8 ball I reckon.

    [​IMG]

    Gisborne, Cow Bay and Banchang were the 3 teams that jumped into the top half of the ladder. Not surprisingly, Gisborne had a huge jump in average weekly team score across the 17 home and away games, increasing by 136 points from last year for an average score of 1321. In comparison, Nuytsland wasn’t far behind with an average weekly score of 1311 (+44 point improvement from last year). Other new finalists Cow Bay and Banchang saw their average scores improve by 86 and 39 points respectively.

    If 3 teams squeeze into the top 8, 3 must make way. This year it was Christmas Island, Serengeti and Groote. Christmas Island and Serengeti saw large declines in their average scores. Groote was unlucky, seeing their average score slightly increasing by 17 points only to miss the finals. The 192 point decline Christmas Island suffered was the largest drop we’ve seen in FU history between seasons, a record previously held by Blanchetown who saw a 167 point drop between years from 2017 to 2018.

    Speaking of Blanchetown, although they were mired to the bottom of the ladder last year, the team rebounded this year with an equal best 136 point improvement in their average weekly score for the year and recorded 6 wins to finish in 13th place. Good things beckon for the Bunyips.

    With 6 seasons completed, I’ve had a look at how many wins a team has achieved in our home and away season versus their ladder position after 17 rounds. The way I look at this next chart is pick a number of wins, say 11, and look down the line to figure out where a team finished, so anywhere from 3rd to 7th. Looking at it the other way, ladder position first then wins, will only confuse things.

    What we see is that a winning season of 9-8 won’t guarantee a spot in the top 8, at best a team has finished 7th and as low as 11th. With a 10-7 record a team has finished anywhere from 4th to missing finals altogether in 9th. To guarantee a finals spot a team needs to win around two-thirds of its games and finish with 11 wins. At worst we’ve seen a team with this record finish as low as 7th and as high as 3rd. Eyeing off a top 4 spot? 13 wins seems to guarantee it...although back in 2014 the top 2 teams only won 12 games and a further 5 teams won 11 games.

    [​IMG]

    Thinking about it more, the only thing us coaches have some sort of control over is who to pick week in week out to try to get the best score we can. Sure you can say wins drives ladder position but it also depends on how many wins other teams have achieved. Last year Pearcedale finished 9th with 9 wins, this year they snagged an extra win and finished….9th. So let’s look at things another way, given a team’s median (as opposed to mean) score for a year, where have they finished the home and away season?

    With a median team score of 1200, the highest a team has finished is 2nd, the lowest 15th. That’s a huge range. You can see for yourself that by roughly aiming for a median of about 1250 it gives you a decent shot of getting into the top 8. But we can see from the chart, even this median score has the potential of a team finishing in 12th (Hughenden in 2015). Think 1300 will be enough? Nope. Groote finished 10th this year with a median of 1292. While the points are outliers on the chart it shows how luck and randomness play a part in SC and our league.

    [​IMG]

    Our league average of about 1200 points per team per week (80 points per player) has been fairly consistent over the 6 seasons. I’ve included a box and whisker plot of our league’s scoring distribution. Rather than me explaining what one is, google will do a better job of it so feel free to take an ad break if you need to. What we can see from the next chart is that 2018 had the biggest range of scores so far. On the flip side, 2019 has the 2nd lowest. What doesn’t change much is the fact that 50% of the scores in a season tend to sit between 1100 and 1300 and the median tends to hover around 1200.

    upload_2019-9-6_15-32-37.png

    Now let’s look at the same thing for 2019 by team, in final ladder position order.

    upload_2019-9-6_13-36-19.png

    In 2019 it’s evident our premiers had an extremely consistent season. A tiny 74 point interquartile range (IQR) and a low score of 1221, the only team this year without a score under 1200. Staghorn was the only other team with an IQR of under 100 with 91 points.

    If you have a close look at the chart, 10th placed Groote sticks out like a sore thumb. I mentioned them earlier, their median score should’ve meant a top 8 finish. A median score of 1292, a league low range of 220 points between their highest and lowest scores for the year. But, it’s a big but and i cannot lie, an equal 3rd worst IQR of 149 points. If you’re scoring anywhere from 1188 to 1337 in 50% of your games, lady luck isn’t always going to smile, give you a wink and her number. Unfortunately for @YAD69 that meant 5 losses where the margin was 60 points or less.

