Help with creating boxers

[Tiger Fan]

First off I want to thank everyone who has been providing us with dozens of new boxer files to add into our games. I want to repay the favour and beginning contributing a few myself. Problem is I have no idea how to start.

Could a few of the veterans provide me with some ideas. Do you simply take a boxer with a similiar record and use that as a starting point and tweak ratings from there? Or do you start from scratch and if so what do you begin with. I am a novice when it comes to boxing knowledge so how much tougher will that make things for me?

Any advice would be much appreciated.


Btw, I have read part two of Jim's article on rating fighters but cant find a link to part one.

thanks

[Mark42661]

Hey Tiger Fan,

When I rate new fighters, I usually copy one who I feel would have similar ratings and then tweak them. During the tweaking process I run alot of test bouts until I get the fighter to where I feel he should be. I like to test the newly created fighter against a real-life opponent already rated in the game until the winning percentage, KO percentage, etc, seem about right. Alot of it is very subjective, but you can usually get pretty realistic results with enough testing and tweaking. I hope this isn't too general an explanation and helps somewhat with your question.

By the way, I enjoy following along with your replay.

Mark

[Dave1877]

The manual is invaluable for creating boxers, so check that out. Also, look at the fighter's record. If he's won all of his fights by KO, this might mean his hitting power should be high. But be careful, you need to look at the quality of his opponents before making a judgement. Stuff like that is what you need to look for. Try boxrec.com. Check how many times the fighter has been knocked down for example, then rate him accordingly.

[IceTea]

I think everyone has their own style. I like to give a fighter split ratings against boxers/sluggers 10-8 9-7 when I can. I think it adds much enjoyment to the process of reply. I try to read as much as I can about each fighter I rate before doing so. An earlier posted stated that an inspection of quality opp is mandatory before issuing a punching power rating, I agree with this statement. I normally take the fighters best fight and their worst fight and consider them outlyers and throw them out of the evaluation process, only using what it left over.

MJ

[Tiger Fan]

Okay, read the manual and the pointers above. I have selected a fighter I will try and rate. He is not in my database but did have a large number of fights so I want to add him to my 1940's replay.

The fighter is Al Mobley (boxrec id 013096). A career record of 55-38-10 with 15 ko's.

Among fighters in the game he fought
Lost 6 rd TKO Sugar Ray Robinson, lost on points to Ike Williams and Freddie Dawson. Fought Eugene Hairston 3 times (a draw and 2 losses on pts)

Other than Sugar Ray and a couple of fights at the end of his career his losses were all by decision.

So my thinking is I will use Hairston as a basis for rating Mobley. Not a great power puncher I would guess by his knock out ratio I assume I should rate him high in defensive categories to mimic going to a decision so often.

I will update you on my progress and may have some more questions. Btw, if anyone else has rated him already let me know and I will pick someone different.

[crusadecat]

I like to look at opponents the person has fought and how the person did against them. I'd suggest doing playtesting of the fighter to see if you are happy with the the results of if the ratings need tweaking. Feel free to make plenty of mistakes, its all fun and educational.

[Tiger Fan]

Okay here is a follow up question.

I tweaked the ratings down to get starting numbers for Mobley and then simmed 100 fights against both Sugar Ray Robinson and Freddie Dawson

Here are my initial results
Robinson who won by TKO in the 6th in real life was 97-2-1 against Mobley with 75 ko's with the average number of rounds in a Sugar Ray win being 5.89

Hairston in real life was 2-0-1 against Mobley with both wins on points. In 100 test matches Hairston was 69-20-11 against Mobley. Hairston won 6 times by KO or TKO while Mobley won 8 times by stoppage.

My question is where do I go from here. I have a couple of more fighters in the game who fought Mobley I can test with but from a quick glance my results seem pretty good. Is there a next step (keeping in mind I know little about Mobley's boxing style) or do I consider the file pretty much complete?


My test Mobley is attached for anyone interested.

[Tiger Fan]

Couple more real life opponents I tested

Frank Szuzina who he fought to a draw twice late in his career
In 100 test bouts Szuzina won 48, Mobley 43 and their were 9 draws

Freddie Dawson who he lost on points too in the middle of his career
Dawson won 95 bouts, 12 by KO with Mobley win 2 and their were 3 draws

Otis Graham, who was 2-2 against Mobley with all going to decision in real life
Mobley won 43, Graham won 36 and there were 21 draws in my test. Each fighter had a pair of knockout wins in 100 bouts


Ike Williams who beat Mobley on points in real life. In 100 test sims Williams won all 100 including 50 by knockout.


