Re: Synchronated Updates 08-09

Is in planning stages.

I would like some feedback on the following, which are changes under consideration:

- use of football manager skills, scaled to make the correct value for player, rather than samey skills for players in same position. will make player skills more realistic and varied. possibly won t be used for the smaller leagues as would take too much time.
- prices of players to be changed to transfermarkt value/4 (or maybe even /5) rather than the current transfermarkt value/3 , meaning great players are harder to come by and the only 5-star players in the world would currently be C.Ronaldo, Messi, Kaka (and possibly, but probably not, Gerrard), the others would be 4 1/2 stars and a lot of 4 1/2 star players would become 4-star, etc.
- possibility that player value can be a little different from his actual calculated skill-based value; making it possible that you can find bargains and overpriced flops.
- always allocating the player his best position even if it doesn t fit perfectly into the team s current formation; allows for more realism again, and better possibilities for team changes.

Thanks for your help
Sync

First of all: Synchronated, you re the best. I can t help telling this every time.

- use of football manager skills, scaled to make the correct value for player, rather than samey skills for players in same position. will make player skills more realistic and varied. possibly won t be used for the smaller leagues as would take too much time.

Well, I agree on this: I too proposed using FM data:
forum.sensiblesoccer.info/index.php?topic=6082
I hope that trying my file could be useful (I also have another dozen of teams which represent the best players of each confederations - UEFA, CONMEBOL- etc. -, with some limitations - no player from national teams in WC06 quarter.finals -).

- prices of players to be changed to transfermarkt value/4 rather than the current transfermarkt value/3 , meaning great players are harder to come by and the only 5-star players in the world would currently be C.Ronaldo, Messi, Kaka (and possibly, but probably not, Gerrard), the others would be 4 1/2 stars and a lot of 4 1/2 star players would become 4-star, etc.

Good for me. Still, I would prefer FM data, for realistic skills.

- possibility that player value can be a little different from his actual calculated skill-based value; making it possible that you can find bargains and overpriced flops.

This would be great, and in the spirit of original SWOS. Producing two versions of the update, one with bargains and flops and one with skill-based value, would be too much work, wouldn t it?

- always allocating the player his best position even if it doesn t fit perfectly into the team s current formation; allows for more realism again, and better possibilities for team changes.

This is a tough point, because SWOS management of positions is not good, it s too rigid. I would give prioriity to player s best position more than to team formation, but I understand the reasons for the opposite position.

Thanks, again.

Good for me. Still, I would prefer FM data, for realistic skills.

Transfermarkt has far more realistic/reality-based transfer values than FM IMO, they are not perfect but it s better if the values would come from there.

Good for me. Still, I would prefer FM data, for realistic skills.

Transfermarkt has far more realistic/reality-based transfer values than FM IMO, they are not perfect but it s better if the values would come from there.

Sorry, I was obscure: I prefer going for abilities (speed, passing, etc.) from FM. For prices, I have no preference: you can automatically compute them, or take them from FM, or from Transfermarkt; if you say that Transfermarkt is better than FM, I trust in you .

But I want to propose another option: you can convert a player s current abilities (0-200) in SWOS price. Just divide by 4, and you have a number between 0 and 50. Now, this is the same scale SWOS uses for prices - there are 51 different prices, see keeper s table here: yodasoccer.sourceforge.net/fformat.htm

Good for me. Still, I would prefer FM data, for realistic skills.

Transfermarkt has far more realistic/reality-based transfer values than FM IMO, they are not perfect but it s better if the values would come from there.

Sorry, I was obscure: I prefer going for abilities (speed, passing, etc.) from FM. For prices, I have no preference

Or were you talking about overall values, not just prices? I.e.: do you prefer the first (abilities from FM) or the second (price from Transfermarkt, then the editor automatically computes the abilities) of the options you wrote down in the original post?

Ideally you would have it set up so that the FM skills will give you a price close to the transfermarkt price; maybe having to adjust one or two skills but not much.

Kaka under your system would read something like 5-5-3-3-6-6-6, which only gives him a SWOS price of £1.9m... so you have to adjust the scale until he is £12m or £15m, because that s the value of the best players in the world in SWOS, so it means 7-7-6-6-7-7-7 in the end anyway.

Also you have a problem just using the seven Passing/Long Shots/Heading/Tackling/Dribbling/Pace/Finishing stats from FM: just using these stats, Ibrahimovic, Villa and Gerrard are much better than C.Ronaldo, Messi and Kaka.

I think maybe I don t trust FM stats enough to work directly from them... I m just looking at C.Ronaldo and his stats for these 7 things are way way low (passing=11, long shots=11 etc) and he would come out as an average player who is just fast, and worth about £1.8m.

It seems to me that using these stats directly can result in very unrealistic player values and abilities. If you search for average players between 140-150 on FM, Belhadj, Marco Cassetti, Chamakh, Recoba, Esposito, Faubert, Nicolae Dica, Numancia s Julio Alvarez, Rivaldo, Di Michele, Serginho, Arshavin, Kennedy Bakircioglu, Emana, Fernando, Ariel Ortega, and many many more players would end up being better than CR, Messi and Kaka (based on these 7 skills). And that s just a small sample from that skill range... so you would need to take the FM skills and scale them up/down to match the player s overall ability anyway.

