Stats Only

[saturn2187]

So I think I'm on board with creating a Stats only league, with one caveat. I've always played Historical, and enjoy the players.


Is there some way to sprinkle some random real players into the draft each year? For example, maybe create a draft pool that has 30 rounds - a random batch of real historical players and the rest filled out by fictional?

I understand that they will not contain high school/college stats.

[Mr. Marlin]

If you use feeder leagues under the league set up options there is a check box for "Import Random Historical Rookies". If you don't use feeders then click the same box in your Major League's options.

[curt]

Where are the settings for stats only?

[The Wolf]

In this thread.

[marc5477]

I tried it... broke the game in about 6 seasons. The problem is the AI sucks the big hairy meat ball. When you set its evaluation to stats, it ends up with Willie Randolph making near the league max salary. You can alleviate this somewhat by giving AI teams more money than they can every use... but that makes playing somewhat meaningless. As a side note, I tried 0/50/30/20. Did not work well at all. The AI needs help.

If I had to do it again, id say turn off all rating scales so that you cant see them. Turn off historical rookies to make sure you have no advantages. And give yourself about half as much budget of the league average.

The bad news is that this is not realistic at all. In real life, we have some idea of talent. I mean... when a healthy guy can bench 350 lbs out of high school, you have a clue about power even if he hit no homers in his high school career. It just means he is undeveloped or had horrendous coaching or played behind the coaches favorite (usually) or an even better player who might have been more developed. Those ratings are usually not that bad.

[marc5477]

Originally Posted by saturn2187

So I think I'm on board with creating a Stats only league, with one caveat. I've always played Historical, and enjoy the players.


Is there some way to sprinkle some random real players into the draft each year? For example, maybe create a draft pool that has 30 rounds - a random batch of real historical players and the rest filled out by fictional?

I understand that they will not contain high school/college stats.

You can use historic rookies and there is an option that has the AI valuate the historical rookies higher to balance the things out. But after the draft... the AI has no clue. Thus you will see a lot of HOFers in FA at low prices and the AI will not pick them up. The problem is, you, a human, knows these guys are super stars, while the AI only sees stats. I actually made a request that the AI use the same custom draft filter on free agents to insure some big stars dont get undervalued. There was a season in one of my games where I had a real stupid team. Ken Griffey Jr (off a injury plagued season), Frank Thomas (after he had a so so season), Edmonds, Glavine and bunch of others (i think I had 9 or 10 super stars on that team) and none of them made much money at all. It was a laughable season and I started a new league after that.

[curt]

Originally Posted by The Wolf

In this thread.

Sure. Make me search.

[The Wolf]

Originally Posted by marc5477

I tried it... broke the game in about 6 seasons. The problem is the AI sucks the big hairy meat ball. When you set its evaluation to stats, it ends up with Willie Randolph making near the league max salary. You can alleviate this somewhat by giving AI teams more money than they can every use... but that makes playing somewhat meaningless. As a side note, I tried 0/50/30/20. Did not work well at all. The AI needs help.

If I had to do it again, id say turn off all rating scales so that you cant see them. Turn off historical rookies to make sure you have no advantages. And give yourself about half as much budget of the league average.

The bad news is that this is not realistic at all. In real life, we have some idea of talent. I mean... when a healthy guy can bench 350 lbs out of high school, you have a clue about power even if he hit no homers in his high school career. It just means he is undeveloped or had horrendous coaching or played behind the coaches favorite (usually) or an even better player who might have been more developed. Those ratings are usually not that bad.

Two things: first, you weren't doing it right (Willie Randolph?) and second, you don't really sound like you're cut out for stats only, which is of course fine,

[The Wolf]

Originally Posted by curt

Sure. Make me search.

Yeah, we deliberately hid it in the middle of the first page.

