Strength of Schedule of the 2017 AFL fixture

As with previous years, we’ve taken a look at the AFL fixture in terms of how the lopsided match-ups impact on strength of schedule.

Here’s the summary:

  • North Melbourne have the easiest fixture while GWS have the most difficult set of double-up opponents.
  • The Bulldogs’ effort winning the premiership as 7th placed team has slightly upset the AFL’s grading of team strengths.
  • Fremantle’s fixture is anomalously tough for a bottom side and the Saints can feel aggrieved from among the middle tier teams.

And some caveats:

  • These are all based on several measures of strength of opponent, with no reference to factors such as breaks or home ground advantage.
  • These assessments will inherently be different after the season has been played out – for example, Fremantle turned out to have the hardest fixture in 2016 after being projected to have a basically average one.

Strength of Schedule

Strength of schedule (SOS) is a method used in US sport to compare results in competitions with uneven fixtures. Our post on the 2016 fixture goes into more detail, so we will keep it brief here.

The NCAA use SOS to determine entry to Bowl games and the College Football Playoff, the NFL use it to break ties for determining draft order. It is essentially the win-loss ratio of a team’s set of opponents.

In the context of the AFL, where the fixture sees 18 teams play 22 games, which results in a slightly uneven draw. More specifically, it means that each team played 12 teams once and five teams twice. HPN can use SOS to assess the deviation of each team’s draw from a perfectly average opponent set. This table of numbers is sorted from hardest to easiest fixture (red is bad, green is good):

sos1

The table shows three measures of opponent strength. The first is just based on win-loss record.

The other two use more advanced methods to account for how teams “should” have performed. Those are Pythagorean expectations approaches pioneered by Bill James and here they’re based on points scored, and on scoring shots. Matter of Stats has a thorough explainer on the topic, but for our purposes it’s sufficient just to say that the two methods account for luck.

Points for/against accounts for the inherent luck in close games, and then scoring shot accuracy factors in the way accuracy tends to mostly (but not entirely, at least with Sydney’s defence) be based on luck and subject to reversion to the mean. We think adding these gives a more accurate picture of team strength.

We can also convert the above ratings to “games advantage” or a fixture effect on expected win-loss, as shown below, to give a more intuitive view of the degree of unevenness in a 22-game season:sos2

GWS faces the strongest set of return opponents while North Melbourne faces the easiest. Effectively, GWS’ draw puts them about 1 expected win behind North Melbourne compared to a completely even draw.

North Melbourne’s only “top 6” double-up is the Bulldogs, who of course actually finished 7th (and who were a bit lucky in close games to get to 15 wins) before their improbable premiership run. This leaves North with a very soft looking set of opponents for a middle-bracket side.

The impact of accounting for luck is visible in these calculations. Teams that had good luck in close games include Hawthorn, St Kilda and the Bulldogs; the three teams that the Swans play twice. This gives Sydney a much easier draw projected on Pythagorean expectations than on actual recorded win-loss records. Similarly, the Bulldogs play Sydney and GWS twice, both teams who are among those with the biggest gain in strength on those luck-based expectations.

To get a fuller picture, let’s now factor in 2016’s uneven fixture and add the two fixture-effects together, to see the cumulative impact for 2017:

sos4

North Melbourne, who had a murderous fixture in 2016 and a much softer one in 2017, should be expected to be 1.3 wins better off just from the fixture effect. At the other end of the scale, the Saints and Giants look to be the worst off, both of whom move from below-average opponent strengths to above-average strengths.

How effectively did the AFL handicap in 2017?

The AFL’s goal in assigning double-up games is to handicap the fixture. The top 6 teams from last year play more top teams, and vice versa. This has the effect of giving top teams harder opponents and increasing competitiveness.

In a completely fair fixture where everyone played everyone else once (or twice), the only fixture effect would be not playing one’s own team. So the fact that top sides have a hard fixture despite not playing themselves signifies that the handicapping is working.