The Williams result is the only one that doesn't fit properly as there were far too many ko's I would suspect. Or do I assume that is not a bad number because in real life Williams won nearly half his bouts by KO and maybe he was having an off night in his real life decision over Mobley?

[LeeSkye]

Taking a quick look at Mobley's record shows he was never KO'd until the end of his career and Sugar Ray Robinson only stopped him on a TKO before that. That means the guy has a pretty good chin. If Ike Williams keeps KOing this guy then I would give him a better chin to prevent that or maybe give him a better ability to "Absorb punishment". Once you keep him in against Ike you may have to go back and tweak his CF or PL etc to get him back to the accurate results you have been getting with the other opponents. And don't forget the CL-clinching factor. I don't know much about this guy but he stays in fights - much like Ruiz does - and clinching can negate the others guys power and killer instinct-maybe this guy is just a lighter Ruiz!

All in all you have done a good job rating this fighter IMHO.

[SAL]

Looks like you're on the right path Tiger Fan, test,test,test,test,test.........
When I rate a fighter, I look for all the boxers already rated in the game that he has fought and run at least 100-200 test fights against everyone of them and then start tweaking the new fighters ratings until I get pretty much the same result in replay as in real life. When I feel I'm close with the ratings, I'll set up the same bouts again but run 500-1000 test fights and look for at least 75% or higher the same results for replay as in real life.
Just for fun I downloaded your Mobley file and ran a few test fights. Don't take this as the gospel because I don't know anything about Mobley, but I changed his PL/CP to 36-38, AP 4 to 2,KO 2 to 1, DEF -3 to 0, raised some of his stratagies and his fighting on the ropes to a 1. This changed his overall rating to a 5. When tested against Ike Williams, I selected the box adjust for weight differences in the scheduler and Ike destroyed him 88-10-2 but with only 6 by KO. Without the weight diff. checked, Williams won about the same amount of times but with 59 KO's. Looks like you did a good job on Mobley. Looking forward to adding your new fighters to my game.

Scot

[erickdamac]

Leesky I must Admit I love You website, I like how you have real life events in the News, such as fighters death, which occured often in those times, Keep up the good work.

[Tiger Fan]

Okay, took Sal's and LeeSkye's advice and did some more tweaking plus an hour of testing and I think I have Mobley ready.

Here are my final test results
vs Sugar Ray Robinson - real life Robinson was only man to TKO Mobley prior to late in his career. Robinson won by 6th rd tko

My test results Robinson wins 984 fights or 98.4% with 44% by KO

Ike Williams - in real life Williams won a tight 10 round decision
My test results Williams wins 87.7% of time with 4% draws. Williams wins by decision 77% of time

Franz Szuzina - in real life fought to pair of 10 round draws late in Mobley's career.
In test Mobley wins 44.5%, Szuzina wins 40.5% with 15% draws

Otis Graham - in real life Graham and Mobley were 2-2 against each with all going to scorecards
In my test Mobley wins 49.5% of fights with Graham taking 36.9%. Remainder are draws. Perhaps Graham is underrated as I have every other opponent pretty much in line except for him?

Eugene Hairston - in real life Mobley had a draw and 2 losses by decision.
In test Hairston wins 74.5% of time with Mobley taking 15.3% and remainder are draws.

Freddie Dawson - In real life Dawson won on points against Mobley
In test Dawson wins 88.7% of fights with 84.1% by decision.


Except for Graham I believe I have all the opponents in line with Mobley. Only Robinson knocked him out in testing with any regularity. Williams is down to under 10% knockouts from the close to 50% I was at with the earlier version. So I think Mobley is ready for release. Thanks to each of you who replied for your help. I think I am slowly getting the hang of this.

I will post Mobley in the non-heavyweight fighter thread. Any changes you might suggest after seeing him are greatly appreciated.

[Gunsmoke]

Well Done TF, the ratings look really solid. The rating of fighters I find to be just a matter of having a good test bed of fighters (real opponents and some iconic style type fighters) and good deal of hard testing.

Mark