Historic Rosters (2008-09)

-

Last edited by dwayne2005 on Sat May 29, 2010 4:50 am, edited 1 time in total.

-

Last edited by dwayne2005 on Sat May 29, 2010 4:50 am, edited 1 time in total.

-

Last edited by dwayne2005 on Sat May 29, 2010 4:49 am, edited 1 time in total.

-

Last edited by dwayne2005 on Sat May 29, 2010 4:50 am, edited 1 time in total.

Can anyone explain to me what does the rating "Envelope" do.

-

Last edited by dwayne2005 on Sat May 29, 2010 4:47 am, edited 1 time in total.

This is a rearchived version, now using 7Zip instead of IzArc, but still in the .zip format.

http://rapidshare.com/files/290302281/NBALive2003Rules.zip

I followed an online tutorial on how to make these 'trainers' using TSearch, but TSearch was a load of rubbish and done nothing so I scrapped one part of the tutorial that may have, in fact, been vital. For the last 2 hours, I've been trying to decipher someones tutorial for another value search program. Worst tutorial ever. And I thought mine were bad.

I need to know if the values in this little utility work for anyone else, whether any fields fail or not. I may require 'pointers', which I'm still yet to fully understand.

Hi, I wanted to know if there is a program like this for the later NBA live games, in particular 04, 06 and 08.

I never had any feedback, so I didn't continue with it. I'm sure it didn't work. I had planned to do it for other games (I have Live 2004 and 06, but not 08). Even if it did that file needs to be drastically changed. Having fields constantly refresh chews CPU like mad.

-

Last edited by dwayne2005 on Sat May 29, 2010 4:48 am, edited 1 time in total.

-

Last edited by dwayne2005 on Sat May 29, 2010 4:48 am, edited 1 time in total.

-

Last edited by dwayne2005 on Sat May 29, 2010 4:46 am, edited 2 times in total.

about Allan Rockets and Magic Woolridge. Dallas Comegys? He will be Mavericks Comegys

-

Last edited by dwayne2005 on Sat May 29, 2010 4:46 am, edited 1 time in total.

Great correction.

---

This contains my .bkr for human/computer ranges. You can range values to be different for CPU and human player. This approximates the 'reality' settings in later Live games, the way it has computer and human values. The only condition is that there is no trading into your selected human team or traded out, since this applies directly to the player ratings and computer/human players will be calculated differently. In the default provided human/cpu range, FG shooting is set to 15-35 for human and 30-50 for CPU. FT shooting is set to 0-99 (Human) and 20-99 (CPU). Blocks are set to 0-5 (Human) and 0-99 (CPU). Jumping is set to 0-75 (Human) and 25-99 (CPU). The default is preset for the Lakers (any season), and there is a provided 'mostly complete' roster in the Examples directory with these settings.

In one of my most recent games, I played v. the Sacramento Kings. Shaq had 25 pts / .688 from the field / 12 reb / 7 blk / 3 stl. Kobe had 24 pts / .500 from the field. Shaw had 9 pts, George had 8, Fisher had 6, Walker had 5 and 3 blk, Horry had 4 and Fox had 4 and 7 ast. For the computer controlled Kings, Stojakovic had 31 / .667 from the field / 10 reb / 5 stl. Webber had 22 / 8 reb / 3 blk. Christie had 13, Bibby had 8 from .286 shooting.

Sacramento won 105-91. i shot .486 to .471 from the field, .364 to .375 from 3, .685 to .800 from FT. Rebounds it was 35-38, blocks 13-8, steals 5-15, assists 30-26, turnovers 19-7.

Overall, I am a pretty pathetic 2-7, including 3 losses to the Spurs. I am shooting around .500 from the field, with some games up to .600. In one game, Shaq went 21/23 which would probably be his career record. The way I figure, the league high team shooting percentage of .500 is counterbalanced by my desire to get the easy bucket. This desire results in an ugly amount of turnovers, which results in more shot attempts from the CPU. A lot of turnovers effectively translates to a higher opposition FG%. So being league best is counterbalanced by being league worst in turnovers. I am trying to find where the balance lies between the two in my settings. Even though I do poorly in the steal department, in one game Derreck Fisher had 11 steals for me so I'm also trying to get some consistency. I am also blocking twice as many shots as the CPU, even though their block rating is about 20 times higher and jump rating is 25% higher. in my last game, however, Yao Ming blocked 9 shots for the computer. I am not opposed to it sometimes being disturbingly high for CPU since it seems there's very little I can do to fix the block amounts for human players, unless I diminish jump further.

I have hopefully fixed up the recalculation stuff (I say hopefully because i lost my original changes). There were calculations that went above and below the 0-99 thresholds. I've now reformatted all values to have IF commands for Open Office.

As usual, the ranges are set to 'simulate' on Superstar difficulty, 10 minute quarters. inside shooter is a nightmare if you wish to take frequent mid-ranged jump shots. This was necessary to get inside FG% to more realistic levels considering the ease of getting in low and close to the basket. That said, i don't take mid-ranged jumpers very often and just 2 in my Sacramento game, both which were left open and had no difficulty going in.

Last edited by dwayne2005 on Thu Jul 15, 2010 6:42 am, edited 1 time in total.

It now has an ADD SEASONS.exe in the MISC dir, where you can derive the season from Basketball-Reference. The 2009-10 season is included as an example in that archive, so the conversions not really useful until after the new season starts. It is not the same as previous conversions, so I don't want to automatically include it with the rest of the files. There are differences in main player ID's which are needed for various operations. It also has a much more simplistic position format (G,F,G-F,C...) that means a lot more players are going to get point guard/small forward/power forward than their 100% accurate position. On the other hand, it has benefits such as the proper weights of players and correct singlet numbers.
http://www.mediafire.com/file/j2n1j2mmn ... sR2m8a.exe

I am still experimenting around with the 'realism ranges'. Even after all this time, I can't find a balance.

Last edited by dwayne2005 on Thu Jul 15, 2010 8:50 am, edited 1 time in total.

Did u see my retro rosters thread? Can I use your rosters as a base?

You can use any of my files if you understand what to do with them (the Open Office spreadsheets and .bkr operations). By default, I would recommend 'no', that would be a bit of an embarrassment to me as they are neither fun nor remotely realistic. They should be ranged and set the way you want them first. There are still way too many things to iron out before I can recalculate all the player ratings in all the seasons to 'realistic' or decent ranges, and even then I wouldn't be happy with it, which is why I now have an option to set different ranges for human team, which also makes it not fully compatible with other projects unless you want to give a file for each team.

You can do whatever you like with it (no permission required, no mention of me required), these are just warnings. I think, ideally, you should look into ranging the values to what levels you want to set. If you want it to be more arcade basketball, then set them accordingly. Realism (my goal for so long now) is hard to obtain in Live 2003, if not endlessly futile. Right now, for instance, I have the shooting values ranged from 5-30 for human and 20-45 for CPU. I had a 15 range before (rather than 25) and found that even if I set the rating down to the bottom (0-15), I can't get realistic shooting percentages for earliest seasons (eg. 1946); I simply adapt my timing and begin shooting 40+ % if you play right up to the hoop (usually 60% in the 1st half, suggesting below that is psychological; forget mid-range jump shots in most cases), where it should be 20-30%. If I set the range too high, now 25 points, the lower end shoot below expectations because the timing is more different between players and harder to adapt to.

I set the jump ranges from 50-99 for CPU and 0-50 for human, by I still get a crazy amount of blocks. I'm looking into adapting that to 0-25 for human and 75-99 (haven't been able to fully test it yet because of my ADD SEASONS.exe effort). The result of doing so will be with the aim of realistic ratings, but also the loss of visual realism as the other players jump so much higher.

Ideally, if you were to use it for your project, I recommend setting the one settings for all CPU and human players ranges but I would hesitate to use my supplied range as its aim is realism and that's not really all that possible when just calculating one lot of ranges for human/cpu. There is something of an arcade range file given, ExampleRANGEADJUSTMENTS[2003] Live 2003.bkr, which is sort of an approximation of the default values in Live 2003 (it sets the maximum high ratings to the highest ratings against multiple seasons, so is not accurate in that respect, but then lifts the bottom rating to give ratings in keeping with the average rating in Live 2003), but I haven't bothered to play with it.

I forgot to mention one thing. If you use teams across seasons, ideally you should use the Pace Adjust .bkr's in combination with ranges (calculated in their rightful seasons first to determine the right FGA compared with league average). But I am yet to find the maximum values for those values. It'd give a fairer representation of modern basketball v classic basketball. (If you include just '83-98, it would be a negligible difference.)

Yes, '83-98.

But next one will be '68-83.

In any 15 year period, the pace will be mostly negligible. I should work out the pace corrector maximum values anyway, and may devote what time I have not playing Monkey Island 2 special edition (which I'm downloading atm) to that, as there is a possibility someone may copy and paste those teams to a broader range.