    What the chart also shows is that Eden were a little unlucky to finish last. 5 teams couldn’t crack 1300 points this year, Eden wasn’t one of them. Eden had the 4th lowest median score, but their IQR sat higher than 5 other teams. If the ladder was based on this chart, Eden would probably slot anywhere in between Hughenden and Serengeti.

    In terms of total points scored during the H&A season, we saw 5 teams break into the top 20 efforts across 17 rounds. Again, Groote was very unlucky. However, it also shows how unlucky it was to declare one of Gisborne or Nuytsland as runner up after their mammoth efforts this season.

    [​IMG]

    On the flip side, 4 teams were in the bottom 20 total points scored. I hate this table, Cow Bay appears 5 times.

    [​IMG]
     

    Attached Files:

    Last edited: Nov 21, 2019
    • Like Like x 5
  4. fresh

    fresh Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,055
    Likes Received:
    1,065
    Finals

    As mentioned previously, we saw Gisborne, Banchang and Cow Bay make their moves into the FU finals at the expense of Christmas Island, Groote and Serengeti. As destiny would have it, both Gisborne and Cow Bay broke their FU finals cherry with each other. While @Tomster enjoyed himself, @fresh was left unsatisfied. @choppers passed out early in the first week which makes it, including FU finals virgin Hughenden, 6 teams yet to win a FU final. @Batfink continues to wait for the perfect opponent to experience finals for the first time while @insider has gotten lucky and has had finals action every year.

    Excluding our first year in 2014, over the 4 years from 2015 to 2018, half of the premiership teams were in the top 8 the previous year and the other 2 (Groote in 2015 and Serengeti in 2017) won it after missing finals the previous year. Gisborne in 2019 continues the trend of a team outside of the top 8 winning the big one in an odd numbered year. That of course means that since we have an even numbered year next year, one of the teams currently in the top 8 will win it in 2020.

    [​IMG]

    There wasn’t much separating Nuytsland and Gisborne over the 3 weeks of finals. Nutysland scored 3996 points and Gisborne 3973. Blue bars in the below chart indicate premiers.


    [​IMG]

    The 23 point difference (blue bars) was the closest difference between 2 FU Grand Finalists seen. The 22 point margin (orange dot) was also the closest Grand Final we’ve seen, bettering last year’s 35 point effort between Ararat and Korumburra.

    [​IMG]

    So we’ve seen that aiming for 11 wins should give you finals hope. Once you make finals, it’s a whole new ball game. To have a shot at winning, an average score of 1300 seems to be par. Sure we’ve had 3 grand finalists with an average score under 1300 across the 3 weeks but that’s the minority of what we see.

    Continuing on with more box and whisker plots, here’s one for finals scores over the 6 FU seasons. No real surprises at first with the first week of finals having a larger range of scores, with this range gradually shrinking over the 3 weeks as the weaker teams get knocked out. The median also slowly goes up from 1276 in week 1, to 1295 and then 1305.

    [​IMG]

    But what’s more interesting is only looking at the distribution of winning finals scores. Median scores across the 3 weeks of finals: 1322, 1335 and 1333. Not only that, score under 1100 in week 1 then you’re gone. Score under 1200 in week 2, your season is over. 1300 points won’t guarantee a win in week 3 but it’ll give you a damn good shot at holding up the FU Premiership Cup.

    [​IMG]
     
    Last edited: Sep 6, 2019
    • Like Like x 5
  5. fresh

    fresh Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,055
    Likes Received:
    1,065
    Club Records

    We saw quite a few club records achieved this year. I’ve simply highlighted any reference to 2019 in yellow so that these stand out.

    A few things I’d like to point out. We saw a number of lowest winning scores by club this year. This shouldn’t be too surprising after seeing the distribution of scores this year in our league being quite low.

    We saw 4 new club record winning streaks: Cow Bay, Gisborne, Nuytsland and Pearcedale. NKOTB Cow Bay and Gisborne recorded a nice increase in their respective wins columns this year so this isn’t too surprising. We also saw 10(!) clubs record their longest losing streak this year. Blanchetown’s 27 game losing streak finally came to an end. Groote and Ararat suffered 3 losses in a row for the first time (get the violins out). Pearcedale’s record 4 losses in a row came in the last 4 rounds, and both Eden and Pakenham copped 12 losses in a row.

