FPL Prediction Tool: Gameweek 10

Welcome again to our weekly series where we’ll be presenting you with the expected points of all the players ahead of Gameweek 10 using clean sheet and goal probabilities as calculated by the bookies, as well as our prediction tool

Our tool was presented to you at the beginning of the season so you can find the methodology and explanations of values here >> https://fplconnect.blog/2018/08/15/fpl-prediction-tool-gameweek-2/

NEW FEATURE 

Predicted assists are now included in our tool this week for the first time to give a greater all-round feel and more accurate reflection to our predictions and calculations. Below you can find a simplified description of our expected assists calculation methodology.

We are using two main statistics to calculate expected assists for each player; past assists data and goals scored by the team past and expected data.

First step is to take the amount of assists from the player from last season and from the current one and to combine it together with goals scored by that player’s team. That way, we are able to get a percentage which is; player’s expected assists per goal scored by his team. So if player X has 20% expected assists per goal scored by his team, that means on average, he is expected to assist one in every five goals that his team scores.

In the second step, we retrieve an average number of goals that each team is expected to score next Gameweek using bookies odds. That way, we will have the amount of goals expected to be scored by a team and the expected number of assists for every goal scored by a team, for each player.

To visualise it better, if we have expected number of assists for player X for every goal his team scores equal to 20% and expected number of goals for his team for the next Gameweek using bookies odds equal to 2, then we can expect this player to get on average 0.4 assists in this game – this gives 0.4*3= 1.2 points for assists expected.

Of course, the whole model is more complicated, as it also takes into account amount of minutes played by a player (obviously you can’t get an assist if you don’t play) as well as goals scored by a player (no points for assisting your own goal).

We also have some ideas for improving this model even further but it requires some more resources which we don’t have access to yet. However, we believe that our current model gives viable results, whilst keeping it fairly simple at the same time.

Data from bookies

In the ‘Data from bookies’ sheet, you can find all data that was in our Excel file in previous weeks, like clean sheets probabilities, goals probabilities and predicted total points calculated from bookies odds, as well as our new feature – expected points for assists based on our new model

‘Your team’ guide

In the ‘Your team’ sheet, you can choose your team and see what your expected score for that week would be as calculated using odds from the bookies.

Step 1: put your whole team of 15 players into column B by using the drop down boxes provided.

Screen Shot 2018-08-30 at 15.08.04

Step 2: choose which players will be in your starting 11 in column D by manually entering ‘1’s into the row that corresponds with the player you want to start and put ‘0’s into the row corresponding with the players you’re going to bench.

Step 3: choose your captain in column E by manually entering a ‘1’ into the row corresponding with the player you want to captain, and ensure the rest display ‘0’s.

Finish: once you’ve done that, your total expected score should be displayed in column H (cells H2 and H3 show the same expected score as for some people H2 cell was showing an error).

Screen Shot 2018-08-30 at 14.32.10
Example – Simon’s team

Blog team

In the last sheet, there is a table with predicted score and line-up of our FPL Connect team based on the prediction tool (you can read more about the team at the end of this article).

Please remember that this is a work in progress and we will continue to make improvements and add in more data and features as we go along.

Top 5 Goalkeepers

GW10 GK

Top 10 Defenders

GW10 DEF

Top 10 Midfielders

GW10 MID

Top 10 Forwards

GW10 FWD

Full spreadsheet link with all the players and the ‘Your Team’ feature

Download yourself a copy below and see how many points your team is expected to return based on clean sheet and goal probabilities from the bookies.

Excel version >> fpl 18-19 gameweek 10
Google sheets version >> https://docs.google.com/spreadsheets/d/1a1uTAs1CGa50mUEJEIlo8HdupMg-BnOywLGhYIw2Fyc/edit#gid=1476184659

FPL Connect Team

Our team was built basing on our expected points calculations. We wanted to incorporate an element of realism with this, as we know we’re unable to create a new team every week to take advantage of the results of our tool, so we are managing the team ourselves, but the spine of it was built based on the results of our predictor tool.

We will make transfers each week as normal and will place the captain armband on whichever attacker (MID or FWD) has the highest expected score, as per the results of our tool.

Gameweek 9 score

So, even though our score was not so high this time, it was enough to give us a sizeable rise in overall rank, moving from 93k to 59k overall after nine Gameweeks.

Once again, only about half of our players delivered however, average returns of the most popular players were quite low, so it wasn’t detrimental to our performance.

Our defence provided us 19 points with Robertson and Begovic getting clean sheets, as well as bonuses. Besides that, Shaw, who was only on the bench this time, once again had one of the lowest results and will probably be sold soon.

In the front, we’ve got some nice results as our midfield got 18 points thanks to Bernardo and Salah’s goals and our front three gave us 15 points however, that was mainly due to Agüero being captain, as Ings and Jimenez were disappointing and accumulated only 3 points between them.

The captain choice was good, as Agüero got us another goal however, this time, Salah got a few more points than our captain. Kane and Hazard’s low scores, who were the other 2 players in the 3 highest captained players, aided us in beating the average score for the week.

Gameweek 10 team

This week, we don’t have too many difficult choices to make. Our team is as shown above for now and we haven’t done any transfers yet, but we have 2 free transfers available after rolling the free transfer last week. We need to wait with regard to transfers though, as Hazard may have an injury that could keep him out this weekend and we would like to have full information from coaches first to decide about our potential transfers.

Salah is expected to get the most points of our team, so he will be the one who gets our armband. He is even more convincing now having scored in GW9 as well as grabbing two goals in the Champions League game last night.

So far, Ings, Hojbjerg and AWB are on the bench due to them having the lowest points potential however, we are still not sure which players will stay in the first team, as we might decide that our newly transfered player is good enough to get a place in the squad. We will maybe transfer out Shaw looking at his last returns, but this week he has a really nice fixture and players like Walcott or Ings haven’t been in great form lately. We will also keep Hazard even if he is ruled out this weekend, as the potential injury doesn’t sound serious given that Sarri has been talking about getting him ready for this weekend.

Our assists model is finally done and included in the calculation, so our predictions should be even more precise. This will make players like Robertson or Doherty to be higher on the lists due to their assists potential.

Once we know more (from press conferences and other reliable sources) we can decide our moves. I will provide an update on my Twitter account once we’ve decided.

According to our calculations, our team should get at least 72.94 points, but that’s not including some factors (like saves or bonus points, but also not including cards and some other factors like penalties saved/missed). This is a great predicted score and we will try to make it even better with well planned transfers.

All the data used for calculations were taken from: oddschecker.com; if some odds were not provided, then data was taken from WilliamHill.com

Line-ups were mostly based on previous GWs line ups.

If you want to get in touch with our data analyst or you have any further questions about this article or the tool, then please follow Piotr on Twitter here >>https://twitter.com/TheFPL_Analyst

keep-calm-and-thanks-for-reading-7

2 thoughts on “FPL Prediction Tool: Gameweek 10

Add yours

  1. I always wait for the prediction tool every week. I hope you can continue it. Thanks for the blog, very useful.. I know a lot of people out there also use the tool. Really appreciate of what you do in this blog.

    Like

    1. Thanks, that’s great to hear! Article will not be out this week but the tool is ready – you can find Google sheets version on my Twitter. If you know other people that use the tool – let them know about that too!

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at WordPress.com.

Up ↑

%d bloggers like this: