BuzzerBeater Forums

Ok, i will say this again. There is no way you can predict a players skillset. You can study match ratings and stats from 2 seasons if you like, but you will never get an accurate reading of a players strengths and weaknesses in his abilities. Therefore the best you can do is work out a players rough ability by his salary, assuming that he is evenly balanced (possibly factoring strengths if stats consistantly favour a certain tendancy ie fg%, fg3%, assists etc).

Since the best (although flawed, still the best) method is a lot of relying on the opponent being relatively balanced, the next biggest factor into a players strength is gameshape. Now it is known to you that gameshape and DMI and salary are relative to eachother, and I know that the game manual plays down DMI (it also says 2-3 zone is increased inside defence even though there is a large number of people who say it is useless), but I will try my best to explain how I use them all together as a scouting tool...

1) DMI=salary calculated by the code. Correct?
2) DMI is affected by gameshape, which is a major factor in a players strength in a match. Correct?
3) DMI is affected by a players skills increasing or decreasing. Correct?

If DMI was only modified by training, then why is it affected by gameshape? When you do the math, it does not always come out exact, but when you consider that gameshape has sublevels it is very close (especially since it is more consistant at Proficient which is the highest point). Now a players DMI is also affected by a pop in skill according to what it would make the players salary if it was reset that day... So with 3 different factors related to a players strength, how do you not see DMI as a guide to how strong an opponent is on gameday?

Lets take your original example. Player B drops to 440000 DMI opposed to 380000 DMI in your opponent. If those players were evenly balanced, I would therefore predict that the difference in gameshape would have made the contest between these 2 players almost a dead heat. Yes, player B would have his arse handed to him, but in the world we live in nobody is priveleged to such information when lining up against an opponent. Now in my experience I would say that the difference in gameshape and the percentages that my theory calculates strength would be pretty much on the money. Feel free to do the math yourself on a few players and see if im making it all up.

So in conclusion to your disbelief of the quote...

So in saying that, not only would I say DMI is an indicator of how good a player is, but I would say it is the single best indicator available!

I would say that there is enough evidence there to prove DMI is an indicator of how good a player is (as also stated in the quote from the game manual), but also just how effective it is in getting the average strength of your player against the strength of your opponent on matchday... If you believe you have a better available system I would like to hear it.

I would say that your assumption that players are evenly balanced is wrong. Most players in this game are not balanced. This is easy to see from a brief glance at the TL. Secondly, as I said before, salary is not really an indicator of how good a player is. As Naker Virus said, a player could have 17 JS and 5 everything else. His salary would be about 21k. You could make a guy with 8 js 8 jr 8 od 8 passing and 5 everything else and his salary would be about 8k. He would the 21k player every time.

The quote form the manual says that the DMI is "a very rough indicator." That means, "not very good." In fact, it goes on to warn you about putting too much stock in the Deliberately Meaningless Index.

To answer your question-
1) salary does not accurately assess a player's strength on the court.
2) GS can be seen on the player profile. translating it from a single digit number to a 4,5,or 6-digit number seems silly.
3) knowing that a player has trained in the previous week tells you nothing about his skills except that at the last training update one or more of them have gone up by some fraction of 1. It might be in ID for a PG which wouldn't really affect the game in which said player would be an opponent.

I think a better way of judging a player's value is to check his profile and take a look at his stats, to whatever degree of depth you choose to.
Which one of these players do you think has the 38000 DMI, and which the 54000? They both rate at 11.0 on the season, but one is better in every single statistical category.
7/20/2010 30 4-7 1-2 1-2 1 4 1 0 0 0 1 10 11.5
7/17/2010 36 4-13 2-8 1-2 1 9 3 1 1 1 1 11 11.5
7/15/2010 48 14-26 3-7 2-3 2 14 7 1 3 2 2 33 11.0
7/13/2010 12 2-4 1-2 0-0 0 5 1 0 1 0 0 5 13.0
7/10/2010 21 9-13 2-3 0-0 3 7 4 1 1 0 6 20 13.0
7/6/2010 39 5-16 1-3 2-4 0 7 1 1 1 0 3 13 9.0
7/3/2010 33 7-15 1-4 1-2 0 1 2 0 1 1 2 16 10.0
7/1/2010 13 6-11 3-6 2-2 1 2 1 0 1 0 1 17 12.0
28.6 0.456 0.364 0.500 0.8 5.5 2.0 0.5 0.8 0.3 2.2 12.5 11.0



7/20/2010 37 8-15 2-5 4-8 1 8 7 1 1 0 0 22 11.0
7/17/2010 28 4-8 2-5 3-4 1 4 0 5 3 0 1 13 11.5
7/15/2010 14 2-5 1-2 0-0 0 3 2 2 1 0 0 5 10.0
7/13/2010 28 11-22 5-10 2-3 2 3 5 1 3 1 0 29 11.5
7/10/2010 20 6-16 0-1 0-0 4 11 1 1 1 0 0 12 12.5
7/6/2010 34 6-22 1-6 1-2 1 5 2 0 2 0 2 14 9.5
7/3/2010 25 11-15 2-3 4-4 3 7 2 1 0 0 2 28 11.5
28.6 0.469 0.400 0.667 2.0 6.3 2.8 1.5 1.7 0.2 0.8 19.7 11.0

Amazing that they exactly the same average minutes. I didn't notice that until I posted here.
Anyway, if you want to use DMI as an indicator of a player's strength that is fine. If it works for you, even better. I am just saying that I don't personally put any stock at all in it when I scout an opponent, and don't believe it to be intrinsically valuable to the success of your team. I have done okay without ever looking at it as a scouting tool, so for me it has absolutely no importance.

