PS: Apologies for the low screen share quality. Only figured out how to do it properly after the recording. 😭 We didn't realize during the recording and it was too late to fix it. I will make sure the quality is much better next time. -Sertalp
Thank you all for your time on doing this podcast, it was truly informative and an eye opener for me. Aside from the immense knowledge in analytics, it’s also the willingness to share and teach that inspires me. Cheers guys!
Lads, as a novice statistician fumbling through an economics degree, a conversation around 50 minutes in peaked my interest around when waiting for more information to use chips turns out sub optimal. Are we talking here about the difference between Type I and Type II errors (the former an error of action which presumably is much more visible and therefore the focus). I don’t think I’ve seen anyone historically review for T2 errors, if that’s what they are. As someone who has always been a ‘wait for more info’ guy, I’d love to know if I’ve been playing sub optimally all this time! Love the pod btw!
Great stuff guys. Just one question. Coming from betting syndicate background I use a sheet which for last 6 GW's incorporates how much supremacy and goals to take off or add to teams with nothing/something to play for. In last few GW's these figures can be pretty drastic. As a result I try to keep a lot of my chips for later down the line to fully expose nothing to play for teams. Is this something Jonny and Alaister factor in? Thanks
I think this is an interesting point and runs counter to concerns that some FPL players have about playing chips late in the season due to "on the beach" concerns about rotation. In addition to what you've highlighted about supremacy swings, there is a well-documented increase in average goals in GW38 (which should extend to higher expected FPL points also) and this may extend to other teams with nothing/something to play for in the preceding weeks also. Whether this factor is big enough to warrant delaying chip use, even when optimization using strong projection data says otherwise is less clear though. Based on experience from prior seasons, the best models available already tend to model a higher number of expected goals in the final week of the season, and they also incorporate data from bookmakers and spread betting sites, so it should be factored in at some point.
@@alastairmckeever5105 Yes that's another thing to factor. In our seasonality graph (weekly goal data based on 10-12 years) the last few GW's goals go up significantly with the GW 38 being by far the highest in the entire season. Also what is worth targeting, that shouldn't put FPL managers off are home side players when both teams have nothing to play for. Goal quotes in those games go up significantly as long as it's not a derby match and supremacy favours the home team by at least 10 ticks on average. Great listening to you guys.
Thank you so much Sertalp and the guys! Layman question - Can I just download the solver and use it, or do I need to learn Python or something to use it?
No, you don't need to know Python, but setup might be a bit difficult. There is a Google Colab version of the solver, check this link, there are instructions for running it without installing anything: colab.research.google.com/gist/sertalpbilal/4c8f22e1eb640b6912c5d485405ef1ce/fplsolver.ipynb
@@FPLOptimized thank you so much!, will give it a spin I do want to learn python and what better way than to learn with football! Any basics course you recommend?
@@aungeer87 Any beginner course will cover the same basics, so any of them would be good. I suggested a few people "100 Days of Code" on Udemy (by Angela Yu) and heard good experiences. It is a more practice-based course. Check it out.
PS: Apologies for the low screen share quality. Only figured out how to do it properly after the recording. 😭 We didn't realize during the recording and it was too late to fix it. I will make sure the quality is much better next time. -Sertalp
we're blessed to have people like you in our community. thanks guys, brilliant pod
Thank you all for your time on doing this podcast, it was truly informative and an eye opener for me. Aside from the immense knowledge in analytics, it’s also the willingness to share and teach that inspires me. Cheers guys!
That bookcase is gorgeous
Great Pod. Superb information and viewpoints. Love the 'CUPSETS' line. Thats going to stick forever.
First analytic pod I listened to and already got a feeling if I am going with analytics it is going to help my daily life and work too 🤔
Ooo I'm excited for this one. Don't really know how to optimise chips!
This was a great listen.Great guests too 😍
Lads, as a novice statistician fumbling through an economics degree, a conversation around 50 minutes in peaked my interest around when waiting for more information to use chips turns out sub optimal. Are we talking here about the difference between Type I and Type II errors (the former an error of action which presumably is much more visible and therefore the focus). I don’t think I’ve seen anyone historically review for T2 errors, if that’s what they are. As someone who has always been a ‘wait for more info’ guy, I’d love to know if I’ve been playing sub optimally all this time! Love the pod btw!
Officially my favorite FPL pod. Well presented & informative! I'm interested in playing analytics FC way.
Great stuff guys. Just one question. Coming from betting syndicate background I use a sheet which for last 6 GW's incorporates how much supremacy and goals to take off or add to teams with nothing/something to play for. In last few GW's these figures can be pretty drastic. As a result I try to keep a lot of my chips for later down the line to fully expose nothing to play for teams. Is this something Jonny and Alaister factor in? Thanks
I think this is an interesting point and runs counter to concerns that some FPL players have about playing chips late in the season due to "on the beach" concerns about rotation. In addition to what you've highlighted about supremacy swings, there is a well-documented increase in average goals in GW38 (which should extend to higher expected FPL points also) and this may extend to other teams with nothing/something to play for in the preceding weeks also. Whether this factor is big enough to warrant delaying chip use, even when optimization using strong projection data says otherwise is less clear though. Based on experience from prior seasons, the best models available already tend to model a higher number of expected goals in the final week of the season, and they also incorporate data from bookmakers and spread betting sites, so it should be factored in at some point.
@@alastairmckeever5105 Yes that's another thing to factor. In our seasonality graph (weekly goal data based on 10-12 years) the last few GW's goals go up significantly with the GW 38 being by far the highest in the entire season. Also what is worth targeting, that shouldn't put FPL managers off are home side players when both teams have nothing to play for. Goal quotes in those games go up significantly as long as it's not a derby match and supremacy favours the home team by at least 10 ticks on average. Great listening to you guys.
This is probably not the nerdiest video on youtube, but it's up there
Best compliment I have received today 😅
Thank you so much Sertalp and the guys! Layman question - Can I just download the solver and use it, or do I need to learn Python or something to use it?
No, you don't need to know Python, but setup might be a bit difficult. There is a Google Colab version of the solver, check this link, there are instructions for running it without installing anything:
colab.research.google.com/gist/sertalpbilal/4c8f22e1eb640b6912c5d485405ef1ce/fplsolver.ipynb
@@FPLOptimized thank you so much!, will give it a spin I do want to learn python and what better way than to learn with football! Any basics course you recommend?
@@aungeer87 Any beginner course will cover the same basics, so any of them would be good. I suggested a few people "100 Days of Code" on Udemy (by Angela Yu) and heard good experiences. It is a more practice-based course. Check it out.
When I watch this video on 1080p your faces are very sharp however the slides you screen share are still blurred. Great content though thank you
Apologies about it, I realized the issue after we completed the recording.