Here’s how the handicapping looks for 2017:

sos5

By deducting the “not playing self” effect, the top and bottom sides adjust their SOS ratings to some extent.

The biggest anomaly is Fremantle being handicapped harder than Port and North Melbourne, having been given almost a worst-case fixture within the bounds of the AFL’s policy. The Dockers drew the two strongest middle bracket teams (North and the Eagles), while many bottom sides get only one. They also get Richmond, the strongest bottom bracket side, and Geelong, one of the stronger top sides. This is nearly a worst case fixture – only replacing Essendon with a stronger bottom-6 side could make it tougher for them.

How did 2016 projections go?

How accurate will this all be? Last year we projected North Melbourne and Adelaide to have the toughest draws and St Kilda the easiest. Here’s how that fared:

sos3

In actuality, North Melbourne’s draw got harder by virtue of the improvement in their double-up opponents Adelaide, St Kilda, the Bulldogs.

On the other hand, Adelaide’s opponents West Coast and Fremantle fell down the ladder, opening up their draw compared to initial expectations. Other teams such as West Coast and GWS also benefited from the decline of Fremantle.

At the other end, the improvement in the Giants and Crows made a number of fixtures harder than projected. Fremantle’s -0.9 expected win change from pre-season projections is due both to those two teams as well as their own collapse changing the “don’t play selves” component of schedule strength.

By design, projecting based on 2016 cannot predict the unexpected changes in strength and fortune that happen every year. Nobody can see the next Fremantle-style collapse or the next bolt up the ladder coming.

What we can say is that the AFL’s fixture handicapping continues to, mostly, do what it tries to do. It turns an uneven fixture, forced by the ugly task of dividing 17 opponents into 22 matches, into a tool for slightly improving competitive balance.

Winners and Losers of the 2016 #AFLTrades Period

Earlier today HPN recapped every trade that happened across the free agency period. If you want to recap that thrilling two weeks of your life in about 15 minutes, that’s the place to go.

Using those numbers, we’ve calculated who are the ultimate winners and losers of the trade period. Without further ado:

balance-table

Winners

North Melbourne

North embarked on a series of low risk, high reward trades for young players who have yet to live up to their full potential. For former number 7 pick Paul Ahern, they gave GWS pick 69. For Nathan Hrovat, they gave up the equivalent of pick 81. And for Marley Williams, they gave up practically nothing. With a number of self-inflicted holes in their best 22, at least two of these three players might be able to step in and contribute right away. Regardless, the Roos pretty much can’t lose these trades because they gave nothing up in them.

The losses for the Roos during the period was that of Daniel Wells, an aging star, and Aaron Black, a player seemingly without a place at Arden Street. It is unlikely that either would have played in the next Kangaroos premiership sides. With a touch of luck, one or two of those recruited above just might. Considering their odd delistings at the end of the home and away season, the Roos did as well as they could to get themselves back on track for 2017 and beyond.

Carlton

Carlton had an extremely delicate situation to balance through the trade period, and whilst they didn’t get the biggest deal that they had on the table done (Gibbs), they did very well out of the period. With Tuohy wanting to leave Princes Park, Carlton had to try to extract as much value as they could for a player that could walk in the PSD. By getting Smedts, they get a potential replacement (if he can get fit). On paper they lost this trade, but won every other.

But the real magic for Carlton was on capitalising on the Hawthorn fire sale, and the continued pilfering of GWS offcuts. Rhys Palmer played in a final this year, and has been a very good contributor for GWS over the past few years. Getting him for nothing was a bargain, as he’ll likely get games right away for the Blues. Caleb Marchbank was in the round 1 GWS side this year before getting hurt, and just couldn’t find his way back in to a dominant squad. And Jarrod Pickett did well in the NEAFL in 2015 before being hurt for most of 2016.

