Great notes from all / appreciate the comments.
The general response continues to be in favor of more reflective/useful commenting should a negative reaction occur between players.
Two additions:
(1) 'include comments' - Believe most on the site are in agreement that any form of a comment associated with a red thumb would be of more use.
The actions therefore are supporting either (a) current method (no force to provide a comment, or if desired can forward an IM) or (b) force a comment with red thumb.
Those responding in this dialog also seem to desire to retain the anonymity of the red thumb giver, and perhaps an impression that any comment will be for benefit.
To restate: The general posting was for two ideas - one, to examine the usefulness of the current green/red thumb system, and two, to reduce the misuse (or improve).
Much of the focus of many of the responders is in reaction to red thumbs offered often in heat of battle, reactively, and with no useful content for correction, etc.
The idea of promoting a comment action along with a red thumb is positive; however, for those not interested in changing their reactive thumb behaviors the likelihood of useful comment from this is probably the same as being forwarded an IM after a red thumb currently (in other words, very low to zero).
The forcing of a comment prior to issuing a red thumb would in most of the reactive situations (again, the primary focus of why looking at the system) lead to flame IMs just to get the thumb issued.
Currently, IMs are not anonymous and thus not used for this effort. While recognizing some folks are sensitive to this point, am not a fan of the anonymous thumb or comment regardless, so have no issue with current state IM exchanges to clarify any play issues. Given this, I would send a IM instead of a red thumb first (key) to clarify, etc. and then decide on either a red thumb or no issuance of a green. Note this does not change in any system suggested.
(2) 'percentages' - Appreciate the ongoing looks at what the math might be with the current or any alternative solution. Yet another quick try here:
The usefulness of a player's general ranking noting both green and red thumb quantities again assumes everyone playing invokes a ranking (gives a thumb).
The reality is that a sufficient number of players do not even apply a green thumb (even though as discussed many default greens occur), lending dilution to the rank idea.
As of this writing, I have played 1644 games with a reputation based solely on only those willing to have ranked my play (634 green/13 red).
While positive (thanks everyone - 634/647 is a 97.9% positive...), when viewed against all games I've played, it is 38.6% (634/1644).
The difference is that many of the games played are with players that have already ranked me, therefore this is a less than useful metric if it is not evaluated every game.
However, if one was able to identify exactly how many 'non thumb' games I played, then a better total (# games with all players who chose to rank or not) could be used.
As a second point, the worry of what such a ranking may mean as a lower number relative to the current reputation ranking should be clarified.
In a generic example, if a player plays 200 games and is ranked for, say 150 games (good or bad), the 50 players not providing a rank are likely to not provide a rank regardless of the system and can be held constant (yes, a changed system may or may not encourage more/less participation, but for now we'll let them be).
Using the current system, if in the 150 games a 'good' player received 120 green thumbs and 30 red thumbs, a 120/150 or 80% 'good/bad' rating results, as well as a 120/200 or 60% 'ranked positive for all games played'.
Note that the metrics mean different things (you did not suddenly lose 20% 'goodness' as a player, but maintained a 60% 'all games played' rating).
Using a green-only system, the 'good' player would be ranked 'good' in 120 games, rendering the same quantity number (120) and the same 120/200 (60%) all game rating.
While the second metric is lower than a comparison to good vs bad, the overall impression should be taken as more impactful given the player encouraged 120 fellow players to consider them 'good'. This helps underscore the points made earlier on 'promoting the positive' vs. 'avoiding the negatives'....
(2.5) In reading the above - yes, to make a better ('best'?) metric useful for the green/games, a thumb vote would be suggested for all players after every game.
This is likely challenging and perhaps tiring (how did every player perform every time vs the current general long term feel of a player) but does work.
The bigger point was to try and show that one's collective 'goodness' does not get diluted if the system gets a change, just how one interprets it.
Short takeaways: If the reactive red thumb comment-givers can be considered to provide useful comments via even anonymous IMs, would be useful.