[The Wolf]

Originally Posted by marc5477

You can use historic rookies and there is an option that has the AI valuate the historical rookies higher to balance the things out. But after the draft... the AI has no clue. Thus you will see a lot of HOFers in FA at low prices and the AI will not pick them up. The problem is, you, a human, knows these guys are super stars, while the AI only sees stats. I actually made a request that the AI use the same custom draft filter on free agents to insure some big stars dont get undervalued. There was a season in one of my games where I had a real stupid team. Ken Griffey Jr (off a injury plagued season), Frank Thomas (after he had a so so season), Edmonds, Glavine and bunch of others (i think I had 9 or 10 super stars on that team) and none of them made much money at all. It was a laughable season and I started a new league after that.

And this is why you always use fictional players.

[marc5477]

Originally Posted by The Wolf

Two things: first, you weren't doing it right (Willie Randolph?) and second, you don't really sound like you're cut out for stats only, which is of course fine,

I think you misunderstood... it was too easy due to how badly the AI handles stats only. By broken, I mean my team was so overwhelming that the game was no longer worth playing. It was like the All American Basketball Team at the Olympics. It was just a funny joke.

I just loaded the save to grab the numbers. I actually played 5 seasons in that game and simmed 3 at the end without even bothering with any transactions and automating the draft and minors. In that last 4 seasons my team had between 100-130 wins and won 2 pennants. No big deal right? Well the payroll was about $14m... The league average is about $30m (just eyeballing it) and the top team (Yanks) had a payroll of $54m. Still, all this would be ok until you see my farm and that was the death knell for the AI. Their farms stunk while mine had a AAA team that could have probably won a pennant or 2 at the major league level.

I can discuss the strategy (upon request) but I really dont want to spoil the game for those enjoying it. Bottom line though, it was too easy. The AI need to use ratings in its valuation else it will make really bad decisions. Right now I have it set to 30/25/25/20 and I still have a few problems with the AI overpaying some player but its not nearly as bad. Most salaries are definitely in line with performance. The only problem now is the common issue with the AI not signing superstar FA's even when the teams are flush with cash. I dont know how to fix this but luckily it is not happening that often any more. I did have a few strange seasons where Ruth and Hornsby went to spring training unsigned but it looks like the problem has sorted itself out since then and I am going into the mid 30's now.

Side note: Willie is just an example of how bad the AI is when you use stats only. Granted, the AI always does something stupid no matter what setting you use, but in this case, Willie had back to back good seasons with over 45 steals and near .290/.390. Of course, after that, he fell off the face of the planet but the AI had him contracted on an escalating contract for 6 years with a player option. He was making an average of about $7m a year which is near top dollar in that league (super star typical set to $6.0m). No offense to Willie fans (he was a good player) but he is not top dollar worthy. My guess is that if the AI were using ratings then Willies salary would probably be around $2.5m-$4m like most good to star players.

[marc5477]

Originally Posted by The Wolf

And this is why you always use fictional players.

Yep, especially if you want to do an all stat league. In a normal league with ratings, this is not as big of an issue if you can get the finances, AI settings and draft pool under control. My current game is working very well with lowered PCM's, adjustments to financials, and AI set to 30/25/25/20. In this league, historic rookies have been a boon to the AI because they always draft them high and 95% of the time, they are picked up in FA at good salaries when they are at that stage. Havent really had an issue since the Hornsby, Ruth incident in the early 20's.

[The Wolf]

Originally Posted by marc5477

I think you misunderstood... it was too easy due to how badly the AI handles stats only. By broken, I mean my team was so overwhelming that the game was no longer worth playing. It was like the All American Basketball Team at the Olympics. It was just a funny joke.

I just loaded the save to grab the numbers. I actually played 5 seasons in that game and simmed 3 at the end without even bothering with any transactions and automating the draft and minors. In that last 4 seasons my team had between 100-130 wins and won 2 pennants. No big deal right? Well the payroll was about $14m... The league average is about $30m (just eyeballing it) and the top team (Yanks) had a payroll of $54m. Still, all this would be ok until you see my farm and that was the death knell for the AI. Their farms stunk while mine had a AAA team that could have probably won a pennant or 2 at the major league level.