The two most visible changes that arise from not using a pace corrector is in the jump and speed ratings. I actually like players who played during a slower period of the game to be slower than their historical counterparts, it is also an important part about the formulation of the pace of the game for their respective seasons. But when combining teams from different eras? Problems arise, especially with jump.

The Pace Adjustments have been done. Way too much work and repetition to get the values. It now means once the Pace corrector is in, the players ratings will be stretched so that the top value = 99. Unfortunately, it is only for the per minute seasons, so doesn't include seasons prior to 1951.

Without pace adjustments, most top players are 'classic' players like Wilt, Archibald, Robertson and Russell. With pace adjustments, it's a different story and brings them closer to level, but gives the edge mostly to the later players.

Dunkability: Allen Iverson - This is the most ridiculous rating, and I never took it seriously; how do you formulate it from stats, anyway? I done it instead with athleticism in mind. The player who gets the top rating after pace correction is Allen Iverson. All the small players did, which shows my method was a failure here. But it's not that big a deal...

Stealability: Tiny Archibald - Unfortunately, this is a second failure. Not a good way to start, huh? But it is based on seasons prior to 1973-74 when steal stats were kept. I came up with a rating that I had hoped would mean Jerry West and other top offensive players were at the top of the rating, since West is considered the best thief in the game. Unfortunately, when pace is factored out, Tiny takes the title. And going by his post 1972-73 stats for steals, he was nothing remarkable. But when you try to guess such attributes from absent stats, there is probably a few correct players as well.

Blockability: Mark Eaton - This seems perfectly fair.

Orebability: Jayson Williams. Possibly wouldn't be the case if I knew the offensive/defensive split for players like Chamberlain and Russell. Moses Malone was also a strong showing.

Drebability: Bill Russell - I was expecting Wilt to be #1, but evidently not when pace is factored out. I was kind of hoping Dennis Rodman would challenge the top spot, but there is no doubt he is strengthened after pace adjustments.

Balability: John Stockton - Based on assists per game, during an average paced period of basketball, no surprise it would be Stockton who gets the crown.

Offability: Allen Iverson

Defability: David Robinson - No surprise.

Speed: Michael Adams - Speed is a ridiculous but an essential rating. While I don't believe he is the quickest player by a long shot, going by all around play I think it is probably fair for the overall 'sim'. Magic was a quicker player without a doubt, and he gets dethroned after pace correction, but then I believe overall he might not have been as direct as frequently as Adams in his peak season. It'd be equally ludicrous to get Magic scoring 40 night in, night out.

Quickness: John Stockton - I am a little surprised. It was a long time ago I done this rating, but I believe I calculated it to offensive caliber based mainly on points, thus giving the lead to scorers not playmakers.

Jump: Dennis Rodman - At once correct and incorrect. I don't believe Rodman was that great a jumper, but it doesn't mean this is wrong. The game doesn't have an attribute for arm span, and jump can make up for it, so I think overall its fair.

Dribble: Tyrone Bogues - No surprise here, since there is no essential change. Bogues is the #1 assist:turnover player in the games history and assist:turnover was used in the computation; since turnovers weren't kept until 1977, the pace difference in seasons was negligible and he still owns the top rank.

Dstrength: Dennis Rodman - The worm again. Not the strongest player, but once again difficult to derive from player stats, and he was a physical player.

Primacy: Wilt Chamberlain - But after pace correction, Iverson starts to come pretty close to the main go to man, but ball hogging has its costs since he was nowhere near as reliable with shooting the ball as Wilt.

Shooting, fatigue and hardyness ratings are not part of the pace adjustments.

(http://www.mediafire.com/file/x23b8hl5t ... sR2m8b.exe ... just .bkr alterations, btw; and somehow I inserted the 103.66 value in the wrong line in the block category)

Last edited by dwayne2005 on Mon Jul 19, 2010 2:31 pm, edited 1 time in total.

AI? Best dunker?
Adjustments will be necessary! good luck!


PS: Did you get my PM?

Adjustments may be impossible. I had originally set it to a PPG rate, so offensive players were best dunkers. But it's equally absurd to suggest Jerry West and Pete Maravich were great dunkers, and Iverson wouldn't be far behind. I couldn't figure out anything logically sound, so I ratioed it to height like my jump rating under the notion that players like Manute Bol, Shawn Bradley, George Muresahn, etc are the worst dunkers, and the result is that Iverson and other smaller players dominate (Spud Webb is not too far behind). Prior to 1979, the leader is 5'9" Calvin Murphy. The fact is, there is nothing that can conveniently identify dunkers. I'm thinking that it might actually fairly represent athleticism to some degree, but it's not a rating I think is necessary at all. I don't even know how the rating factors into play. And besides, these smaller players are not going to get the dunks anyway, so to some extent it might be leveled out a bit.

