PREDICT Football Matches With POSSESSION | Excel Tutorial 

Download excel ?

@@jurajpodmanicky6382 Yes, signing up will give you downloadable access to this spreadsheet 👍

@@excel_ladz download complete program possesion distribution

@@jurajpodmanicky6382 hey, please i'm finding it difficult to download the spreadsheet, do you mind sharing?

Table as shown in the video is not in the site.. there are tables for different stats and one team may have more stats than the other in the main league table making it hard to transform the table.. kindly share the exact link

Thanks lad 💪💪💪. The inspiration for this video was actually from a subscriber in my last 'World Cup' video. So, if you have any suggestions please don't hesitate to write them in the comments lad.

Thanks lad, make sure to subscribe if you haven’t already!

@@excel_ladz of course lad

Thanks lad, I’m very happy you took value out of the video 💪.

Thanks lad, I appreciate the support. Subscribe for more!

Hi lad, your relative Attack/Defense Strength Rating Model is exactly what I’m looking to publish within the next couple of weeks 🔥 Make sure you’re subscribed to catch this lad 💪

Thanks for watching lad 💪

G’day lad, you’re totally right 🔥 This video was more for beginners to understand the process of using linear regression to estimate an output based off of an input 👍

Thanks for watching lad 👍. Unfortunately I don’t think I’ve had the privilege of meeting your nephew.

That's great news lad! Thank you for watching 👍

Hi lad, thanks for watching 💪 The data on the EPL Data worksheet is from Football Reference where I’ve gotten the stats from the EPL Squad Possession table there. In the Simulations worksheet, the data table created is just 10,000 simulations of the scorelines based on xG (Expected Goals). This is from S11:S12 on the EPL Data worksheet. Now, these simulations you can create totally by yourself using the formula outlined in the video. I hope all goes well lad 👍

No, sorry lad. Just try and have a watch through the video and copy the steps shown. If you run into a problem, just chuck a comment and I’ll comment as fast as I can.

Yeah for sure lad 👍 Here’s a great website for pulling in these stats for the NHL: www.hockey-reference.com/leagues/NHL_2024_skaters-advanced.html

Hi lad, yes unfortunately I’ve had to do that. However, newer versions of Excel can allow you to do this, while still using a Power Query to make the data update automatically 🔥 First you would have to make two separate queries; one for the normal table with the goal difference, points and etc, while the other query would have the possession data. Then in Power Query Editor, you can merge these two queries (as long as one column is constant or common between the two tables, which is the teams list). This will become a new query in which the data can update by itself 👍

Hi lad, thanks for watching the video 🔥 In this particular model, there is no difference between the ‘What If’ and copying down the 10k simulations. The reason why I wanted to include it in this model is because ‘What If’ analysis is a really valuable tool to have when running simulations that have multiple formulas and calculations. For example, using a linear regression model to simulate an entire league table using the Binom.INV function, then simulating that whole season itself 10,000 times (which only ‘What If’ has the power to do). I’m hoping this explanation helps lad, sorry if I’ve caused some confusion with doing the What If rather than copying down the Binom.INV function 👍

@@excel_ladz no worries it’s a great explanation. I’ve always wanted to do simulations like this but have never had the skills or knowledge so thanks.

Hi Sara, I have a video predicting the World Cup on my channel but it does not use linear regression. You could use linear regression in predicting the World Cup, but there is a lack of data available compared to the Premier League. This is because national teams do not play a lot of matches relative to English Premier League teams, meaning the dataset is smaller and perhaps less reliable. Please subscribe for more content like this 👍

Hi lad, yes there is 🔥 The average goals is basically a scale to adjust a team's xG for the match. Let's take an example of Everton vs Manchester City. Man City has an ATT Rating of 2, meaning they score goals 2x more than average in the league. Everton has a DEF Rating of 1, meaning they concede an average amount of goals in the league. Multiplying these two together, Man City will be expected to score 2*1 = 2 times more than the league average in their game against Everton. If the league average goals were 2, than Man City would be expected to score 4. If it was 0.5, then they would be expected to score 1 👍 Vice versa for Everton's xG: Everton ATT * Man City DEF * League Average Goals

G’day lad, thanks for commenting. If you haven’t already, please consider subscribing to the channel! To your question though, any significance F below 0.05 is considered fine and acceptable to be used in a predicative model. In regards to creating sequences without the most recent version of Office, just type in 1, then 2 below it, then 3 below that cell. Then highlight this range and drag down for 1,000 cells (or rows) and it will create a sequence from 1-1000. Let me know if you have any more questions lad.

Absolutely lad, as long as the dataset has around 20+ games, it should be accurate enough 👍 Thanks for watching!