Fremantle

After suffering a fall from grace last year it was clear that Fremantle needed to get some reinforcements if they were to hold any finals aspirations next year, and in this period they did just that. The Dockers finally brought Cam McCarthy through the door, at a much lower asking price than last year. Then they got a durable 23 year old triple premiership player (albeit a polarising one) for a second round pick.

At the end of the period they paid slight overs for Joel Hamling, but if he kicks on it will be a very good gamble for little outlay. Finally they grabbed a 23 year old KPF for very little (Shane Kersten), in perhaps their best deal of the period.

Fremantle also bet on themselves being better next year than they were this year, by trading their 2017 second rounder to Gold Coast for pick 35 this year.

The big out during the period was Chris Mayne’s departure to Collingwood, but the compensation pick that they got in return should cover the loss (it was on traded for Hill). Last year Fremantle lacked for tall targets and run, and these trades addressed both issues for very little in the way of costs.

St Kilda

At HPN we firmly feel that Jack Steele is a star in waiting, and that the HPN values underestimate his potential. Regardless of that, the Saints paid a pretty fair price for the Belconnen product. The Saints then committed daylight robbery on Hawthorn, in a frankly confounding trade. Finally, St Kilda paid very little to lure Koby Stevens away from the west of Melbourne, in what could be a massive steal. A very calculated trade period.

Losers

GWS

HPN technically has to call GWS losers, but in reality they did exactly what they wanted to. With such a talented list players are bound to want to find new clubs for more senior opportunities, and by getting enough picks for academy points, they’ve set themselves up for the next cycle. If there’s one criticism of the GWS trading strategy it is that they don’t quite extract full value when trading picks out, and that they don’t try to get the most for their exiting players. That said, given the amount of assets they have they don’t really need to.

GWS probably paid overs for Deledio, but the Giants can only play 22 players at once, and making sure that they maximise their premiership opportunities is their key goal. Deledio helps them do that where a 2017 first rounder probably doesn’t.

Western Bulldogs

The Bulldogs are in here as losers, but they really weren’t involved that much. The Dogs got rid of three fringe best 22 players (Hrovat, Stevens and Hamling) for a bit less than they were probably worth, but then they brought Travis Cloke in for a bit less than he should be worth. All in all, it’s hard to criticise them too much considering they just won a flag for the first time since decimal currency was introduced.

Hawthorn

Hawthorn not only offloaded two solid veterans (Lewis and Mitchell) for essentially nothing, but they basically sold the farm for just two players (O’Meara and Mitchell). They are completely out of the 2016 draft, and they are almost out of the 2017 draft. There appears to be no plan B here.

We analysed the series of loss-making moves Hawthorn made to convert their existing assets into O’Meara. The short version is that after getting Mitchell a bit cheaply, Hawthorn spent the trade period making loss-making exchanges to convert their second (36), third (54) and fourth (72) round picks, Brad Hill, and next year’s first round pick (~14), into O’Meara.

After the signing of the new CBA, every club should have excessive TPP room, so clearing contracts for TPP purposes doesn’t ring true. Vickery is a handy acquisition, but Hawthorn seemingly got the worse end of several deals here. A mild shock.

The 2016 Trade Period in Graph Form

This graph breaks down the net trade balance for each component of the trade period in bar form. Each 10 points roughly equates to a game (adjusted for elite weighting). For example, GWS lost nearly 950 games of future player production in this trade period.

Carlton, Fremantle and North win big for gaining future player value, GWS got the most current draft pick value and Gold Coast concentrated on future draft pick value. Also note how Port has essentially eschewed the 2017 draft in favour of the 2016 one.

net-trade-balance

For a more holistic look at where each club stood at the end of 2016, the next chart is for you.

stock1

Player stock is an estimate of future game potential of the 2016 playing lists. Total stock adds the draft picks clubs had before trading began. Note how far ahead of the pack GWS is – with only Gold Coast in sight. At the other end of the scale, Hawthorn, Geelong, Fremantle and Carlton appear to be aging quickly.

