Why You Should NEVER Use Statistics In OOTP
ฝัง
- เผยแพร่เมื่อ 20 มิ.ย. 2023
- Statistics are the backbone of baseball. But in OOTP, they can be misleading and harm your ability to assess players. Here is why you shouldn't consider them in player evaluation.
It’s simple! Ootp is a game and not real life, therefore the ratings of a player drives his stats. Never vice versa! So simple yet so many people fail to grasp it! Great video 🎉🎉
I do agree with most points made in this video and for the most part I agree with not using stats as some are flawed based on the game engine.
One key point not brought up is the delayed scouting reports. I've seen numerous examples of hitters drafted in rounds 6 or later with 35/50 power current/potential from my scout with a very high accuracy report within the previous few months to go on to hit a crazy amount of HR's in rookie or low A. Than I request another report and it still doesn't come back as improved from 35/50 rating but all of a sudden in the offseason it jumps like crazy to 55/60. It's clear from those examples that the on field performance predates the scouting report changes so for those examples statistics are invaluable. In your examples in the video you were talking from a 100% scouting accuracy examples but in franchise mode almost everyone I assume doesn't play with 100% accuracy on so my above example becomes very frequent.
I think statistics are helpful. The mistake is looking at a partial season or single season of statistics and drawing conclusions. If you look at the last 3-4 years of stats and weigh them together you often get a more accurate picture.
Players change over time. Using multiple seasons might well have you evaluating a player before he gained 10 points of power, or lost it. By the time you have a big enough sample size to determine a player, he's not the same player, because he isn't static. Not to mention that even over the course of 4 seasons, if the player were to not change at all in rating, you are STILL more likely to vary from the player's talent in stats than your scout is if you have an at all respectable system. Like I said, scouts are hardly ever more than 5 points off, and are often exactly on point for MLB players, even in the worst built scouting systems.
Anyways, my point here is that statistics are always a trap in OOTP and never as reliable as scouting.
@@sgtmushroom any thoughts on how ratings translate to stats? IE what impact does 1point of eye have on BB rate and thus WAR or WOBA or whatever catchall stat? i'm pretty new to ootp but beginning to dive into the stats and my next task whenever i have time is to try to correlate all this tournament data i have with individual ratings. i suspect i'll find similar things that everyone else does with their "secret models." also wondering if these correlations are 1:1 with perfect team and offline mode because the latter seems like a much easier way to collect useful (non-defense) data.
@@nickrockway472 I would recommend looking at my model, MOPS. It should be linked in one of my videos. The model is outdated (built for the 21/22 engine, not 23/24), so just remember K's are less valuable now, power slightly more so. It also inflates batting average. DM me on discord if you want more information.
Thanks for the video. I'm curious how focusing on ratings affects navigating prospects through the minor league system. I previously used statistic thresholds for promotions, but I agree that a rating threshold is ideal. Seems tough/daunting to figure out these thresholds across the different minor levels.
Statistics are important for the sake of player morale. If players have poor morale, they'll have a much tougher time developing/adapting to new levels. Therefore, it doesn't really make sense to promote players who are struggling statistically. That said, ratings are probably the most important benchmark for promotion.
I will say I've had players that have had what I would consider weak ratings wise that consistently performed throughout multiple years. He never looked like a pitcher i would consider good but always had an ERA under 4 and into the very respectful low 3s. His FIP was always good too. I could start him and use him in the pen it was amazing. But it could be a lot of different things that could help that. Pitch type, arm angle and things like that. So you have to use multiple years and ratings. A lot of players can have a big year and it shouldnt be held as a constant thing.
thats most likely because they had good mental attributes... the ones that contribute to their personality. an average player with high mental attributes can outperform a good player with low mental attributes. a player with high intelligence, high workrate and the ability to hand success/failure etc will get the most out of their ratings where as someone with bad mental attributes can often underperform.
Good video, but you should caveat that this only applies to the default mlb roster and historical players. For fictional/ootp generated players, statistics tend to be pretty predictive
Agree with historical if you've got recalc on, since that is how it works. However, I hard disagree for fictional (and historical without recalc, though admittedly a little less so). Fictional games (and future live games) function very similarly to live games, with almost every principle applying reciprocatively. The engine at its core is still the same and development/scouting/performance still functions the same, so there's no reason for this concept to be any different.