It is obvious which player is who or else you wouldn't use it in your arguement. Fact remains however, that can't discredit my theory as we do not know the DMI of the players in every match, nor that of their opponents as it changes from week to week. And if anything your example is saying that the best rounded player is best everytime? Wouldn't that mean that if my calculations work (which you have not shown how they don't) that by preparing for a well rounded player, you are going to be better off if the player does have 1 very high skill and the rest mediocre? If so, as long as you can cover a balanced skilled player in the equation, I would say that you have covered the worst case scenario and almost guarantee a good result. And looking at those stats you could never predict the players exact skills. If both have had proficient gameshape and either one drops to strong, then with skills that close you would say gameshape would have a massive factor in the matchup.

I really don't understand what grounds you are trying to say my theory is wrong on? So far your only arguement is that a player that is well rounded will beat a 1 skill wonder, which you will find I have said several times that the best you can do is assume just that. And just because you have done "okay" without it as a scouting tool, does not mean it isn't effective!

If you see a guy with 380000 DMI at GS 9 and you say to yourself, 'Oh my player at GS 9 is 540000... I'm gonna run my offense to exploit that mismatch" and it turns out that your opponent's guy is the MUCH better player even though he has lower salary and DMI, you're going to have problems.

Like I said, if it works for you, great. You seem to be having middling success so far.

If my 540000 DMI player has a massive has 1 over the top skill, I know that he is a player that is suited for one role and one role only. Its much easier to scout your own players.

Actually if you can expect to be in a league where everyone has roughly similar players

Would that it were so... Actually in our PL I would say the players are fairly even. I haven't used DMI as a scout tool in that, but maybe this week I will give it a try.
According to Pablo Ignatio Martinez the figures are 7*salary at strong and 3*salary at respectable.
Due to injuries etc. I happen to have a number of really crappy game shapes on my team at the moment and I would guess that respectable is closer to 4*S, average is closer to 3*S, and mediocre is aboutr to 2*S.

How looking at DMI is any different from looking at GS is a mystery to me still, since A) the number is "give and take" and it is dependent on salary as well, which we know is a poor indicator of a player's on-court performance. I suppose if you had reams of data and could accurately predict GS sublevels using it, it might be more helpful (And I admit that it is helpful to some as is) but I haven't ever taken GS sublevels into account in my scouting preparations.

I can't believe how hard it is to explain such a simple theory...

I would guess that respectable is closer to 4*S, average is closer to 3*S, and mediocre is aboutr to 2*S.

That is why I said it is "around" the figures I stated. Do the math against all players on your list. You will see that they differ due to sublevels. You do have a player whose is respectable with a DMI pretty much on 4x the salary. Your other respectable is pretty much 3x. So that would factor a 10% swing in skill with the same gameshape if they did have the same skills. Same rules apply for all levels of gameshape with the higher levels being greater affected than the lower ones.

How looking at DMI is any different from looking at GS is a mystery to me still, since A) the number is "give and take" and B) it is dependent on salary as well, which we know is a poor indicator of a player's on-court performance.

Looking at DMI any differently from gameshape is the whole thing I have been trying to explain, yet you don't look past the fact of a players skill. So...
A) The number is "give and take" according to sublevels. I don't know how many times I have to repeat that. I have repeated that many times in this thread, this is the second time in this reply alone and I know that I will have to do it again.
B) YOU DO NOT HAVE ACCESS TO AN OPPOSITION PLAYERS SKILL THEREFORE CANNOT FACTOR INTO YOUR SCOUTING. This thread is about DMI and its ability to predict strength... NOT SKILL! It is about predicting a players skill according to the effect current gameshape has on your opponent assuming (which is all you can do) his skills are balanced according to salary!

I suppose if you had reams of data and could accurately predict GS sublevels using it, it might be more helpful (And I admit that it is helpful to some as is) but I haven't ever taken GS sublevels into account in my scouting preparations.

Once again, if you divide a players DMI by his salary you get a number that is relative to his gameshape. How is that not an accurate (provided the player has not been trained in which it would still serve the same purpose of showing the players relative strength) calculation of gameshape sublevels? This is not about some program that claims to predict the absolute foolproof system of scouting. You challenged my theory of DMI being one of the best indicators of strength available when you have not only never bothered to look into it, but you are looking for some bigger picture to prove me wrong when my theory is so simple that anyone could use it... Just because you have never taken GS sublevels into account, does not mean that it is not an effective tool!

Last edited by Pablo Ignatio Montoya at 7/23/2010 2:37:16 AM