Finally, this graph shows how well each club did out of the trade period.

stock2

Despite their massive loss through the 2016 period, GWS sits comfortably ahead of every other club in future game stock, illustrating the sheer quantity of assets they hold and why they’re fine letting them go cheaply.

Hawthorn, sitting at the bottom already, somehow further decreases their future game stock in the desperate pursuit of one more flag before their window closes.

Every trade that happened during #AFLTrades 2016

Phew! After a slow start to the trade period, the pace seemed to intensify as the deadline drew closer. As this post will cover all 39 trades, and the HPN valuations of the trades, we’ll keep the intro brief. For a full recap of all movement on an easy to read webpage, check out Draft Guru. As always:


NOTE: We do not use the AFL draft points system to rate players or picks. We instead use our own system that attempts to compare both players and picks directly, using a common currency of “quality-adjusted expected future games output“. Fuller discussions of method and theory are located here and here. Hit us up on Twitter if you’d like to discuss.


Cameron McCarthy to Fremantle

mccarthy4

Verdict: The trade made values McCarthy as having negative value considering the other assets involved. If McCarthy can regain his form from the first half of 2015, Fremantle will have a great deal on their hands.

Brad Hill to Fremantle

bradhill2

Verdict: Brad Hill is a very polarizing player it seems – more people complained with our rating of Hill than any other rating to date. At just 23, Hill looks to be a solid contributor for the best part of the next decade. HPN rates Fremantle as massive winners here (again).

Pick swap between Brisbane and GWS

brisgwspickswap

Verdict: A nearly perfectly balanced exchange of picks that seemingly gives both clubs what they want.

Jack Steele to St Kilda

steeltrade

Verdict: Both clubs receive something that they will be able to get more use out of than the other club could. St Kilda will give the talented Jack Steele plenty of game time, and GWS can get extra bidding point value out of the pick due to the 20% point discount in bid-matching.

Sam Mitchell to West Coast

smitchell2

Verdict: Did this one actually happen? Did Sam Mitchell really leave the Hawks? A club legend moves west to begin his eventual transition to the coaching box, and both sides are reasonable about the value of him. Perhaps the most reasonable trade of the period involving a player so far.

Tom Mitchell to Hawthorn

tom-mitchell-final

Verdict: Hawthorn decided they couldn’t go on without a Mitchell in the middle, and Tom is a massive upgrade to their long term prospects. Hawthorn get an absolute bargain here for the talent young player, giving up only their first round pick for a player likely to produce a lot more value than that pick would have (the pick swap doesn’t matter).

Pick swap between Hawthorn and St Kilda

stkhaw

Verdict: What were Hawthorn thinking here? Hawthorn pay massive overs in a desperate attempt to get a pick in order to facilitate the O’Meara trade. St Kilda downgrade slightly, and get two good chances this year and one very good one next to further strengthen their list.

Paul Ahern to North Melbourne

ahern-trade

Verdict: GWS cast aside damaged goods for essentially nothing, North get a shot at a recent high draft pick with basically zero downside.

Pearce Hanley to Gold Coast with Port Adelaide involved

hanleytrade1

hanleytrade2

Verdict: Gold Coast overpay for a bloke with a short shelf-life, demonstrating that they will pay a pick premium in the pursuit of established players. Brisbane win big in expected future output.

Port Adelaide move more value into this year’s draft. Contrary to popular perception, looking at the actual productivity of the two draft picks they obtain shows Port don’t really go backwards here unless they tumble a long way down the ladder in 2017.

Jordan Lewis to Melbourne

lewistrade

Verdict: The Demons get a steal as Lewis is traded for the equivalent of pick 84.

Joel Hamling to Fremantle

hamling-trade

Verdict: Fair for a player who has had one great season but not a long track record. The balance of this trade depends on how we view the risk of Hamling failing to live up to his 2016 performance.