I can discuss the strategy (upon request) but I really dont want to spoil the game for those enjoying it. Bottom line though, it was too easy. The AI need to use ratings in its valuation else it will make really bad decisions. Right now I have it set to 30/25/25/20 and I still have a few problems with the AI overpaying some player but its not nearly as bad. Most salaries are definitely in line with performance. The only problem now is the common issue with the AI not signing superstar FA's even when the teams are flush with cash. I dont know how to fix this but luckily it is not happening that often any more. I did have a few strange seasons where Ruth and Hornsby went to spring training unsigned but it looks like the problem has sorted itself out since then and I am going into the mid 30's now.

Side note: Willie is just an example of how bad the AI is when you use stats only. Granted, the AI always does something stupid no matter what setting you use, but in this case, Willie had back to back good seasons with over 45 steals and near .290/.390. Of course, after that, he fell off the face of the planet but the AI had him contracted on an escalating contract for 6 years with a player option. He was making an average of about $7m a year which is near top dollar in that league (super star typical set to $6.0m). No offense to Willie fans (he was a good player) but he is not top dollar worthy. My guess is that if the AI were using ratings then Willies salary would probably be around $2.5m-$4m like most good to star players.

Sorry, this whole topic went right over your head. You don't use stats only with historical or current leagues, you use it with fictional leagues. It didn't work because you did it wrong.

[marc5477]

Originally Posted by The Wolf

Sorry, this whole topic went right over your head. You don't use stats only with historical or current leagues, you use it with fictional leagues. It didn't work because you did it wrong.

You are wrong bud.

Historic and current are not what broke the game. The results would have been the same even if all fictional. The triggers that caused the game to fall apart were not the players specifically, but rather valuations by the AI on its own teams. Recall, outside of the draft, the AI does not see "historic" or "real" players, it only seed numbers. Willie Randolph could have been named Weird Al Yakovich and the results would be the same.

What historic does is give the player an edge due to knowing the players. It changed nothing in the AI behavior. Further, I never said I had recalc on so we are using "historic" very loosely here.

You can use historic just fine with stats if you turn recalc off. It is actually a load of fun that way because you get to use all the famous players but you never know what the results will be. For instance, in my current league, the best closer in all baseball is... Ben Tincup. In real life, he barely played a few seasons. Babe Ruth is never going to get near his home run record due to injuries and not getting signed the 1st time he was drafted. Ty Cobb is still playing at age 44. Jim Bottomly hit for a record .448 BA in 1924 (in a league season that averaged .238). And the single season home record is held by Akihiro Okada... an international FA who hit... 86 home runs. Yep it was nuts and no I did not have him on my team. The next closest leader that season barely had 40. Too bad he entered at age 30.

The problem is with AI valuations in stats only mode.

[BIG17EASY]

Originally Posted by marc5477

You are wrong bud.

Historic and current are not what broke the game. The results would have been the same even if all fictional. The triggers that caused the game to fall apart were not the players specifically, but rather valuations by the AI on its own teams. Recall, outside of the draft, the AI does not see "historic" or "real" players, it only seed numbers. Willie Randolph could have been named Weird Al Yakovich and the results would be the same.

What historic does is give the player an edge due to knowing the players. It changed nothing in the AI behavior. Further, I never said I had recalc on so we are using "historic" very loosely here.

You can use historic just fine with stats if you turn recalc off. It is actually a load of fun that way because you get to use all the famous players but you never know what the results will be. For instance, in my current league, the best closer in all baseball is... Ben Tincup. In real life, he barely played a few seasons. Babe Ruth is never going to get near his home run record due to injuries and not getting signed the 1st time he was drafted. Ty Cobb is still playing at age 44. Jim Bottomly hit for a record .448 BA in 1924 (in a league season that averaged .238). And the single season home record is held by Akihiro Okada... an international FA who hit... 86 home runs. Yep it was nuts and no I did not have him on my team. The next closest leader that season barely had 40. Too bad he entered at age 30.

