Step 4 of 5 complete✅ The Transfer Solver now handles the Assistant Manager chip! This was the biggest hurdle, so it's great to have it done even if it's a few days later than initially targeted. On the flip side, I'm expecting Sensitivity & Chip Analysis to be back... sooner than originally plannedshow more

Other notes: - Solve speed seems to be about the same - Small patch applied which will improve solver performance - If you come across a bug it would be great if you can reach out via email. I test as much as I can, but not quite like the userbase putting it through its paces

Benelux Extension 💙 We are releasing our fourth stage of Soccer Bot™ Universal+. We now offer 15 leagues (7 new) from the Benelux countries (🇧🇪🇳🇱🇱🇺). Try it out today. Try Soccer Bot™:

Not working

Should be back now

Good stuff. No increase in FT value?

It is definitely warranted to think about this, the AM chip certainly will have "some" upward impact on FT value. Extended rambling thoughts below, The scale of this upward impact requires some head scratching. The average value of an FT over the remaining window might increase by as much as ~0.35 EV (~0.15-0.20 over season), however I suspect we need to think about this differently for two reasons. 1. In my mind, one of the powers of FT value is as an abstraction that allows for sparse plans with "key" moves highlighted- the sparseness creates a more robust plan. With that in mind to retain the same level of sparsity, would actually be to take the Nth percentile of FTs used over the season as the reference point 2. The other power of FT value, is to help the literal trigger of a move in the live GW. Again it seems to me that we want that trigger set based on the Nth percentile (this becomes particularly clear if we imagine a world where an exaggerated case where a single magic 1,000 EV transfer will exist, our trigger point should not be based on averages) So for this reason, I am more likely to view AM as discarding of 1 or 2 of your weakest remaining FTs and replacing them with 1 or 2 very strong moves. Which could also be imagined as going from moving the benchmark from the typical 10th best remaining FT to the typical 8.5th best remaining FT. (Can say 20th FT => 18.5th from the full season perspective) My uneducated expectation here is this would likely be meaningfully a lower in scale than the impact to the mean (maybe +0.10-0.15, until AM finished). I would also expect the bump in value to be more stable over the season, relative to mean if we need to consider this for future seasons. Through simulation it should be possible to create an actual framework to examine this

What's the viability of coming up with something that works out best chip strategies overall?

Thanks for the updates, I seem to have lost the Chip Timing function? is that expected?

Yep, as it is dependent on the same solver it's temporarily down until made fully compatible with the new changes Will have it back up and running as soon as I can, chip analysis feels as it should be prioritised ahead of sensitivity analysis at the present moment

@FPLresearch Missed the Empty Slot > Arne Slot wordplay 😔

@FPLresearch 😂😂