Michael Hibberd to Melbourne

hibberdtrade

Verdict: Fair trade, but likely to favour the Demons if Hibberd’s year away from football hasn’t negatively affected him

Dion Prestia to Richmond

prestia-2

Verdict: Just fair, but Richmond’s bet on themselves rising up the ladder doesn’t change the value of their second round pick very much, so a limited upside versus the straight swap of 6 for Prestia.

Toby Nankervis to Richmond

nanktrade

Verdict: Very close to the mark for a developing  ruckman low on the Swans’ depth chart, based on his former draft pick.

The Suns and Bulldogs swap picks

dogssuns

Verdict: This is a simple 2-for-1 trade, a win-win for both parties’ draft goals.

Jarrod Witts to Gold Coast

wittstrade

Verdict: Witts might represent the Suns’ best chance at a top shelf ruckman, and this trade is slightly in their favour.

Nathan Hrovat to North Melbourne

hrovat-trade

Verdict: A new start for Hrovat. Nearly a freebie for North, and the Bulldogs do not attempt to extract much value in this move.

Shane Kersten to Fremantle

kerstentrade

Verdict: Quite clearly Geelong didn’t rate Kersten, as Fremantle steal for a song, or part-thereof.

Tuohy to Geelong, and Smedts to Carlton

tuohysmedtsfinal

Verdict: Geelong get a creative runner off half back, and Carlton get a bunch of hope.

Josh Caddy to Richmond

caddy2

Verdict: It seems that Richmond are getting a really good deal for Caddy here, even without considering the exchange of later picks.

A Richmond win, as Geelong tries to claw back into the draft.

The Giants and Swans swap picks

swansgiants

Verdict: Notionally quite unfair but Sydney only lose value they weren’t going to use.

Jarryd Lyons to Gold Coast

lyonstrade

Verdict: Lyons has been traded for the lower end of the reasonable range. Fair trade, but Gold Coast might end up ahead in the long run.

Patrick McKenna to Melbourne

mckenna-trade

Verdict: The trade nets GWS a tiny upgrade on Academy points and presumably disposes of a player who is surplus to requirements. Very clear win for Melbourne. Or more likely Casey.

Nathan Vardy to West Coast

vardytrade

Verdict: Token pick for nearly valueless player.

Travis Cloke to Western Bulldogs

cloketrade

Verdict: The Dogs should get more out of Cloke than out of pick 76.

James Stewart to Essendon

stewarttrade

Verdict: It’s a wonderful trade for Essendon, as Stewart goes to star in the Essendon story. Vertigo.

Aaron Black to Geelong

blacktrade

Verdict: Pick 92 comes after the point at which picks likely have literally zero value. Most picks at this point aren’t used, and we’re unsure if North will be the exception to this rule. Black is treated like a virtual delisting.

Will Hoskin-Elliott to Collingwood

whe2

Verdict: It’s unlikely that Hoskin-Elliott would have cracked into the Giants best 22 this season, but has a clear place at Collingwood. His best is very good – but his worst was sub-NEAFL level. This deal is all contingent on Hoskin-Elliott’s improvement. If it continues to stagnate, this is a really smart deal by GWS.

Jack Frost to Brisbane

frostupdate

Verdict: The Lions have done really well out of this deal, and the Pies leave us scratching our heads. Frost might not set the world on fire, but he should fill a role for the Lions, and right now that should be enough.

Lynden Dunn to Collingwood

dunn2

Verdict: Dunn should play every week at a vaguely AFL-standard level in whatever role is needed down back, and that’s what the Pies need. Dunn has been valued by this trade as being virtually worthless, meaning that Collingwood might have a steal on their hands if they can get even 10 games out of him.

Marley Williams to North Melbourne

williams2

Verdict: This isn’t a trade.

 The Hawks and Blues swap picks

carlhawswap

Verdict: Carlton take Hawthorn for a ride, but Hawthorn get what they need.

Rhys Palmer to Carlton

palmer2

Verdict: THIS IS NOT A TRADE!