The problem is with AI valuations in stats only mode.

I haven't had a chance to dive into stats only yet, although I plan to start over the holidays. So I may be speaking out of turn here. But here's the thing that jumps out to me in what you're saying.

AI evaluation is fully customizable, so maybe you need to adjust those settings to stop the problem you're having. Considering there have been lengthy discussions about stats only on these boards for the past few years and nobody has ever had the problem you're having, I'm guessing that if it's not your AI evaluation settings that need to be tweaked, it's another setting that may not be right.

[The Wolf]

Originally Posted by marc5477

You are wrong bud.

Historic and current are not what broke the game. The results would have been the same even if all fictional. The triggers that caused the game to fall apart were not the players specifically, but rather valuations by the AI on its own teams. Recall, outside of the draft, the AI does not see "historic" or "real" players, it only seed numbers. Willie Randolph could have been named Weird Al Yakovich and the results would be the same.

What historic does is give the player an edge due to knowing the players. It changed nothing in the AI behavior. Further, I never said I had recalc on so we are using "historic" very loosely here.

You can use historic just fine with stats if you turn recalc off. It is actually a load of fun that way because you get to use all the famous players but you never know what the results will be. For instance, in my current league, the best closer in all baseball is... Ben Tincup. In real life, he barely played a few seasons. Babe Ruth is never going to get near his home run record due to injuries and not getting signed the 1st time he was drafted. Ty Cobb is still playing at age 44. Jim Bottomly hit for a record .448 BA in 1924 (in a league season that averaged .238). And the single season home record is held by Akihiro Okada... an international FA who hit... 86 home runs. Yep it was nuts and no I did not have him on my team. The next closest leader that season barely had 40. Too bad he entered at age 30.

The problem is with AI valuations in stats only mode.

No, the problem is the wrong way that you did everything. YOU DON'T DO IT WHERE YOU HAVE PRECONCEPTIONS OR FOREKNOWLEDGE ABOUT THE PLAYERS. The only problem here is you and your own failures; the game works fine, evaluation works fine and stats only works fine. You just have to be able to follow directions.

[marc5477]

Originally Posted by BIG17EASY

I haven't had a chance to dive into stats only yet, although I plan to start over the holidays. So I may be speaking out of turn here. But here's the thing that jumps out to me in what you're saying.

AI evaluation is fully customizable, so maybe you need to adjust those settings to stop the problem you're having. Considering there have been lengthy discussions about stats only on these boards for the past few years and nobody has ever had the problem you're having, I'm guessing that if it's not your AI evaluation settings that need to be tweaked, it's another setting that may not be right.

You are sort of right. But there is nothing you can do about the AI core if you make it use stats only. You are seriously handicapping the AI because it does not seem to have the ability to think "an extra move ahead." Here is how I broke my league.

When evaluating players, what do humans do? Sure, we look at stats but we also look at player size, strength, speed, etc. In an all stats league, you obviously cannot do most of that since you are depending entirely on their historic stats. So if you are smart, you basically assume right from the front that you will never pay top dollar to anyone and you will play based on probabilities. So how do you build a great team in an all stats league? Simple. Play the numbers.

--------------- Spoilers ----------------

Rule #1 - Your farm is the key. Spend on dev and coaching. NEVER bring a player up before they are 24 years old. Reason? 1st you want the stats. 2nd, in their second major league year they will be 26 and you will have huge leverage over contracts since they are on the see-saw of make or break in their careers. You will be able to lock them into 10 year contracts for about 30%-50% of their asking price. Its important to keep them under control (see #2 below). Dont worry about being locked into long term deals. Remember, you are paying them peanuts anyway and if they ever start to stink badly, place on waivers or trade away for fodder.