Hi Sarge, quick question: let's say I play on "very low" scouting accuracy (but have an ideal HFT scout, of course). Will the scout's evaluations of players be less accurate (even when it says "very high" for accuracy), or does the "very low" setting simply mean that he will need to re-run his reports more often?
There have been a couple instances in my save where a player *really* consistently (talking 3+ seasons) under- or over-performs his ratings. E.g., a player who is in the 60-65 range for contact and avoid K's and ~55 range for everything else (20-80 scale) consistently putting up sub-100 OPS+ seasons. I don't know if it's just 0.1 percentile luck or if my scout can just be wrong about a guy.
Regarding the role of stats, I do think they have some value over a large enough sample. Obviously, I agree with your point on Gimenez that the 100% scouting accuracy ratings are more predictive than past performance (this is not debatable), but I have encountered some players for who either a) seem to always underperform their ratings, as referenced above, or b) have a weird ratings profile where I cannot immediately set an expectation for what his performance will be based on his ratings. And in these cases, I do think stats are useful. But I don't disagree with anything you said in the video.
Scouting accuracy is how long it has been since your scout scouted that player, and how many times that player has been scouted recently (more times scouted = higher accuracy). A scout's ratings determine how close their scouting will be to a player's ratings, while your scouting budget determines how often players get scouted. So if you have a "very low" scouted player, that just means it's been a while since he's been scouted (however, if changes happen in that time, and it's very possible they do, you won't have that information until running another scouting report).
In response to your point about ratings not always being predictive and players chronically underperforming or overperforming expectations, and I said this in another comment on this video, it could be something particular about the player's ratings, his morale, or interactions with coaching/other players on the roster. There is usually an explanation for it, or of course, it is possible that you just have either very bad or good luck, since it really isn't THAT improbable given how variable OOTP is, and given the hundreds of players in the MLB at any time, it shouldn't be a surprise someone is having that.
Furthermore, in regards to full seasons of stats, they're still going to be more variable than effective scouting. There is just almost no circumstance where it makes any sense to look at stats, period.
@@sgtmushroom Thanks for the reply, to clairify, I mean that I play on the "very low" scouting accuracy game setting, like you recommended in a video a while back. Does this increase the likelihood that my scout will consistently overrate or underrate a player, or does it just affect how often scouting reports need to be run?
@@paulg8314 Ah, got it. Yes, with lower scouting accuracy, there is going to be a little variance in scouting reports from a player's true talent that wouldn't be there. Ultimately, however, it's still going to be reliable most of the time (especially with investments in the scouting budget and a good scout).
Great Video, I generally agree that stats don't matter, but with some exceptions. Because there actually are things you can not see with scouting. In the editor you will see things like HBP rate and triple rate for hitters, which you can't see with scouting. Same goes for pitchers with their wild pitch rate and balk rate.
In the example of Andres Gimenez, he should not hit .328, it's just good babip luck for that season going his way and it boosts his OBP and Slug, making his stats inflated. No there is no way I'm trading for him in that case with inflated value. But I'm sure if you look at his HBP you will see that he gets hit a lot, in my simulations Gimenez constantly gets HBP 30+ times a year, compared to 5-8 for the normal everyday player. A walk is as good as a single, and a HBP is as good as a walk. He draws 20-30 more HBP than your average player and that is a huge contributing factor to his OBP. Therefore if he is having a season where he hits like .230 with an average walk rate, but still has a good OBP. I know that it is because his HBP rate is high.
Same goes for pitchers, some pitchers have good ratings but never good stats. Especially randomly generated players can have wacky Wild Pitch rate and Balk rate. I see a pitcher with bad FIP and ERA for the past 3 seasons but my scout has told me he has always had good ratings the past 3 years? I have to look into the stats, is he having 30+ wild pitches and 10+ balks a year in 150 innings and that's why he is struggling? Or is his catcher and defence or even manager strategy just bad?
The point is that yes stats in vast majority of scenarios don't matter, but there are cases worth looking into.