Pickett and Marchbank to Carlton

pickbanktrade

Verdict: Marchbank and Pickett are each being valued as still worth top-ten picks, due to having been recently drafted with picks 4 and 6. They’re being let go for a lot less than that. Marchbank particularly showed a lot of promise in 2016. Carlton make big gains in isolation on this trade, GWS lose more players cheaply and gain academy bidding currency.

Jaeger O’Meara to Hawthorn

jom-trade

Verdict: Pretty fair in isolation, but this trade is part of a bigger picture. Hawthorn took resource-conversion hits in trades with Fremantle, St Kilda and Carlton to make this happen.

Port Adelaide and Sydney swap picks

portsydney

Verdict: This is perfection. We are not worthy.

Koby Stevens to St Kilda

stevens

Verdict:  The net difference in the pick swap suggests a valuation of Stevens of around 244 points, or an expected 24 games of value. This is obviously a bar which Stevens should pass handily.

St Kilda have a potential steal here.

Fremantle and Gold Coast swap picks

sunsfreo

Verdict: The Dockers get a higher pick this year in exchange for a second-rounder next year.

Brett Deledio to GWS

delediotrade

Verdict: Richmond are entering a semi-rebuild phase, and it is unlikely that Deledio will be there to help them win in the future. But GWS are already there, and Deledio will step in right away and contribute.

A big Richmond win, but if Deledio helps the Giants win a flag it will be just the cost of doing business.

TRADE: Deledio heads to GWS, not Geelong #AFLTrades

At the start of the trade period one of the bigger “bombshells” was around Brett Deledio’s potential move to Kardinia Park. After going through several permutations, the deal to get Deledio to Geelong appeared dead.

And then GWS stepped in, and used Geelong’s pick for them.

DeledioValue.JPG

HPN is still confirming the details of this trade, but we believe that the 2017 1st rounder involved is the Geelong pick. Regardless of specifics, Richmond got a great deal for their aging star, and GWS got another elite player that will help them win now.

delediotrade

Richmond are entering a semi-rebuild phase, and it is unlikely that Deledio will be there to help them win in the future. But GWS are already there, and Deledio will step in right away and contribute.

Verdict: A big Richmond win, but if Deledio helps the Giants win a flag it will be just the cost of doing business.

TRADE: Gold Coast likely plan to go into points deficit for Brad Scheer #AFLTrades

After completing the O’Meara trade Gold Coast acquired pick 10 and apparently decided they didn’t need pick 35 any more. As such they’ve moved resources over to 2017 via Fremantle’s second-round pick.

sunsfreo.PNG

The Dockers get a higher pick this year in exchange for a second-rounder next year.

Gold Coast now take picks 4, 6, 8, 10, and 73 into the draft. This has interesting ramifications for Academy bid-matching. Jack Bowes won’t last to pick 10, but assuming Brad Scheer makes it past pick 10, they’re seemingly willing to go into points deficit in 2017 to match for him.

If Scheer goes in the first round after pick 10, the value will come off Gold Coast’s first pick next year. Otherwise, they now have four second round picks including their own next year and the value will be deducted from there.

Verdict: Fair and novel.

TRADE: Koby Stevens let go for not much #AFLTrades

Koby Stevens has become a bit of a forgotten man as the Bulldogs rose to premiership success. He has found injury his biggest obstacle at the Bulldogs, missing half of 2015 and chunks of 2016 with lung, foot and abdominal injuries, after a solid run of 20 games in 2014.

stevens.PNG

The trade here undervalues Stevens’ likely 100 games of future return over the next 6 or 7 years. St Kilda have rather astutely secured themselves a bargain in the midst of the Bulldogs premiership dream. Stevens, who played VFL during the Bulldogs’ finals campaign, should find a regular spot in a less developed St Kilda midfield.

stevens.PNG

The net difference in the pick swap suggests a valuation of around 244 points, or an expected 24 games of value. This is obviously a bar which Stevens should pass handily, even at just 14 games a season.

Verdict: St Kilda have a potential steal here.