Last year I developed a new metric to evaluate teams called Net Defensive Points (NDP). Simply, NDP looks at all games a team plays and then calculates the defensive FPL points a defender would have earned if they played every minute. For a clean sheet, 4 points is added, for every 2 goals conceded in a single game, -1 points is added. This gives an average defensive return rate that ignores player rotation, early/late subs, and other point sources and sinks (Cards, attacking returns, etc.) and evaluated teams’ defensive FPL point potential on an even field. Because it uses game scores only, it can also be done for past years with little difficulty. I have NDP for the 2013/14 season onward and wanted to look at 2018/19 and see if there is anything we can learn from it to help us in FPL going into 2019/20.
One more piece of information to note, I have broken teams into two categories:
- Big Six: ARS, CHE, LIV, MUN, MCI, & TOT
- Little 14: Everyone else.
With the exception of the occasional outlier, this division seems to group teams fairly well and provides a balance between small enough groupings to make the data usable, but also large enough groupings to make it significant.
Total Net Defensive Points
Let’s start off with a plot of NDP based on the final standings of the 2018/19 season. This chart includes the total NDP earned by each team and breaks them out by opponent type (remember that there are fewer games played vs Big Six clubs).
Over the season, LIV and MCI were the clear standouts with CHE, TOT, and EVE the only teams close with almost half as many defensive points. FUL and HUD ended up losing more points through conceding goals than earning points by keeping clean sheets. It is also worth noting that (if you ignore ARS’s and MUN’s 1 point) only five teams managed to earn defensive points against Big 6 sides. Finally, CAR ended up eleventh overall for NDP earned, but this was not enough to keep them out of the relegation zone.
Basic Defensive Points
The next graph adds in points for minutes played and starts to better reflect the base points a defender would earn from each team (I have called this Basic Defensive Points, to differentiate it from NDP). Adding in the minutes played points also makes all the values positive and easier to compare. To convert back, just subtract 2 from the totals. The figure below shows the same data in both charts, but the left one is sorted by final league position while the right is sorted best to worst by Basic Defensive Points per game.
There is a lot to take away from these charts, so let’s get started.
- Only three teams (MCI, LEI, and WHU) managed more BDP per match against Big Six teams than against Little 14 teams. EVE and MUN very very close as well
- ARS and MUN had a very poor defensive season finishing 10th and 12th for BDP/game, respectively
- EVE fared the best of the Little 14 teams finishing 4th overall.
- Teams like CRY, NEW, and WOL provided over three points per game against Little 14 sides and would have been decent options to play for favourable fixtures.
Now looking back six years, below are the average Net Defensive Points per game for the Big Six and Little 14 teams for each season.
After several years of increasing defensive returns, the Big 6 teams had a downturn. The Little 14 teams have been less consistent year to year, but returned the fewest points in the six year history. To look at the contributing factors, I plotted each team’s BDP from 2017/18 (x-axis) versus 2018/19 (y-axis). The left chart shows the season total and the right chart shows the BDP in games versus Little 14 teams. In both charts, dots above the line indicate an improvement from in 2018/19 while dots below the line indicate a decline. The relegated and promoted teams were combined based on their BDP rankings.
From what the figure shows, LIV and MCI both improved and moved into a commanding lead. CHE and TOT saw a sight decline, but are still in line with last season’s returns. ARS was a decent dip in returns, but MUN had a massive fall from top of the league at 3.76 points per game to twelfth and 2.29 points per game. Removing MUN and averaging the other five Big Six sides results in 3.39 PPG in 2017/18 (a drop of 0.06) and 3.40 in 2018/19 (a rise of 0.19). With the lower NDP returns for TOT and ARS, it isn’t fair to blame MUN for the entire decline, but it does show what a single team in a small data set can do to averages.
