Re: Synchronated Updates

Was happy to fix scarini s problem over email - if you encounter the same problem just let me know.

Right, the spreadsheet - it s coming along nicely. If anyone has any other suggestion for carrying persistent attributes with only simple Excel-can-handle-65,000-of-it-at-once calculation, please let me know/discuss ASAP as this should hopefully be on the go within a week, and ready for merging with some new real-life data in terms of January transfers/February squad lists/players who might be missing

(if we can find a good way to quickly merge with real-life transfer/current squad data, the managers for each league need only be concerned with adjusting/updating players attributes and reputation, meaning those elements can be very much fine-tuned all round!)

If anyone has any other suggestion for carrying persistent attributes with only simple Excel-can-handle-65,000-of-it-at-once calculation, please let me know/discuss ASAP as this should hopefully be on the go within a week, and ready for merging with some new real-life data in terms of January transfers/February squad lists/players who might be missing

I am not sure I have understood the problem correctly, but maybe I have some ideas, or at least (as an Excel-fan) gladly try to solve it.

Well at the moment each team s data looks something like this:



Each player s attributes are numbered from 1 to 7, with the 7 being their strongest attribute and 1 being the weakest.
The sheet combines this with the reputation coefficient (just to the left of the attributes) to generate SWOS attributes which will give a value that corresponds to the coefficient, i.e. everyone with reputation of 550 comes out as £2m in SWOS, with attributes distributed according to the player s strong and weak points.

e.g. Adam Johnson at £2.75m comes out as 7623775 just by increasing the 1-7 attributes, with a maxout of 7, until you get the right number of skill points. Dedryck Boyata at £750k comes out as 3265341. At any point we can simply increase or decrease someone s reputation, and therefore increase or decrease his overall ability, but the basic description of how his skills are distributed, and so what kind of player he is, can still remain the same.

So that is how it currently works (but if possible in the ouput, 7s and 6s will be replaced with 15s and 14s in order to make those attributes searchable in SWOS), and will work fine unless someone pitches an improvement for it or another system

Currently not all points totals are possible (using this simple calculation for each attribute) where the attributes are 1-7. Sometimes one or two are skipped because two attributes increase by a point at the same time. This is the main drawback currently.
Each attribute is currently (reputation/135)*(7/20)*(attribute) since this is the formula that the version of SWOES I am using to import uses. But we don t HAVE to use this, it s just the simplest way, we can bypass it if we have already calculated the values themselves.

Thing is, using a definitive points target is ideal, yes. However I can think of no simple way to arrive exactly at it with certainty (the values in the current sheet are taken from the outcome which occurs for that reputation value, not the other way around, and the reputation values have been scaled to fit that). I can arrive exactly at a points total in my fmtoswos and fifatoswos attibute calculations, but those are very complex and have many iterative steps... perhaps that can be applied to one team at once in an intermediate step but it cannot be applied to the whole sheet. Even 600 or 700 at once and Excel yells that I m torturing it.

I will have a look at making an intermediate step which just makes the points hit a specific target as this is a very good idea and should ideally be achieved before this sheet is put to use. It should also be easier/simpler than fmtoswos, since with the attributes being 1-7 and all being different there is no problem deciding between any two attributes when adding/subtracting missing skill points.

edit: as there are only 49 possible actual reputation values, I can basically do this with lookup tables and an attribute template (or maybe more than one) for each reputation value. Should work. Same result as the current basic calculation really, except that all values will be possible and the templates will be a bit better for SWOS than the basic result currently is - and reputation now just = number of SWOS points.

Yeah if the value is higher than 7 it just reads 7.
I think this formula is about to be abandoned in favour of reputation values that are directly equal to SWOS points, and attributes adding up to the right number of points.

Nested IF functions only allow 8 conditions (and that s if you have no other functions requiring parentheses within them). At least on Excel 2003.
So we would have to do it as a lookup table anyway and then refer to it, and due to the 1-7 nature of all the player attributes the result would again be 50 templates, eg for £3m, the skills are always going to be (for example) 2-4-5-6-7-7-7 or 2-4-5-6-15-15-15, but the order in which they appear depends on the player s preferred attributes and makes most players unique or at least realistically distinctive.
So that is a good method but the result is pretty much the same.
I think now that we have set reputation = skill points, the rest is simple enough as a lookup table.