Currently players are awarded skills which are appropriate to them making sure overall that they match the right transfer value, and for unknown players in lower leagues they are automatically generated by SWOES to match their transfer value, and I think this might continue to be the best way. Plugging 7 skills from FM straight into SWOES does produce the most bizarre transfer values and overall abilities.

To put it more simply, with examples from the current FM data

Great players (180+)
Kaka is 14-15-10-9-17-17-16 (total 98)
C.Ronaldo is 11-11-15-2-19-18-15 (total 91)
Messi is 16-13-7-10-19-17-16 (total 98)

Averageish players (140-150) (just a few examples)
Cassetti is 16-18-15-13-16-16-14 (total108)
Emana is 17-17-12-12-15-16-13 (total 102)
Faubert is 15-18-13-13-13-18-9 (total 99)
Serginho is 18-15-9-12-18-17-13 (total 102)
Rivaldo is 18-18-15-8-18-12-17 (total 106)

Rubbish players (100-110)
Steven Hammell (Motherwell) totals 91, same as C.Ronaldo
Steven Cohen (Ferrol) totals 91, same as C.Ronaldo
Douglas Rinaldi (Watford) totals 95, better than C.Ronaldo
Pelegrin (Salamanca) totals 97, better
Adenon (Le Mans) totals 90

Terrible players!
Rene van de Brink of Elinkwijk totals 97 on the 7 skills. he s 41 and has overall ability of 82 on FM; would be better than C.Ronaldo in SWOS based on those 7 skills alone.
Mark Snijders (age 35, 91 overall, unattached) totals 90

So, just using those 7 FM skills you couldn t make players have the right overall skill on SWOS, unless you applied other factors (scale it to a correct overall value) anyway.

First of all:
1. Synchronated acts like great men do: he listen to any opinion, thinks deeply and judge without prejudices. This is the reason (together with a lot of hard work!) that explains why its updates are so good.
2. Thanks for the feedback!

Now, to the discussion: you definitively have a point. I can t deny the strength of your examples (I can only note that a player with, say, three 7 can be best in his position than a player with seven 4, even the former he had a worst overall value; but this doesn t touch your central point).

Still I keep two points of my argument:
a. With your data Kaka would be near perfect in tackling, so he would be a great center-back or at least defensive midfielder: not realistic. But this applies to original SWOS data as well, so: 1. this is absolutely not a criticism to you; 2. I should argue against the original SWOS data, and this is not easy...
b. It seems that you agree about the fact that differentiating players values is good. Indeed, you already do it (I didn t know it; BTW, I have a question: how do you determine the starting XI and the tactic for each team? You do it very well!)

So, to keep the idea of using FM data, I should go for some tinkering, as you say. Multiplying overall value for each ability (e.g. in passing Kaka 14*190=2,660, while Cassetti 16*140=2,240 - I m not using exact numbers, it s just to have an example), but this still will be different from SWOS: an attacking midfielder could not have the max value in every ability (following this train of thought, the only player with 5 stars could be a total player like Beckenbauer). But see point a above on this.

Another way could be distributing points proportionally: e.g. let s assume that Kaka is 12M because of his overall value, then you distribute the points corresponding to 12M, e.g. 47 points, proportionally to each ability; so, you ll have something like control 9, tackling 3, etc.; you assign 7 to control, and distribute the 2 extra points proportionally again. But I don t know how to automatize this, and I don t know if the result is worth the work.

Currently players are awarded skills which are appropriate to them making sure overall that they match the right transfer value, and for unknown players in lower leagues they are automatically generated by SWOES to match their transfer value, and I think this might continue to be the best way.

If it s not clear enough from my previous post, I want to explicitly write that I agree: if we don t want to change SWOS original data s philosophy (and you should have a strong reason to do it), the best thing is to start from the transfer value, and adjust skills if possible. I.e. what you do!

(Still, I think that trying different ways, even if at the end they must be abandoned, is good: progress by trials and errors, you know.)

Just another question: Currently players are awarded skills which are appropriate to them . Can I know exactly how? You use your subjective knowledge, stats or something other? Here FM could be useful - and this is your original idea, if I m not wrong -, but, as you said, they are anyway a bit strange: see C. Ronaldo.

Another way could be distributing points proportionally: e.g. let s assume that Kaka is 12M because of his overall value, then you distribute the points corresponding to 12M, e.g. 47 points, proportionally to each ability; so, you ll have something like control 9, tackling 3, etc.; you assign 7 to control, and distribute the 2 extra points proportionally again. But I don t know how to automatize this

Well, I ve tried. This spreadsheet:
spreadsheets.google.com/pub?key=pTkVWUJsaIYpcmdjHOTbLZg
does just one iteration, that is good for some players (Kaka, Pablo Alvarez), but not others (Messi, Ronaldo, Bordon), who still have some skills 7; anyway, to have at the end the desired result (no skill 7), it should be sufficient to do other iterations. But I ve not time right now... (edit I ve changed the overall values, and, as a side-effect, one iteration is enough)

If someone (Synchronated?) is interested, I can let him edit the spreadsheet on line, or send him the file.

(I ve taken the SWOS overall value from FM value, but of course you can use Transfermarkt price or any other value.)