The little 14 teams were mostly uninteresting with EVE improving following a strong finish to the season while HUD and BUR returned less, the latter after a stellar 2017/18. WOL and CAR performed well as promoted teams, while FUL returned the fewest defensive points in the league. Casting our gaze back further, here is how the teams performed in the last six years.
Ignoring the big 6 teams for now, there are still some interesting takeaways.
- EVE have consistently been near the top and seem to switch between good and great seasons
- LEI have remained consistent over the years (with a standout title winning season)
- NEW have been slightly improved in the last two years after a year in the championship.
- CRY have been up and down defensively, and 2018/19 was an up year.
- SOU have been poor in the past two years following four great years (missing Virgil van Dijk?)
- BUR had their worst season to date following their great 2017/18.
- BOU have remained consistently poor
- WHU and WAT have been slowly declining over the years.
Getting back to the big 6, there is the same data, but for those six teams.
- MCI and LIV both improved and are returning some significance points.
- CHE remains consistent (outside of the one bad year), but slowly declining.
- ARS and TOT have both been in decline for a couple years. It is worth noting that the decline for TOT is due to their home form. Being in the new White Hart Lane this season may give them a bit of a boost.
- MUN. Last season they were joint top with the highest BDP to date. This season that returned the fewest points of any big 6 club in the last 6 years.
For some additional information, below is the same chart with years of new manager hiring identified with a red border.
CHE is a bit of an oddity, but outside of them, defensive returns tend to increase in the years after a new manager. However, with the small sample size, it might be driven by other factors. That said, it will be interesting to watch ARS and MUN and see how 2019/20 shakes out.
Home and Away Games
So far we have primarily discussed the impact of games against Big Six and Little 14 clubs. There is one more important factor to consider, home and away matches. The BDP from the 2018/19 season are provided below split by opponent type and location of game.
There is a lot of interesting things here, but we do need to be careful with this data because our sample size is getting smaller and smaller. The games against the Big Six clubs represents only 5 or 6 data points and is much more suseptable to variance. For example, EVE recorded 4 clean sheets at home against Big Six teams in the last 10 games of the season. They lost 2-0 to MCI three weeks before this run and 6-2 to TOT in December. I don’t know exactly what caused this impressive run at the end of the season, but I would be careful assuming it will continue through all of 2019/20. That said, some things to note:
- While ARS were poor defensively on the road, their home returns are quite good, regardless of the opponent.
- TOT, CRY, WOL, BUR, and BHA are all examples of teams with three tiers of results: great returns home to Little 14 teams, good returns away vs Little 14 teams, and poor results vs Big Six clubs.
- MCI had the best BDP for a single fixture type, away vs Big Six. They lost 2-0 to CHE, but kept four clean sheets in the other fixtures (The first in GW1 and the last in GW35)
The main thing we can learn from this data is that there will be defensive assets that are fine to play regardless of the fixture, and assets that will provide value if you can only play them in the favourable matchups.
Net Defensive Points and FPL
So, how does all this information impact your FPL?
When choosing your team, There is much more to consider than net defensive points. You also into factor in price, attacking returns, saves, number of team slots, fixtures, and others. NDP also provides a look at what HAS happened, but not WHY or what WILL happen.
So what am I doing with the information?
- I look for past consistency
- I avoid uncertainties (new managers, chance for “bounce back” season, etc.)
- I consider the price and attacking potential of individual players
- I look at the fixtures and potential rotation options
- I listen to the information the community has
- I stay flexible
In short, the NDP data provides a foundation in which I build by FPL strategy. It is helpful to provide stats and avoid biases, opinions, or false memories. While it changes every season based on a huge number of factors, being able to look at the results of the season in context of past performance both by that team and other teams. As the season goes on, it can also be a metric to evaluate a team’s performance vs your expectations.
I have already been using NDP in a number of analyses (most available through my twitter account) and will continue to look at throughout the preseason and into next year. I hope you have found the information helpful and I’m always happy to discuss this article or related topics, so hit me up if you have any questions or comments.