I have a tooth abscess from a chipped tooth that is becoming serious affecting half my face that I have to deal with in the next few days, and am not getting much sleep to tackle things; besides I'm perplexed enough as is, to take up on those few players you need help with at the moment. I won't rule it out, but you may be better off finding someone else at this moment. There are rating calculators out there on some of the other Live games, I'm not sure where you'll find them.

EDIT: Now that I think about it FG%:height might identify dunking better (it'd elevate Michael Jordan, but it'd ever more greatly elevate John Stockton, so that result would be a bit amusing as well unless I made it a 3 way ratio incorporating points or points:minute), but I'm not intent on revisiting the Pace values for as long as I live. The repetition drives me insane. So I'll leave it as it is for now until I adequately give Pace a trial.

The calculator is the Ratings Machine 2.1 and 2.2 by Cerventes with talk of a 2.3 but there are no working links that I could find. The 2.2 files name was RatingMachine22.rar but I can't find anything for it on Google. I'll put your players through my default Live 2003 'action' ratings as a stab in the dark for you. Do you set the years best to 99 ratings? I'll assume you do. I haven't the.bkr formatted data to factor into account team success for 2009-10 stats, either, it'd be pure stats unfortunately.

I've got to do up a new type of .bkr file to deal with making this compatible elsewhere. My ratings are set to 0 (historical lows) - 99 (historical highs), I need to a method to automatically recalculate these ratings to the max and min of each season before I can give any decent ratings. I have the idea down but not keen to work on it right away as it requires more repetition that I'm trying to take a break from.

This is just a throw out of basic figures, there are problems.

OVERALLRTG,N,2,0 PLAYRVALUE,N,5,0 FGPBASE,N,2,0 THREEPTBAS,N,2,0 FTPBASE,N,2,0 DNKABILITY,N,2,0 STLABILITY,N,2,0 BLKABILITY,N,2,0 OREABILITY,N,2,0 DREABILITY,N,2,0 BALABILITY,N,2,0 OFFABILITY,N,2,0 DEFABILITY,N,2,0 SPEED,N,2,0 QUICK,N,2,0 JUMP,N,2,0 DRIBBLE,N,2,0 DSTRENGTH,N,2,0 DHARDY,N,2,0 DSHOOTRANG,N,2,0 FATIGUE,N,2,0 INSIDESC,N,2,0 PRIMACY,N,2,0
Bender Jonathan 35 1691 39 67 88 55 42 38 39 54 33 58 51 62 63 41 52 64 30 24 34 81 51
Brezec Primoz (76ers) 22 1036 30 49 45 44 42 23 43 51 30 56 46 45 57 46 52 90 8 18 20 94 46
Brezec Primoz (Bucks) 22 1036 61 49 7 44 42 23 36 51 30 56 46 49 58 35 52 71 17 13 18 94 46
Delfino Carlos 51 2440 42 70 71 63 65 27 40 71 46 67 64 64 67 58 73 63 90 22 74 58 66
Arroyo Carlos 44 2150 60 54 81 63 50 21 35 54 49 64 51 65 74 41 93 48 87 23 56 40 54
Boykins Earl 42 2044 47 61 84 95 46 20 33 51 45 63 48 72 76 48 68 45 81 24 45 0 56

Only the bottom 3 listings had measurable minutes really; anything that is too low is useless for player ratings, those ratings are pretty random for the top 3 entries, but the 1st one is bound to be more accurate than the two beneath it. Not good, like I said. It won't have any degree of consistency. The overall and total fields are rather basic, but who knows how to calculate those properly anyway? The minimum values may be too high because my stuff isn't readily adapted to this purpose (with the Live 2003 ratings, I set highs to default game highs, but they are stretched out across all the seasons; then I moved up the lower threshold to create an average that was consistent with the default game average. Now the highs are set to season highs not alltime highs, the lower values should be reduced to bring down the average. But since I've incorporated a new .bkr procedure, I would like to address this at some point). This is based on per game ratings (only a portion, I think about 8, are per game, the rest -- physical and shooting attributes -- are either per minute or ratioed). Per minute ratings would require me going back to the high-low settings, but instead of doing that and other things I'm going to try to sleep while these 'termites' eat out my jaw...

Thank you very much and hope those "termites" stop eating you.