How about for a player severely underperforming to what their ratings suggest they should be? Also, how is Jones only a 3.5 star?
I find that I use statistics solely for edge cases and for things that don't show up on the player profile (Hit by pitches and the like). For a high stuff reliever with 40 control, sometimes the control plays more towards 45 and is less problematic than a guy who hovers between 35 and 40. Same with range where I can have a CF whose range appears at 70 but can play towards a 65 (pretty average) or can play towards 75 (gold glove caliber).
Very true, and good point especially about hit by pitch. Though I would generally agree that statistics can sometimes help determine the edge cases.
Depending on your settings stats can be useful as well. There is a percentage for stats, current year, last year, and the year before. I wondered about that myself but it appeared that my players who had a bad year continued to underperform for the most part. I'll be curious to see how things go next season as I've wound up resiging much of the Blue Jays for 2024 (out of a weak FA pool, and what the AI considers overpaid vetrans). I also made some bad blunders in the offseason.
How do you adjust your lineup throughout the season? Or do you bother?
Day to day
Also not even to mention essentially every single weighted stat in OOTP is not calculated correctly
The ratings are what the players are not the stats.
Is there a "form" stat in the background? Can reading stats of a season help you make a decision to sign a player around the deadline based on how he is playing that year?
What do you mean by "form" stat? Stats can kind of help identify deadline targets. If you look for players who are playing well below their ratings, it can be useful to pick them up, expecting that they'll turn their seasons around.
@@sgtmushroom I meant if there was a stat that determines whether there is a higher or lower chance of someone having a good or bad season, playing above or below their attributes for a time. It happens irl, was just wondering if it was modelled in the game at all.
@@NickKrige You might be thinking of morale, which affects player performance. Morale is determined by a number of things, including how the player is doing, how the team is doing, how the player is being used, and how the player feels about transactions of the team. There are some personalities that affect it too. The rest is just random variance. There's nothing "sustained", but OOTP is more than random enough for a player to have wildly up or down seasons, even in a significant sample. I would recommend you visit the editor of a player though and go through it. Everything used in the engine can be found there.
I'm not sure I'm totally understanding the applications of your examples. Your examples seem to show players who, statistically, are outplaying their ratings, but we should discount those stats because of their low ratings (regardless of sample size)?
On the flip side, how do we approach players with great ratings but are, statistically, getting killed? Take them in hopes that a different situation will allow them to thrive?
but have you considered the fun and realism of factoring in statistics to some extent
What would you recommend setting AI trade value to then?
As in the trade difficult setting? IMO, even maximum difficulty isn't really that difficult, so for OOTP veterans I think maximum difficulty is definitely a reasonable "challenge" on 24.
@@sgtmushroomno for the how much they favor ratings, and stats from the previous years
@@Ian00003 Technically you could probably do 100% ratings and be totally fine.
Top scout matters a lot.
In a fictional league where current ratings are hidden and you only see potential, would you still only want a highly favors tools scouting director? Do stats have meaning in that format?
This complicates what the point of the different type of scouts is again. Are they just wrong towards their point of bias? Tools towards 'potential' skill levels, Ability towards players whose ratings are reflective of their current stats? Gimenez in that example I think had a 4WAR season followed by a 6WAR season (estimating, not replaying). If OOTP will create several seasons of powerful anomalies to that extent he's performing past the WAR the ratings suggest, it's no more than a slightly suggestive 'safe' bet I see a lot of variation in OOTP, I'm not sure I see as many as I even do in real life of players with a freakishly strong 2 years. I've had several highly rated players who were duds and I hung on to for too long. I'm not suggesting that Stats are 'useful' just that ratings are maybe not that useful.
so ideally we should use scouts that scout ability not use tools?
Vice versa. Tools scouts over ability
What people often fail to realize is age of player and league level play a factor. I hate using “Ratings Relative to Level” but a 30 year old with average MLB ratings is going to crush A-level players. Thus, artificially inflate their stats.
If a prospect player has hight ratings but is not performing in the minor leagues its more than likely that you have him at too high of a level, for instance if he's in double A then you should drop him to single A, a lot of times people try to progress their young players through the minor leagues too fast.
Stats are useful for inflating my players values in my online leagues in trade talks though, did you consider that? no!
