Produce ranking

[PaulCristina]

Can a ranking be produced? The "model.probabilities" accepts only two teams as input but what if you want to produce a ranking using the games played so far?

[lucasmaystre]

Hi @PaulCristina you can use `Item.predict()` to predict the score of each item and rank them according that score. Something like: ``` ts = [1.0] # Point in time at which you want to make the prediction. res = dict() # Contains predicted score. for name, item in model.item.items(): means, vars = item.predict(ts) res[name] = means[0] for rank, (name, score) in enumerate(sorted(score.items(), key=lambda x: x[1], reverse=True), start=1): print(f"rank {rank}: {name} (score: {score:.2f}") ``` I haven't run this code but I hope it gives you the idea! Lucas

…

On Tue, Aug 10, 2021 at 4:00 PM PaulCristina ***@***.***> wrote: Can a ranking be produced? The "model.probabilities" accepts only two teams as input but what if you want to produce a ranking using the games played so far? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#9>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACD6ZNG23E73MYFOASU5ULT4E5J7ANCNFSM5B4LJKAQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email> .

[PaulCristina]

Thank you