Rule #2 - Do not overplay rookies and young players until you lock them into long term deals. Always bench them if one looks like they are breaking away into super stardom. Wait till they are 26 and see #1 above.

Rule #3 - Play the numbers.

Batters: Look for players with at least 600 AB, .250 BA (or slightly below your league average), .320 OBA, < 80 SO, & good fielding stats (fielding is very important). Dont worry about any other stat like HR, RBI, Runs, etc (though injuries is important). These guys will always be super cheap and odds are, since they have a good eye, they will have break out years often (50% of the time). In break out years, they will hit .275/.350 which is what you want. If you hire 10 of these guys (including bench), usually 5 will break out every season and 5 will be the same or slump. Bench the slumpers, keep the breakout guys at the top of the lineup. You will need 2 power guys at #4 and #5. You can hire 1 super star if you want or just do what I do. .250/.320 player with 20+ HR, and (most importantly) < 60 SO. Yes, low SO's is very important.

Pitching: RELIEVERS ARE THE KEY! Yes, you read that right. Why? Because they are very cheap. You can hire the best of the best for peanuts. There is no excuse for not having the best relievers money can buy on your team. Further, you can always convert these guys to closers and they will cost you very little compared to a developed closer. Relievers are your bread and butter. That said, dont bother wasting high draft picks on them because they are plentiful.

For starters: Ignore wins, & losses (they are meaningless in stats league). Here is what you want. Low number of injuries, a lot of losses, few wins, 220+ IP, 3.50-4.50 ERA, < 4 BB/9. Why? Because these guys, again, will be super cheap and odds are they were just on a very bad team. Since you are loading up on good fielders, they will be a lot better on your team and their ERA will drop to around 3.00-4.00.

Thats it. If you do all the above, you will be so flushed with money, you wont know what to do with it so:

Rule #4 - Abuse your money. Use it on the draft and development. You know all those impossible draftees that the AI ignores due to demands? Yep, you should take them. Raise your development budget, and in 5-6 seasons your AAA team will be better than some AI major league teams. You will pump out one super star caliber player almost every year and if you follow #1 and #2 above, you will have them in their prime, and locked into 10 year contracts. (remember to give young ones minor league extensions to keep them on till 24ish). As you can see, it snowballs after that.

With this makeup, you will play about .500 the 1st 3-4 seasons (1st season usually sucks for me). In breakout years, you will go up as high as .600. Once your farm is pumping out stars in 4-5 years, the game will be over.

--------------- End Spoilers ----------------

There are a few work-arounds if you want to give the AI a chance in stats only mode.

#1, You can basically give the AI unlimited money. In this case, it really does not matter if they over pay for a mediocre players. (does not solve the AI farm disadvantage)

#2 Cut your total income to about 1/2 of league average. There are many ways to do this. (probably needs to be closer to 1/3).

#3 No minor leagues or feeders... kinda lame for an all stats league but since the AI sucks at drafting, contracts, and minor league management, it evens out the playing field... actually this is probably still not enough.

My 2 cents

[The Wolf]

Try that in a fictional league with feeder leagues, scouting set on low and trading set on very hard and heavily favor prospects and see how far you get doing that with just your draft picks. Or would that just be too hard?

Periodically someone shows up here and claims of something or the other is broken, despite the fact that it isn't broken for anyone else. Virtually every time (/wave at Enuttage), it turns out that their computer is broken, their league is broken, their game setup is broken, or they are broken.

[curt]

Originally Posted by the wolf

yeah, we deliberately hid it in the middle of the first page.

bastards

[84Hammy]

I'm with the wolf on this one. I think the biggest challenge is the fictional set up. You very well may have a Willie Randolph on your bench, but if you give him three cracks at the lineup and he can't hit better than .230, you aren't going to keep him in, no matter if your scout tells you he is the next Willie Randolph. But since he is a fictional dude named Glenn Moffatt, you won't feel so bad.