If you're using the default view it doesn't surprise me you don't view stats as useful. In the first case, you aren't showing BABIP in the ratings which can very much artificially inflate a players numbers just like it can in real life. In this case, it looks like Gimenez's batting average is about 30 points higher than his real life 2022 and that's the only reason his war is higher. I suppose he could have also had a better defensive season too but obviously those numbers aren't shown. But his 139 OPS+ is basically identical to his 140+ in 2023. So, assuming he's 2 or whatever wins better is just not really understanding how to use the statistics. There's other stats i'd also prefer to see like ISO, walk rate and k rate but point being it's not enough to just look at a guys triple slash or WAR. You want to know why his WAR was the way it was.
In the second case, you're playing a 19 year old in A+. For reference, this season in the midwest league(A+), the average batter age is 22.3. If i have a player who's 3 years younger than the average age of a league putting up a 102 OPS+ I'm thrilled.
There are numerous players in each iteration of the game who play better than their overall ranking would otherwise say just because the shape of their ratings. Stats are great for finding those types of players but you have to understand what you are looking at and be able to parse out why a player's war is what it is. For example, if you have a slugger with terrible avoid k who has a monster year you should not be expecting that again because K rate is notoriously streaky. Like wise, if you see a guy who has a terrible year with like a .200 BABIP but his other numbers like bb%/k%/iso look fine he's potentially a player the AI will devalue in trade but is a reasonable guess to improve the following season.
Regardless, I wouldn't say stats are useless in evaluating a player. I also find they are very helpful for determining promotion as a prospect. For example, if you have a good prospect with a below average avoid k he may be able to put up a good year at a level with a high k rate. However, if you promote him before his k rate is ready you can ruin the player because that higher level could tank his season the following year.
For the most part, you're at least partially right here. A number of your points are fair, and I probably should've added a disclaimer at the end that players whose statistics are worse than expected can be used to identify undervalued players to acquire, and that statistics can sometimes be useful in prospect promotion.
So, a couple notes here. You are mostly right about the statistics view point. It would be useful to see Gimenez's BABIP, but it is important to consider that players in OOTP can absolutely sustain higher or lower BABIPs with their BABIP rating, and yes, his BABIP overperformance does largely explain why he was so successful. However, it should be his ratings and not the fact that he has an outlier BABIP that would give you that information.
Your second point about comparing Gimenez's 2023 season in this simulation to his real life 2023 is definitely not a good idea. OOTP does NOT equal real life, and comparisons between the two are almost never viable.
Claiming that I don't know how to use statistics because I said he's 2 wins or so better than I'd expect is... wrong? Disrespectful? You take your pick. I don't really know what your point is here since you just sandwiched what seems like a nothing comment in between a relatively well written paragraph? The reason I made that comment is because I was trying to keep this video simple, and making a rough comparison between his overall performance and his ratings was a simple, straightforward way to make my point.
As for ISO, K%, BB%, this claim is a fundamental failure to think about OOTP from the perspective of OOTP rather than real baseball. I would repeat, this is NOT real baseball. Trying to evaluate OOTP players as you would someone in the real world simply does not work. If you really want to evaluate a player statistically in OOTP, you need to do it within the context of order of operations. So walk rate, strikeout rate in at bats (so NOT normal strikeout rate, which is plate appearance based), home run rate when not striking out or walking, extra base hit rate when not striking out, walking, or homering, and then BABIP when not striking out, walking, or hitting for extra bases. This is how the engine works, and this tells you what a player is doing in the context of the engine. Technically speaking, since the variability compounds towards the end, this will actually give you a decent degree of accuracy on player evaluation, but even STILL it's less effective than just using ratings. Trying to use advanced statistical analysis with things that might give you insight to a real player's value like ISO and true K% won't tell you jack in OOTP.
The point about Jones, once again, seems to be based on real baseball principles. This seems to be the running theme here. OOTP is not real baseball, so you absolutely need to be treating it differently. In OOTP, players stop developing almost completely around age 24, and development slows significantly once you get around 22-23, so you need to get your prospects going quickly. Having a 22 year old at A+ basically guarantees that they're a non prospect. So yes, for a top prospect, you'd absolutely expect them to already be at A+ by age 19-20.