    upload_2019-9-6_13-38-33.png

    upload_2019-9-6_13-38-54.png
     
    Last edited: Sep 6, 2019
    • Like Like x 4
  6. fresh

    fresh Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,055
    Likes Received:
    1,065
    League Records

    Before I get into the league records, I wanted to mention that we saw a larger proportion of matches being decided by fewer points. The chart below shows % of matches in a particular margin range, blue being 2019 and grey being all FU years. Across all matches ever, 31% of matches have been decided by less than 75 points. This year that percentage is 39%. 28% were decided by a margin of 10 to 49 points, versus 18% for all matches. No wonder we saw so many club record lowest winning scores and losing streaks.

    upload_2019-9-6_13-44-43.png

    In terms of league records, it was nice to see none of the matches this year made it into the top 20 thrashings. The 423 point margin between Staghorn Eden in round 3 comes in at number 30 but things were closer this year. The Round 9 match between Ararat and Southern Cross was the 5th highest scoring match we’ve seen. We’re still yet to see a match break the 3000 point barrier.

    [​IMG]

    [​IMG]


    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]
     
    • Like Like x 5
  7. fresh

    fresh Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,055
    Likes Received:
    1,065
    Club H2H Records

    Finally, head to head records for home and away, finals and overall. Enjoy.

    upload_2019-9-6_13-45-44.png
    upload_2019-9-6_13-46-1.png
    upload_2019-9-6_13-46-20.png
     
    • Like Like x 6
  8. martyg

    martyg Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,127
    Likes Received:
    519
    WOWSERS... Just wowsers... so much data, so much effort. Well done mate.
     
    • Like Like x 1
  9. YAD69

    YAD69 Moderator Staff Member

    Joined:
    Jan 3, 2016
    Messages:
    1,668
    Likes Received:
    898
    F%$king Awesome!
     
    • Like Like x 1
  10. tyze1

    tyze1 Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,648
    Likes Received:
    1,306
    as a man who likes numbers - I can only applaud - MAGNIFICENT
     
    • Like Like x 2
  11. DamoH

    DamoH Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,022
    Likes Received:
    818
    Love it, even though this season is one to forget for CIFC.
     
    • Like Like x 1
  12. eagle_eyed

    eagle_eyed Training the house down!

    Joined:
    Jan 3, 2016
    Messages:
    1,558
    Likes Received:
    790
    Amazing insights mate! Thanks for putting this together :thumbsu:
     
    • Like Like x 2
  13. Tomster

    Tomster Moderator Staff Member

    Joined:
    Jan 3, 2016
    Messages:
    1,145
    Likes Received:
    622
    This is incredible, thanks a ton @fresh

    Interesting to note the differences and similarities between Gisborne and Nutsy. Both teams finished with similar points but I always suspected Gisborne were more consistent and Nutsy had a greater scoring range - the stats just confirm it. The GF could've gone either way realistically, but I'm quite happy with the end result :D

    Bring on next year!
     
    • Like Like x 2
  14. insider

    insider Moderator Staff Member

    Joined:
    Jan 3, 2016
    Messages:
    1,378
    Likes Received:
    867
    Thanks fresh
     
    • Like Like x 1
  15. wrightbrendan

    wrightbrendan Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    723
    Likes Received:
    645
    Incredible. Top drawer stuff @fresh
     
    • Like Like x 1
  16. eagle_eyed

    eagle_eyed Training the house down!

    Joined:
    Jan 3, 2016
    Messages:
    1,558
    Likes Received:
    790
    Nice crest mate :thumbsu:
     
  17. bryzza

    bryzza Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    1,408
    Likes Received:
    732
    Great work mate, effort is top notch ...........
     
    • Like Like x 1
  18. wrightbrendan

    wrightbrendan Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    723
    Likes Received:
    645
    Haha, thanks mate.
    Was cleaning out an old laptop over the weekend and stumbled across the old design from before the tooserious website update. Good to have an identity back for the Buff! :p
     
    • Like Like x 3
  19. JPK

    JPK Moderator Staff Member

    Joined:
    Jan 3, 2016
    Messages:
    3,852
    Likes Received:
    2,022
    Excellent work @fresh
    Some amazing stats there, an amazing read, but.... and this is just a small but.... surely you can lie.
     
    • Like Like x 1
  20. Mick

    Mick Well-Known Member

    Joined:
    Jan 3, 2016
    Messages:
    952
    Likes Received:
    368
    Unreal! Amazing work! Thanks @fresh
     
    • Like Like x 1

Share This Page