Hi lad, you can email me at ladz.excel@gmail.com where the Excel Sheet Template is available for a small price 🔥. I hope you enjoyed the video 💪

G'day lad, that sounds like a great idea 💡. In the short term, I'll see if I can upload that to the channel 💪

Hi lad, I’d love to help 💪 If you’re interested in joining the Excel LADZ Discord, where I and others will help you with your sports models, have a look at the link below 👇 A copy of the model adjusted for Excel 2016 users and later is also available for instant download 🔥 Check it out here: www.patreon.com/ExcelLADZ

Hi lad, no bother at all! I love answering questions for subscribers and viewers 👍. First of all, I don't think you doing the regression on Windows had anything to do with your error. Make sure when you input the x and y ranges for the linear regression they are correct for your data. If you are copying my data, then enter the formulas in exactly how I've done it in the video. If your problem is the linear regression tab doesn't even appear, then make sure you have the feature installed within your Excel 'Data' tab by enabling it in settings. If you still have problems, comment again lad. The more specific the better my answer will be 💪 .

@@excel_ladz Thanks a lot mate I'll try this out !

Try the link in the description for the data lad.

I feel you can use R. as soon half of the season is done, define training and test data. Its better for prediction

G'day lad, obviously it's much tougher estimating the attacking and defending strength of a newly promoted/relegated club when they haven't played any games in the competition yet. One strategy would be to gather data on the success of teams that are relegated/promoted in the past few years, based off of their points before getting promoted/relegated. Then you could apply those ratings to the new clubs. It's a really interesting topic for me personally, so I plan to release a video on this before the new Premier League season starts.

Hi lad, you could do this by converting the percentages to 'intrinsic' odds. For example, a Man City Win outcome with a likelihood of 50% should be fairly priced at $2.00. If any price above that is offered, the bet should be theoretically taken to maximise expected value. However, the model is obviously not perfect. The closer your R squared value is to 1, the more powerful the model becomes.

Hi lad, the stronger the correlation, the more accurate & reliable the model will be. While a model with a weak correlation could still produce a predicted result, it would be less reliable 👍

@@excel_ladz Thank you for the respond. I've also watched your video about prediction of world cup using probability and elo rating. I've enjoyed that video a lot. Can you suggest me ways to construct a probability with possession-related system? I've calculated the correlation but it shows moderately weak. My teacher don't really encourage making a model just as you mention but instead make a probability. Your help would be really appreciated. Thank you!

Hi lad, thank you for watching 🔥If the input range contains "non-numeric data", then you must have your numbers formatted as text rather than actual numbers. You'll see that the range I've highlighted for the 'Input X Range:' at 2:53 has numbers on the right hand side of the cells. 67.9, 62.7, 61.8 etc are all on the right-hand side of the cell; this means it is formatted in numeric form. In your case, it is most likely that your numbers are on the left-hand side of the cells, and instead formatted in text form. You need to convert these numbers into numeric form, and you'll know it's successful when the numbers are on the right-hand side of the cells. One solution to fix this is to wrap the numbers in the VALUE function in Excel... Let me know how you go lad 👍

@@excel_ladz Thank you for answer, but i have one more problem. At 9:17 =xlookup(P7,teams,$k$2:$k$21) and i get message: It is some problem with the formula try etc etc" why? I did everything the same way you did from the start on the video, i followed every second you did. But for some reason it not working. :( Help please thank you

And i use windows, Excel 2016

Thanks for watching and commenting lad. As to your question, I'm not quite sure what you mean. Are you suggesting a stem and leaf plot held in cell K2?

Sure lad, what part of the video in particular are you having problems with?

La formula RAND() EN español que significa y

G'day lad, thanks for watching! Make sure you're subscribed to be come a lad 💪. As to your question, the R-Squared value in this model is only 0.71 (as explained at around 3:33 of the video) . An R-Squared value of 1 means that all predictions are identical to the observed values. Meaning, the closer you get to 1, the more accurate your model will be. Usually an R-Squared ratio of above 0.90 is considered to be good or a "reliable indicator", meaning that this model (due to its lower R-Squared rating) could potentially lead to inaccurate predictions at a more frequent rate than a model with a higher R-Squared ratio.

@@excel_ladz what factors determine R-squared accuracy?

Factors that increasingly successfully predict the outcome you are measuring by will score higher R squared values. For example, when predicting whether a team will win a match, factors such as possession, shots on target or dangerous passes will lead to higher R squared values (and thus higher confidence in the model) than a stat such as yellow cards. Sorry for the late reply lad, for some reason I don’t get notifications when people reply to an existing comment 😞. Follow me on my Twitter page: @excel_ladz 🔥

Hi lad, thanks for being a loyal follower of the channel. This data (from the 2021 English Premier League season) has been copied from the website Football Reference. The link to the stats is in the description of the video 👆👆👆 .

Hi lad, you can reach out to me on my email: ladz.excel@gmail.com Thanks for watching Excel Ladz 🔥

HI lad, make sure to subscribe if you enjoyed the video! The POSS that comes into my worksheet is the average level of possession that team had scored throughout the 2021 English Premier League season. I then use this stat to predict the outcome of matchups between these teams, assuming there is a correlation between a team's performance and their level of possession.

Have a look at the video I just posted lad 🔥

Ben yapamadım bana gönderebilir misin?

@@Fahri-55 nasıl göndereyim

@@Fahri-55 aslında yüzdelik olarak %70sin üstündeki veriler oda oranlar zatrn 1.25 gibi oranlar