For your point about players outperforming their overall because of their specific ratings... I just have to assume that you've literally never watched my channel before. And frankly, I have to assume you weren't paying close attention to anything I was saying, because a 3.5 star potential isn't a "top player", while Jones absolutely is in my eyes. I am a huge proponent of analyzing players by their individual ratings and strongly believe that a player's overall rating is completely useless and rarely correlates to their real talent. I am very well aware that players with certain batting profiles will perform differently. I have literally no idea what your point was there, because it has nothing to do with anything I claimed in this video.
As for the AI devaluing players who underperformed, absolutely agree. Once again though, I'm not exactly sure what the point is here.
Statistics are kind of important in promoting prospects, because you want to make sure they are able to produce statistically and that they have been producing statistically before promoting them. The first part I'd determine with ratings though, not statistics. The reason for that is that statistics can impact a player's morale, which in turn affects both their ability to produce statistically and develop as a prospect. Strictly speaking, if a player is happy with bad performance and has the ratings to perform at the next level, I would probably still promote them.
How many more videos are coming up before your retirement? Love your content and you will be missed.
I'm planning at least 3 more, with possibly a couple others as well.
We should not evaluate minor league players like the way we do in MLB. For example you can check ERA for MLB pitchers but K/BB is more important than ERA when evaluating a AA pitching prospect. I understand what you mean but 'NEVER' seems a little bit relentless.
Star ratings are also useless
Disagree. Stats are very useful. TCR turned Eric Hosmer from Padres (OOTP 22) into a yearly .300 hitter yet his contact and eye rating remained 55. After 1 year, the scout changed overall and potential ratings to 3 star.
Ratings do not take into account if a player is hot or not. You like to sim years at a time where I play each game (one pitch mode). You lose something simming and not watching/managing each game. Streaks is the big one. For pitchers, times thru the lineup and how well the pitcher is doing at that point in the game can help you make decisions regarding the bullpen. A sim does not even give you the option to do this.
Of course, the biggest stat that matters is the teams wins and losses. Without that, this game is meaningless.
Through the points you used to support your argument here, you can justify nearly any claim. You gave me exactly one example of something that doesn't even seem directly connected and didn't even supply enough details to start, and then you're claiming that stats are useful because of... micro... situations. None of this relates to the point of the video. As for the claims you're making in regards to managing every game, I strongly disagree as well. It is important to remember that you are playing a SIMULATOR, not managing actual players. In real life, streaks might be a legitimate thing due to how a player's mental state can affect their outcomes, having periods of making better connections, etc., but in OOTP, a simulator, there is nothing that "causes" streaks. They are simply the result of chance. The only thing that can cause "streakiness" is morale, and theoretically, that can just be modeled and factored into player evaluation on the surface level. Statistical performance can affect morale, but it's the morale itself that affects player performance, not statistics directly. And as for when you pull a pitcher... that is totally irrelevant to this video. However, "how well a pitcher is doing" is not indicative of future performance at all. Once again, this is a SIMULATOR, and there is nothing in the simulator that causes this. Blowups are typically the result of simple bad luck, and pulling your starter after they give up a couple hits in the fifth inning but only have thrown 65 pitches is not always the smart move. Anyways, yeah, simulating games instead of managing them means you are left to the computer's interpretation of your strategy settings for how it manages your pitchers... but again, how is that relevant? No one is questioning that you would win more games managing your team yourself than letting the AI do it most of the time.
And as for your last point... that wins and losses are all that matters. I can routinely get above 110 wins early in saves, and after a while once I've established my organization I am regularly above 120 and hit 130 often enough. If you are not at that level, do not use this argument with me.
I'm still not really sure exactly what it is the point of this post is though since you didn't really connect to what I said in the video.
@@sgtmushroom We agree to disagree. I did not watch the video at all and only focused on the title because I think the title is ridiculous.
@@BryanBear5050 The fact that you couldn't be bothered to watch the video before disagreeing with it says everything. I won't bother trying to convince you, but please do not post anything publicly where others may see it and be confused unless you are qualified to do so... which in this case your comment is clearly not.
BaseBALLS